In the last two decades, we’ve seen remarkable progress in mental health.

Particularly, in getting people access to care.

What was once a taboo subject is now openly discussed in homes, workplaces, boardrooms and government offices around the world. Stigma has fallen, insurance coverage has expanded, and millions of people who would once have gone without treatment can now access it.

The access movement had a compelling narrative: that the primary barrier to better mental health was access to effective care.

Over the course of the last twenty years, we have seen huge increases in spending and the availability of care. In the US alone, expenditure on mental health and substance-use treatment rose from $41B in 2001 to $140B in 2021, significantly outpacing overall healthcare spend.

Mental Health and SUD Nominal Spending in the US

With this increased spending, we saw an increase in the number of people who received care. In 2022, 23% of US adults received mental health treatment, an increase of four percentage points from 2019, which equates to a change of 21% in just three years.

Mental health treatment access among US adults.

Yet the success of this access movement leaves a hard question to be answered: why do we continue to see such high rates of mental disorders?

In 2023, 29% of US adults said that they had had depression over the course of their lifetime, up from 20% in 2015. Many more statistics on the prevalence of mental conditions look similar to this chart.

Rates of lifetime depression among US adults

The stakeholders who supported the access movement are beginning to reconsider their role in mental health.

Financially, employers, insurers, and governments face rapidly rising mental health costs. Tim Noel, CEO of UnitedHealthcare, recently noted that behavioural health costs are increasing between 20% annually. Employers have expanded mental health benefits, yet many continue to see rising absence, disability claims and lost productivity. Increasingly, they are asking not whether people should receive care, but which forms of care demonstrably improve outcomes and justify continued investment.

Politically, governments and research funders are becoming more cautious. In the US, we have seen significant cuts to research spending and many mental health programmes. Year to date (as of July 2026), the number of grants the National Institute of Mental Health has funded is down 88%, versus the same period in 2024. The Make America Healthy Again movement has brought attention to psychiatric prescribing and concerns about over-medicalisation, particularly among younger people. The political conversation is shifting from expanding access to defining the appropriate role and boundaries of mental healthcare.

Culturally, new narratives are emerging. Some argue that we have over-pathologised normal human distress. Others go much further, questioning the legitimacy of mental illness itself. Many of these arguments don’t withstand scrutiny. But they are gaining momentum and reflect a wider societal unease about what is happening to mental health and whether the existing story still explains it.

Simply continuing to increase access to care will not keep these stakeholders on board.

The field will need to answer some hard questions and create a new narrative that addresses the reality of 2026’s challenges.

Narratives are how complex systems organise themselves. They shape what governments fund, what employers buy, what investors back, what researchers study and what ideas founders pursue. The access movement succeeded because it aligned these groups behind a common understanding of the problem and a broadly shared theory of change. Without a similarly coherent organising story, the mental health ecosystem risks pulling in multiple directions and losing support of key stakeholders along the way.

This doesn’t suggest that support for mental health is disappearing. Across the field, hundreds of thousands of clinicians, researchers, advocates, founders and policymakers continue to dedicate their careers to improving mental health. Millions more support the cause because of their own experiences. The motivation to improve population mental health remains strong.

The question is whether that motivation can be organised around a sufficiently clear account of the problem and the path to progress.

To create a new narrative, the field must explain why psychological distress remains so widespread despite two decades of increased spending and access to care. It must offer a more compelling account of what is driving this distress, and provide a clearer vision of where the next generation of solutions will come from.

There are many potential answers to these questions.

Decades ago, biological explanations dominated public understanding of mental illness, epitomised by the idea that depression resulted from a “chemical imbalance”. That account has largely given way to the biopsychosocial model, which recognises that mental health emerges from the interaction of biology, psychology and social circumstances. It is almost certainly a better scientific model. But it is a much weaker public narrative. If genes, relationships, work, housing, technology and economic conditions all shape mental health, what, exactly, is driving rising distress? And where should society look first for solutions?

Simple explanations, by contrast, gain traction and drive change. Jonathan Haidt’s Anxious Generation movement is a clear example. Haidt offers a straightforward account of cause (social media use) and effect (the decline in adolescent mental health), which has become one of the most influential narratives in global health. Australia has already banned social-media use for under-16s, and the UK is pursuing similar measures. The actual efficacy of these policies is still hotly contested among academics, and yet huge action has been taken.

This challenge is compounded by uncertainty over what exactly is changing. Distinguishing between worsening illness, greater awareness, and changing definitions of impairment is becoming a central challenge for mental health policy.

For example, in recent years, the number of young people reporting mental-health conditions has risen sharply, alongside increases in reported anxiety, ADHD, autism diagnoses and disability. Yet the picture becomes less clear when looking at functioning. Analysis of British data by Christoph Henking and Ben Baumberg Geiger found that while the share of young people reporting a mental-health condition increased substantially, the share saying that a mental-health problem limited their day-to-day has changed much less.

Source: Financial Times

At the most severe end of the spectrum, however, in several countries, hospitalisations for self-harm among teenage girls and young women have risen sharply.

Source: Financial Times

This data supports a narrative of a teen mental health crisis and could be used to support policies like social media bans.

And yet, other analyses have shown that in many Western nations, teen suicide has actually decreased as social media use has risen. This includes the UK, where from 1995 to 2023, the suicide rate for people under twenty fell by 22%.

Source: There’s no “global increase in teenage suicide”. Global Burden of Disease, University of Washington School of Medicine Institute for Health Metrics and Evaluation 2026. Suicides are deaths from “self-harm.” Rates are suicides per 100,000 population age 10-19 by country. “Social media era” is 1995 through 2023, with all countries supplying data for all 28 years.

If this picture feels confusing, it is because it is. The problem is not simply that mental health is difficult to measure. It is that different measures tell different stories. And when we see multiple stories, it’s hard to align on where progress should come from.

What remedy are we proposing?

Over the years, I’ve spoken to researchers, clinicians, policymakers, founders and public health experts about what we need to do to create a mentally healthier world. There is little consensus.

Some look to better pharmacology, others to AI or neurotech, while many argue that, unless we change the social and economic environments in which people live, rising distress will continue regardless of healthcare innovation.

The theory of change implicit in this approach has therefore become more difficult to communicate. It suggests that mental health is influenced by many interacting factors, measured imperfectly and addressed through an increasingly diverse set of interventions. All of that may be true, but it is the kind of complex and nuanced narrative that struggles to organise cultural and economic forces.

Finding a new narrative

While defining specific metrics is challenging, most people would actually find it easy to agree on the world they want to see. It’s one where we limit unnecessary suffering, enable people to function in their families, workplaces and communities, and create the conditions in which people can flourish. Achieving that vision will be difficult. It will require a clearer understanding of the problems we are trying to solve, the outcomes we are seeking to achieve and, critically, where the next generation of solutions will come from. Most importantly, this understanding must be packaged in a compelling narrative that provides direction.

One possible route is to stop presenting mental health as a single problem. Severe mental illness and more common conditions, such as anxiety and depression, share important features, but they are not the same challenge. SMI requires better treatments and support for a smaller group with profound needs. The latter involves a much larger population where the boundaries between healthcare, normal distress and social conditions are less clear. Presenting these as distinct narratives to stakeholders could be beneficial.

Whatever route it takes, the field now needs a new organising narrative: one that is clear, compelling and addresses the challenges ahead. It must explain not only why mental health matters, but where solutions will come from and what role governments, clinicians, employers, researchers, funders and innovators each have to play. The field may have to face some uncomfortable questions about whether it wants to be right or effective.

The access movement demonstrated what a shared narrative can achieve. The next phase of progress will depend on finding one again.

That is all for this week. If you enjoyed it or have thoughts on this topic, please do reach out by replying to this email.

Keep fighting the good fight!

Steve

Founder of Hemingway

P.S. If you find these updates valuable, consider becoming a Hemingway Pro Member. As a Pro Member, you will get access to even deeper industry commentary and will be invited to join our community of over 500 mental health leaders.

In fundraising, your pitch is your story. So what does a great pitch look like for a mental health business in 2026?

There are some general things all founders need to get right when pitching — the internet is full of helpful tips on this topic, and I’ve included the best resources in the notes.¹

But there are also elements of a great pitch that are highly specific to mental health. These take into account the nature of the market in 2026 and the experience of the investors who will be listening to your pitch,

Finding that specific advice is much harder, which is why I’ve written this post. After reviewing hundreds of decks and speaking to several founders and investors, this post captures the top tips I’ve learned for pitching your mental health business in 2026.

Let’s get into it.

Subscribe now

Key Takeaways;

1. Pitch your business first, impact second

2. Keep your personal story short

3. Be different. Don’t fall into the sea of sameness in mental health

4. Address the “mental health is a niche” objection upfront

5. Prove beyond doubt that there is a willingness to pay

6. SHOW ME THE OUTCOMES. Include validated outcomes

7. Don’t be afraid to estimate

8. Show your ABZ plan for commercial success

9. Keep it simple

10. Take away their reasons to say no

11. Up your confidence

1. Pitch your business first, impact second

Many mental health pitches start with the social problem. They show statistics about the impact of mental illness and explain how their idea will solve it. That’s great. But it’s a mistake.

Investors don’t give you money because they care about the problem (even if they do actually care). They give you money because they believe you’ll generate a financial return.

I learned this from Tim Ferriss at the One Mind Accelerator earlier this year. Tim has seen thousands of startup pitches, including many from mental health founders. He says that founders should pitch the business first, and the societal impact second. The reason someone should care is not the same as the reason they should invest.

So start your pitch by showing that you have a genuinely compelling business opportunity. Convince investors that your company can become a great investment. Then bring in the societal impact as the cherry on top.

Stephen Murdoch, an investor, shared his thoughts on this: “Compassion might get you a meeting, but unit economics and clinical credibility that produce measurable outcomes get you funded.”

Also, if someone is going to care about the impact, they’ll care after one or two slides. Any additional slides on social impact are redundant, so get rid of them.

Yes, mental health is an important problem. Yes, investors care about it. Yes, your solution might genuinely improve people’s lives. But your pitch should lead with the investment case, not the impact case.

2. Keep your personal story short

On a similar note, you must make sure your personal story doesn’t dominate your pitch. Many founders start a mental health business because of their personal experience. While that story shows a passion for the problem, it has a limited effect in convincing investors to fund your business. Keep your personal story short and do not let it be the centrepiece of your pitch. Your main pitch should be about your business, the specific problem it’s solving, how you’ll get paid to solve that problem and why you’ll be a giant company one day.

3. Be different

There is a sea of sameness in mental health right now. You cannot fall into it. Investors are seeing hundreds of pitches, and in the mental health space, many of them look the same.

As we outlined in a piece with the One Mind Accelerator earlier this year, AI chatbots and teletherapy are particularly saturated areas of innovation right now. While there may still be room for good companies to get funded here, your pitch must clearly communicate why you are different to all the other companies that investors have seen.

Having a truly unique view of the problem, or a unique product to solve it, is a great way to show differentiation. But it is not the only way. You can differentiate on your commercial model (how you get paid and by whom), or through your distribution (how you reach buyers and users). Perhaps you’ll differentiate on how you actually operate your organisation. I’m still waiting to see a pitch for a mental health company that applies AI within their business to operate at much lower costs than incumbents. I think this is one way an aggregator business could reach >10% net margins in 2026.

You must realise that mental health is now a crowded and competitive market. And it’s not just your business that must be differentiated; your actual pitch must be too. Remember, investors are seeing hundreds of pitches, so you need a way to stand out.

Communicate your unique point of view on the market and have a unique story to tell. Avoid including generic mental health slides. Don’t open your pitch with a slide on the prevalence of mental disorders. Or even worse, the economic cost of mental illness. Firstly, these statistics are irrelevant to the commercial success of your business. And secondly, they are in every other mental health pitch deck!

I’ve found that the best pitches teach the investor something. They give them an interesting insight to consider and make them feel smart.

One way to test if your pitch is unique is to go to each slide and ask yourself, ‘Could this slide be in someone else’s pitch deck?’ If yes, kill it.

Subscribe now

4. Address the “mental health is a niche” objection upfront

Many investors still see mental health as a niche, and one too small to justify venture returns. This is particularly true for generalist investors. You need to address this objection early in your pitch and punt it so far out of the room that it’s no longer an issue. To do so, give them some hard numbers on the size of the specific market that you will operate in.

The mental health market is huge and growing — in 2021, more than $140B was spent on mental health and substance use disorder care in the US. Today, that number is higher (if you can find a reputable source, include it). But more specifically, find the part of the market that you will be fully able to address, show how large it is, why it will grow and why you’ll be able to capture a large part of it.

Also, there are now multiple mental health unicorns, many with annual revenues in excess of $1B. And yet, each only holds a single-digit percentage of market share.

One helpful way to remove this objection is by comparing the mental health market size to another market that investors are already convinced is large — the global ride-sharing market (the arena of Uber, Lyft, Didi and more) is approx. $170B.

5. Prove beyond doubt that there is a willingness to pay

Perhaps the most important part of your mental health pitch is proving to investors that someone will pay for your solution.

Joshua Chauvin, Venture Partner at Awareness Capital, feels strongly about this. “In digital health, the graveyard is full of companies with strong products and no proof anyone would pay”.

Investors are highly aware of this, and so the best mental health pitches answer the commercial question before they have to ask it. As Chauvin explains, “The best pitches are laser-focused on proving willingness to pay early, because everything else follows from it. Including comments like ‘we interviewed prospective buyers, and they’d be interested’ isn’t evidence — a signed pilot with a named budget holder who is committed to putting skin in the game, is.”

Max Rollwage, Co-Founder of Fault Line doubles down on this. He explains that to raise capital, you will need to show that you deeply understand the buyer landscape and can show why they will choose your solution. “Unless you have an extremely crisp articulation for why your buyer will benefit from adopting you, it will be difficult to raise.”

Rollwage adds, “the reason of ‘better patient care’ is usually not a good enough answer as adoption might depend on so many other factors”. Show the incentives of the different parties in your ecosystem, especially your buyers, and exactly how your solutions satisfy those incentives.

FamilyWell is a great example of this. They integrate their behavioural health solution into OBGYN practices. Why do they get adopted? Because their solution actually makes money for the OBGYN practice by allowing them to bill for collaborative care.

6. SHOW ME THE OUTCOMES

Being able to deliver superior outcomes is a competitive advantage in today’s crowded mental health market, and investors care about that.

As Stephen Murdoch explains, “investors need evidence that you’ve found a repeatable way to improve outcomes while building a durable business”.

Show this by including hard data on validated outcome measures. We wrote an entire article on this topic with Jen Huberty, PhD, which explains exactly which validated outcomes investors are looking for at each stage and how you can collect them.

Including hard outcome data is one of the best ways to separate your pitch from others. They increase credibility, show the validity of your solution and signal that you are serious about creating measurable impact.

If Jerry Maguire were a mental health investor, he’d be screaming, “SHOW ME THE OUTCOMES.”

7. Don’t be afraid to estimate

Mental health founders are often hesitant to provide estimations. Many come from clinical or scientific backgrounds where it’s unusual to make approximations with uncertain assumptions. In startups, you rarely have perfect data. However, you still need to give answers to investors. You can’t tell them that you’re not sure what your growth rate will be or how much money you expect to burn. They want certainty, and you need to give it to them by providing clear answers. That requires making estimations, and you should not be afraid of this.

For example, you may not yet know what your actual unit economics are. But you cannot tell that to an investor. You will need to estimate what they will be. To do this, use a logical approach and set of assumptions that you would be comfortable defending if an investor probed.

Investors are smart and know that your projections likely won’t come true. But they need to see that you’ve thought deeply about them. The only thing worse than an estimated answer is no answer at all.

8. Show your ABZ plan for commercial success

When an investor is listening to your pitch, they’re asking themselves, “Can this be a billion-dollar company?” You must communicate with clarity and confidence why you will be. This is the “Z” part of your plan, the final destination, and you must communicate it in a way where it seems like a foregone conclusion.

However, there is another element to your commercial viability that many mental health founders miss - what you will do to succeed in the next 12 months. This is the “A-B” part of your plan.

Max Rollwage, Co-Founder and CEO of Fault Line shared this tip with me. He said, “Be very specific about two points related to commercial viability. First, how will you be a $1B-$100B company in 10 years? And second, how will you create sufficient revenue for the next raise in 1 year?”

He emphasised that you need to provide detail in this commercial plan, not just hand-wavey projections based on capturing X% of the TAM.

“You probably want very specific bottom-up unit economics and, very importantly, a good understanding of the market dynamics and incentives of buyers that would help or hinder you in your adoption.”

8. Keep it simple

They can love you, they can hate you, but they cannot be confused by you. This is another gem I picked up from Tim Ferriss.

Many mental health solutions can be complex, but if your investor is confused by it, they won’t invest in it. They might even recognise that it could be a great idea, but they’ll put it in what Warren Buffett calls the ‘too hard’ bucket, and politely pass on your deal.

Your challenge is to condense everything you know about your business into a single-threaded, easily understood narrative. There must be one thread that ties together the problem, the solution and the commercial plan.

One big mistake I see in pitch decks is strategic hedging. This happens when founders pitch a wide range of options for where they can take the business. They say things like “we will start by selling to private-practice clinicians, but if that doesn’t work, we could also sell to community mental health organisations, or maybe even health systems.” This hedging just portrays a lack of conviction in your main plan. And lack of conviction is instant death in a pitch.

You must keep your pitch simple.

As Joshua Chauvin describes, “the best pitches are ruthlessly simple: no more than 10 slides, and probably 90% fewer words than you think you need.” I can’t emphasise this enough, please cut words from your slides. And, don’t include more than one chart on a page.

Long decks full of words and multiple charts muddy the message. As Chauvin explains, “A dense deck signals a founder who can’t distil their own business.” Not only does that make it harder for investors to evaluate your idea, but it signals that you will have trouble communicating it to other stakeholders, like customers, partners and employees in the future.

10. Take away their reasons to say no

Before founding Hemingway, I worked at a Fintech startup where we raised over $200M. The founder of that business gave me some great advice when it came to pitching investors. He told me, “Instead of giving them reasons to say yes, just take away all their reasons to say no.”²

So why would a mental health investor say no? Many of the reasons we have already discussed in this article. They might say the market isn’t big enough or that they don’t believe your unit economics are good enough. They might say that they don’t think you are differentiated enough or may not believe you have a buyer who will pay for your solution.

As you are crafting your pitch and testing it with people (because you should be testing it), write down the reasons an investor would say no. Then, address these objections in your pitch.

11. Up Your Confidence

My experience is that founders from other industries (like Fintech or SAAS) pitch with much more confidence than the average mental health founder. I’ve seen many bad ideas raise money from such overconfident founders. You can learn from this.

Investors want to feel confident in their decision to give you money, and being highly confident in your own business and your own pitch will rub off on them. So make sure you have an idea and a plan that you deeply believe in. Get people to poke holes in it and then fill those holes. And once that’s done, deliver your pitch with the supreme confidence of a twenty-year-old Stanford dropout with an AI startup.

That’s all for this week. Many thanks to Max Rollwage, Stephen Murdoch, Augusta Peytz and Joshua Chauvin for their contributions to this report.

Got another tip you’ve found useful when pitching? Please share it with me. My aim is to provide content that is helpful for people innovating in mental health, so contributions are always welcome.

Keep fighting the good fight!

Steve

Founder of Hemingway

Notes:

[1] These are the best general resources for fundraising and pitching:

• Y Combinator Guide to raising a seed round

• Fundraising wisdom from First Round

• Fundraising tactics from Kevin Ryan

• How to design a better pitch deck by Y Combinator

[2] I think he learned this from an episode of Billions, but it’s good advice all the same.

[3] I know that raising venture funding is not for everyone. I will be sharing more guides soon for those who can’t / don’t want to raise funding in this way. Just keep an eye on your emails!

Subscribe now

This is a Hemingway Pro article. To get access to more content like this, consider becoming a Hemingway Pro member. You can join here today for 50% off your first 3 months.

In mental health, aggregators are everywhere.

These businesses aggregate the supply of clinicians in a given market and then offer it to people who want care. Many of them are now large, profitable businesses. Some bring in over $1B in revenue each year, and as a category, they have attracted more investment than any other.

But despite the success of some of these businesses, we have yet to see a truly dominant player.

In the US, even the largest mental health marketplaces are responsible for only a single-digit percentage of the market. They aggressively compete for users, clinicians and payers, with no business developing a position where they have true defensibility — and the growth rate and margins that would come with that.

This is not how aggregation plays out in other industries.

Usually, a handful of aggregators (or sometimes even a single aggregator) develop a dominant position with a very large market share. With this dominance, they achieve better unit economics, which further entrenches their leading position. Think of Uber in ride-sharing, Airbnb in vacation rental, Spotify in music, or Skyscanner in flight booking.

This hasn’t happened in mental health. Or at least, it hasn’t happened yet.

The reason is that no one has truly aggregated demand.

In this article, I discuss why this is the case, how aggregation theory works in mental health and what businesses can do to organically aggregate demand and achieve dominance in this space.

Let’s get into it.

Subscribe now

Aggregation Theory is a concept coined by Ben Thompson to describe the success of most large tech businesses in the internet era. According to the theory, the internet drove distribution and transaction costs to zero, which has allowed tech businesses to aggregate the supply and demand within a given industry. With this aggregation, they can create marketplaces, and these marketplaces are amazing businesses — they have high margins, can grow quickly and are defensible.

In the 2010s, many founders were aware of Aggregation Theory and the attraction of marketplace businesses. In the mental health space, some of these founders identified a problem which they believed could be solved with aggregation — finding a therapist or a psychiatrist. The simple idea was to aggregate clinicians (the supply) to make it easier for clients (the demand) to find care. And so these founders started a variety of mental health aggregators. Today, they are the biggest businesses in the industry.

But despite their success, no player has yet established a dominant market share position. Even the biggest players, like Grow, Headway, Alma, Talkspace, Talkiatry or BetterHelp, still own only a single-digit percentage of the entire market. This has constrained their valuation multiples and made it harder for them to exit opportunities, specifically through an IPO. So why hasn’t one of these well-run, well-capitalised businesses achieved a dominant position?

It’s because no mental health business has truly aggregated demand.

Instead, they’ve all focused on aggregating supply. This has served them well to date, but true market power comes from aggregating demand, not supply. The technology investor Bill Gurley explains this well:

“Aggregating suppliers is a necessary, but insufficient step on its own. You must also organically aggregate demand. With each step, it should get easier to acquire the incremental consumer AS WELL AS the incremental supplier.”

When you truly own demand, you develop negotiating power with all players in the ecosystem — including payers and suppliers. Many believed that negotiation power with payers would come from aggregating supply, and that once enough supply had been aggregated, businesses would be able to secure higher rates with these payers.¹ But this has not proved to be the case. Today, many of these businesses are struggling to secure good rates — in May 2026, Aetna actually cut rates for therapists on one of the biggest marketplaces in the US, Alma. Talkspace faces similar rate challenges.

Organically owning demand also has huge implications for client acquisition costs and, in turn, growth. Look at Talkspace. The business now has over 200 million covered lives, meaning theoretically, there are tens of millions of people who need therapy each year and could get it for free (or close to it) through Talkspace. And yet Talkspace therapists see a few hundred thousand clients each year. Why? Is it because they don’t have enough therapists? No, it’s because they do not own the demand. They must pay to acquire each new client amongst a competitive pool. This keeps CAC high and limits growth.

If organically aggregating demand is so important, why haven’t mental health aggregators done it? First, because it’s hard. But more accurately, it’s because they haven’t needed to.

Mental health is different to other markets. Supply is heavily constrained, and payers obscure market dynamics. The supply-demand imbalance has been so great, especially from 2020 to 2022, that all an aggregator needed to do was aggregate the supply of clinicians. They did this by making it easier for clinicians to get in-network with payers and building software to solve some of their problems, like administrative work. Importantly, they also helped them find clients. Once they started to aggregate supply, they were able to sell access to these networks to payers, employers and ultimately, to clients.

But for these companies to achieve their ambitions — faster growth, more impact and exits for their investors — aggregating supply won’t be enough. Without truly owning the demand, these businesses have limited negotiating power with their supply. In mental health, clinicians know they can move to another marketplace or just go out on their own. As a result, businesses must constantly compete to acquire and retain their network of clinicians, which impacts margins.

Some businesses may think they have aggregated demand, but they haven’t. They’ve just intercepted it.

This becomes obvious once you look at how these businesses acquire clients. They run ads on Google or Facebook. They get therapists to include links to their platforms on their Psychology Today profiles. They have referral pathways with the health plans, employers or health systems they work with. But none of these tactics organically aggregates demand. The demand exists on someone else’s platform, and they intercept it. You can build a good business intercepting demand like this.² But without organically aggregating demand, it is very difficult to develop a dominant position with better pricing power and better unit economics.

This is different for businesses which can aggregate demand organically. In this scenario, users come to the marketplace first. They do this because they find so much value in the product itself that it would be silly for them to go anywhere else. Because they get so much value, they stick around and do things that help drive organic growth — like leaving reviews and telling their friends.

The important word here is organically.

To get users to come to a marketplace of their own volition, businesses must create genuine surplus value for the user. This is the nut mental health aggregators are yet to crack.

Bill Gurley describes how the best marketplaces create surplus value for the user by creating experiences that were not previously possible: “Great marketplaces do not simply aggregate a market; they enhance it. They leverage the connective tissue to offer the consumer a user experience that simply was not possible before the arrival of this new intermediary.”

Mental health aggregators have made it easier for users to find a therapist or a psychiatrist. They’ve also made it cheaper for most users to get care by allowing them to access in-network clinicians.

But is the user experience of getting care through an aggregator really that much better than what it used to be? Have they created a user experience that wasn’t possible before? Are they solving the most important problems for users?

The real question is, how much more value does a user get from going through an aggregator versus the alternative of the non-aggregated experience (going directly to a therapist)?

When the delta between the aggregated and non-aggregated user experience is small, aggregators will be unable to organically own demand at scale.

And when you don’t own the demand, you will always be competing for supply. So what can these businesses do to create more surplus value for users?

Add more supply?
In other industries, you can create value for users by bringing more supply to the network. The version of Amazon where I can buy pretty much anything is much more valuable to me than an Amazon where I can only buy books. But in mental health, this isn’t really true. A client just needs to find one therapist to work with. Providing some choice and ensuring a user can find a therapist covered by their insurance does add value. But aggregators can achieve this with relatively few therapists in their network. The minimum threshold of supply for an aggregator to be valuable is quite low. After that threshold is met, there’s limited marginal value to a user in having more therapists in the network. This is one of the reasons we continue to see new aggregators pop up, which are capable of competing with much larger incumbents.

Make the supply cheaper?
Providing economic value to users is another approach to improving their experience. Spotify made it much cheaper for me to listen to music. Mental health aggregators have done this well. Their strategy to get clinicians in-network with insurers has made care available at a very low or no cost to users.³ The challenge is that this has now become table stakes for the market, and there’s nowhere left to go. The only way to add incremental economic value to users would be to pay them to go to therapy.

The thing users actually care about… to get better
What these businesses must do is find the most important problems users have when it comes to their mental health journey and solve them. More than anything, people want to get better. They want high-quality care. They want an organisation to hold their hand through the messiness of their journey. They don’t want to have to question all of these things; they just want to trust that they are getting the correct, high-quality care and that this will make them better.

If an aggregator could deliver on the promise of making people better, imagine how much demand would flood to them. And it would all be organic. This organic aggregation of demand would then propel the aggregator into the dominant market position we have discussed throughout this article.

When I look at the aggregators today, many are still focused on the supply side. They are implementing technology to reduce admin burden, to improve note-taking or to address other supply-related issues. But a greater focus on solving demand-side problems will provide much greater return. Investments in building consumer-facing AI products are a sign that some businesses are starting to think this way. But what else can they do to solve the core problems users care about? Those who take this job seriously will reap huge rewards.

I’m keen to hear your thoughts on the role of marketplaces in mental health and how they are developing, so feel free to comment below and share your reflections.⁴

Keep fighting the good fight!

Steve Duke

Founder of Hemingway

Notes:

[1] In a 2023 piece, investors from a16z argued that mental health marketplaces would benefit from network effects. Specifically, they said that as marketplaces onboard more clinicians, they will have greater negotiating power with insurers and can use that power to drive up rates. But this hasn’t happened. A lot of the negotiating power still lies with the insurer because (a) there is a lot of competition among mental health businesses and (b) the product is largely commoditised. In May 2026, Aetna cut rates for therapists on one of the biggest marketplaces in the US, Alma. Talkspace faces similar rate challenges. Based on my own analysis, the average revenue they receive per session (a very good proxy for rates) declined from 2024 to 2025, and their gross margin has also been declining as a result. They openly admit that this is a consequence of pursuing insurers as their growth strategy. Greater supply has not translated into better rates.

[2] Of course, not all businesses are aggregators. LifeStance, for example, is a very large provider of outpatient mental health, but they do not use technology to aggregate the supply or demand of their market. They take a pre-internet approach to building businesses — vertically integrating supply (clinicians) with distribution (their clinic footprint) to provide a service. In other industries, traditional businesses like LifeStance have suffered much more at the hands of technology-focused aggregators. Have you checked in on your local travel agent recently? Businesses need to make a decision whether they truly want to be an aggregator or not. If they do want to be an aggregator, they need to organically own demand, and that will require creating surplus value for their users. If not, they will need to be very wary of any businesses that do.

[3] It should be noted that while this is good for users, it is not actually making care less expensive; it is just reducing the out-of-pocket expense for clients. To reduce the cost of care, businesses would need to actually create additional supply. But they are not able to do that — they can’t increase the number of accredited clinicians. Of course, some are attempting to do this using technology.

[4] Are aggregators good for the mental health ecosystem? Proponents will say they increase the efficiency of the market, help improve access to care and make life easier for clinicians. Some clinicians are apprehensive about their role, however, citing concerns about their role as an intermediary in care journeys and exploitative tactics around rates. If these aggregators become rent-seeking, then it will likely be bad for clients, clinicians and the broader ecosystem. We would not want a situation where aggregators have some sort of regulatory capture or monopolistic situation over the supply of clinicians, which would allow them to extract rent. But this is currently not the case. There is a significant amount of competition, and clinicians can move freely between them or find clients through other channels. But if a business organically aggregates a majority share of the demand, could this change? Perhaps. The important thing in this scenario would be to ensure that such aggregator(s) do not take a rake too far and use their market power to reduce clinician rates or impose other practices that would be detrimental to clinicians or clients. It’s also worth noting that the whole point of this article is to say that this situation would only occur if an aggregator is creating genuine surplus value for clients. This would be a good thing — clients deserve a much better experience of care and, importantly, much better outcomes. We would just need to ensure the power that comes with this is not wielded against clinicians or others.

In 1939, Howard Florey was trying to figure out how to produce enough Penicillin to run a trial in humans. It was a scientific and manufacturing question. But Florey’s genius lay in realising he needed to innovate in how he was approaching finding an answer. He recruited an interdisciplinary team to his lab at Oxford and had them work simultaneously on their respective areas of expertise — something that wasn’t really done at the time. This process soon provided an answer to their Penicillin production problem allowing them to make enough of the substance to test it in a human.

A few decades later, NASA were trying to put a man on the moon. To do so, they had to invent modern project management just to coordinate the 400,000 people working on the problem.

In 2026, the field of mental health faces another hard question: Can we use Artificial Intelligence to improve people’s mental health?

There are many technical and clinical aspects to this question that we still don’t understand. Teams are building on a rapidly changing frontier technology, the evidence base is still emerging, the stakes are high and there are multiple success criteria to consider. Because of this, the old product playbook no longer works. Leaders realise they need to redefine how their teams tackle this question.

Over the last few weeks I’ve been speaking with product and clinical leaders from organisations like Spring Health, Headspace, Grow Therapy and Slingshot to understand more about how they approach this problem.

In this edition of The Hemingway Report, I share seven insights into how these leading organisations are redesigning their teams and processes to build mental health AI.

Let’s get into it.

Many thanks to the contributors to this article:

• Gijo Mathew, Chief Product Officer Spring Health

• Jenna Glover, Chief Clinical Officer @ Headspace

• Fay Kallel, Chief Product & Engineering Officer at Headspace

• Bre Vergess, VP, Head of Product @ Grow Therapy

• Kevin Ramotar, Director of Clinical Product & AI @ Grow Therapy

• Matt Scult, Principal, Clinical AI @ Grow Therapy

• Derrick Hull, Clinical R&D Lead @ Slingshot

• Dana Udall, Healthtech Executive, Psychologist and Former CCO @ Nourish and Headspace

Building Mental Health AI poses a new set of challenges for product teams. As Gijo Mathew, Chief Product Officer at Spring Health puts it:

“
The biggest learning is that building AI for mental health is not just a technical challenge. It’s a product, clinical, safety, and trust challenge all at once.

Gijo Mathew, Chief Product Officer at Spring Health

Leading organisations are therefore writing a new playbook for building mental health AI products. From organisational structure, to hiring, to governance and building trust, here are seven lessons for building a high-performing mental health AI team.

1. Build AI-native, cross-functional teams, with clinicians embedded from day one

Every leader emphasises the importance of including clinicians from day one.

In mental health AI, product decisions are clinical decisions. Tone, timing, response length, how validation is delivered are all aspects that shape a person’s psychological experience. A product team making these choices without clinical expertise in the room may build something that feels polished but has no therapeutic grounding. In the worst cases, it could even be harmful.

As Jenna Glover, Chief Clinical Officer at Headspace says: “Remember that every design choice is an intervention. There has to be intentionality about why each choice was made and whether it leads to the intended outcome.”

Glover continues: “It’s not enough to layer clinical expertise on after the tool is built. It has to be there from the start”.

Glover’s colleague Fay Kallel, who is the Fay Kallel, Chief Product & Engineering Officer at Headspace, emphasises the importance of embedding these clinicians into cross functional teams to building mental health AI teams and aligning these teams behind shared customer outcomes: “The strongest teams are organized around customer outcomes, not functions or feature areas. When product, engineering, design, data, content, and AI are aligned around a shared problem, decision-making is faster and accountability is clearer.”

Contrary to what some may believe, including clinicians early can actually speed up product development. Kallel argues that when organisations treat clinical review as a final approval step, it can create friction, lead to worse products, wasted work and slower product development. “When clinical expertise is present during problem definition, design, and experimentation, you catch the right issues earlier and build experiences that are safer by design, not just safer by review.”

“
Safer by design, not just safer by review.

Fay Kallel, Chief Product & Engineering Officer at Headspace

While clinician involvement and cross functional teams are consistent traits across all the organisations I spoke to, they have taken different approaches to implementing these principles in their orgs.

Spring built an AI-native team from the ground up. This included recruiting AI specific roles like AI Product Managers, Prompt Engineers, ML Engineers, ML Operations, AI Trust and Safety Engineers, and AI UX Researchers who evaluate whether designs are fit for human use.

Clinicians sit directly inside this team, integrated as collaborators from day one. As Mathew says: “We bring product thinkers, clinicians, engineers, researchers, and operators together around one question: how do we use AI to remove barriers to care without ever losing sight of the person who needs help?”

Grow Therapy took a different route. They created a dedicated Clinical Product function that sits within their broader product org. Bre Vergess, Grow’s Head of Product speaks about when she realised they needed to move beyond consultative involvement and to build out this Clinical Product team. “…as AI became a bigger part of what we do, the stakes for getting it right only went up. Having clinicians available to consult wasn’t enough, they needed to be full-time builders, sitting side by side with PMs, designers, and engineers, with all the context and the right level of influence to shape what we create.”

“
Having clinicians available to consult wasn’t enough, they needed to be full-time builders.

Bre Vergess, Head of Product at Grow

Grow’s Clinical Product team is staffed entirely with licensed clinicians and is responsible for ensuring that product development is grounded in clinical best practices, enhances patient safety, and supports evidence-based care at scale. The head of this team reports directly to the VP of Product and each Clinical Product team member serves as a designated point of contact for specific product work-streams across Grow.

Headspace include clinicians in their cross functional teams and embed them across their product org. Glover says that the team building their AI companion, Ebb, is roughly 50/50 clinical to product/engineering, with clinical expertise woven throughout the team.

2. Hire solutions-oriented clinicians with product sense — and product people with clinical curiosity

What types of people do leaders look for when building their mental health AI teams? Clinical expertise and technical skill matter, but they aren’t sufficient. The best teams hire people that combine domain knowledge with a unique set of traits that allow them to work effectively in this emerging, high-stakes field.

What to look for in clinicians.

Vergess looks for solution oriented clinicians. “The archetype we were looking for wasn’t just clinical expertise, it was clinical expertise combined with a solution orientation and real product sense…”

Vergess notes the importance of having genuine ‘product sense’ alongside clinical expertise: knowing how to identify priorities, communicate crisply across disciplines, and show up as a clinical voice that product and engineering actively seek out at key moments.

Dana Udall, a Healthtech Executive, Psychologist and Former CCO @ Nourish and Headspace, recommends screening clinical hires for risk tolerance, cross-functional willingness, and flexibility. She prioritises candidates who have a strong clinical point of view and who know how to have their concerns understood and accepted by the wider organisation. She also looks for people who are flexible and open to having their mind changed.

Grow deliberately recruit clinicians with different therapeutic orientations to make sure they have scientific breadth and avoid tunnel vision from a single clinical perspective.

What to look for across the broader product team.

Kallel reflects Udall’s desire for open-mindedness in her teams. “The strongest team members are curious, collaborative, data-informed, and comfortable operating in ambiguity.”

“
The strongest team members are curious, collaborative, data-informed, and comfortable operating in ambiguity.

Fay Kallel, Chief Product & Engineering Officer at Headspace

AI fluency is also important for Kallel — not just for engineers, but for clinicians, product managers, and designers working on these products. The field is moving so fast that people who can’t engage with the technical realities of how AI systems work will struggle to contribute meaningfully to product decisions. As Kallel describes: “AI fluency, adaptability, and a continuous learning mindset are becoming foundational traits across every discipline.”

Deep technical expertise matters too.

While all leaders recognise the importance of these softer character traits, Derrick Hull, Clinical R&D Lead at Slingshot reminds us of the importance of hard skills in this domain. Building purpose-built mental health AI requires a specific set of technical skills, especially in machine learning. Hull’s belief is that teams that underinvest in technical expertise and rely on prompting alone will hit a ceiling in what their products can do.

Becoming a Hemingway Pro Member

If you’re enjoying this article it’s probably because you want tactical information on building a successful mental health organisation. If you’d like more of that, and to meet peers also building in mental health, consider becoming a Hemingway Pro Member.

As a member, you’ll join over 450 other mental health leaders and get access to exclusive content, events and resources.

Join Hemingway Pro

3. Empower clinicians to build, not just review

Embedding clinicians is important. But top teams have moved past using clinicians solely for product reviews or requirement generation. Their clinicians do actual product work.

At Grow, the Clinical Product team is hands on. The team writes prompts, runs evaluations, creates prototypes, designs QA architecture, and analyses data. They operate much closer to engineering than a traditional subject matter expert function, bringing clinical expertise directly into the technical work rather than handing off recommendations for others to execute.

While Grow’s Clinical Product team is somewhat unique in how hands-on they are, Kallel thinks similarly about empowering clinicians at Headspace: “Clinical partners aren’t reviewers of finished work, and product teams aren’t simply executors of clinical requirements. Both are involved early in defining problems, evaluating opportunities, and shaping solutions.”

“
Clinical partners aren’t reviewers of finished work, and product teams aren’t simply executors of clinical requirements.

Fay Kallel, Chief Product & Engineering Officer at Headspace

Spring take a similar approach with clinicians shaping evaluations, safety prompts, and feature prompts directly alongside product and engineering partners.

4. Build trust through shared work, not review gates

Building innovative products requires a high degree of trust within teams and across the organisation. This trust is best built when clinical and product teams actually work together on the same problems.

Mathew describes how this happens at Spring: “Trust is built through shared work — reviewing real member conversations together, writing prompts and evaluation criteria jointly, being in the same room for the difficult discussions. When both sides are looking at the same outputs and accountable to the same quality bar, the division stops being meaningful.”

“
Trust is built through shared work.

Gijo Mathew, Chief Product Officer at Spring Health

Kallel at Headspace takes a similar approach. “One of the biggest lessons I’ve learned is that trust between clinical and product teams doesn’t only come from governance structures, it comes from how people work together every day. We’ve been very intentional about establishing workflows where neither group is brought in at the end of the process.”

Grow describe building a culture where Clinical Product isn’t a separate conscience sitting outside the rest of the organisation. They’ve found that the best work happens when clinicians feel empowered to flag a genuine clinical concern and equally empowered to find a creative solution — rather than defaulting to a hard no. Trust at Grow builds in both directions: when Clinical Product flags something early and it turns out to matter, that earns confidence with the broader product team. When Clinical Product enables other teams to meet their timelines while maintaining conviction on safety and quality, product managers gain confidence in what they’re shipping (and in the value of the Clinical Product team).

Aligning around a shared metric also helps improve collaboration and trust. At Headspace, that metric is member trust. As Kallel explains: “When trust becomes a shared goal, many of the hardest tradeoffs become clearer. The conversation shifts from ‘Can we ship this?’ to ‘Will this strengthen or weaken the trust our members place in us?’”

Shadowing and rituals can also be helpful. Udall warns that when escalations are the primary mode of contact between clinical and product, the relationship can turn adversarial. “If that is the bulk of the interaction, you are going to obliterate trust.” Her advice is to shadow each other, build shared failure reviews, and approach the other team with curiosity.

As an example, Headspace run a “Trust Tuesday” session where product, clinical, operations, and legal discuss priorities and emerging issues. Spaces like this where teams can openly share what’s on their mind and also be transparent about their concerns and mistakes are critical to building trust.

5. Create clinician buy-in by leading with clinical reality

Getting clinician buy-in is critical to product adoption. Clinicians outside the core AI team are evaluating your product through the lens of patient safety, clinical quality, and whether it makes their work easier or harder. The teams that get buy-in from their broader clinical network realise this and address these topics directly in their communications. They don’t lead with AI hype.

Spring Health found that conversations with clinicians work better when they start with the clinical problem, the workflow reality, and the guardrails — not with just pure excitement about the technology. They recommend approaching conversations by sharing the clinical problem the product is meant to address and exactly how the product is meant to do that. Spring also found that transparency about limitations builds more credibility than polish. Being direct about uncertainty and failure modes earns more trust than presenting the technology as fully solved. Hype-driven messaging and framing AI adoption as inevitable causes clinicians to disengage quickly.

Grow emphasise the importance of meeting providers where they are on their own journey with AI. Some providers are early adopters, others are cautious, some are vehemently opposed to AI. Grow compares it to telehealth adoption, which was slow before the pandemic. Some providers have now fully transitioned while others still prefer in-person work. Grow lead with clinical credibility, acknowledge concerns openly, and respect different practice philosophies. What doesn’t work is dismissing concerns, over-explaining, or moving too fast.

Udall recommends making the co-creation process visible. Involve clinicians in development from the start, then have those clinicians present to the wider organisation about what they contributed and what changed as a result.

6. Formalise your safety governance before you need it

Every team we spoke to has built formal governance structures for safety and quality decisions. The structures differ but the underlying principle is the same: don’t leave safety to informal judgment, and agree on how decisions get made before a crisis forces you to improvise.

Headspace created a Clinical Safety and Experience Council (CSEC), a governance body that brings together clinical, safety, research, and product perspectives to make decisions about Ebb’s safety and quality. As Glover describes: “When we started Ebb, we had different groups working together, but there’s so much happening in the space that we decided we needed to bring a council together that could put all these inputs together and make decisions. That became our Clinical Safety and Experience Council.” The council discusses both micro-level aspects of Ebb’s behaviour (how the AI responds in specific conversation types) and macro-level changes happening in the field (new research, regulatory developments, strategic questions, etc.). Glover gives an example of a recent CSEC discussion: “A topic that came up recently was global expansion. When is it okay for us to release Ebb in different countries? How does safety vary by culture? And how does motivational interviewing, in terms of how Ebb is currently trained, fall flat within different cultural contexts?”

“
There’s so much happening in the space that we decided we needed to bring a council together that could put all these inputs together and make decisions.

Jenna Glover, Chief Clinical Officer at Headspace

Spring Health created an AI Governance Board that brings together security, clinical, legal, AI safety, and compliance perspectives. It sits independent of the product org. This body defines what “safe to deploy” means, and that definition cascades into everything across the org. Mathew explains why their AI Governance Board is independent of the product org: “Feature teams move fast — that’s their job. Safety sets constraints they operate within. When safety is owned by the feature team, it tends to bend to shipping pressure. When it’s independent, it doesn’t.”

“
When safety is owned by the feature team, it tends to bend to shipping pressure. When it’s independent, it doesn’t.

Gijo Mathew, Chief Product Officer at Spring Health

At Grow, readiness to ship is determined by predefined success criteria, including benchmarking tests, internal review, and red-teaming by an outside organisation. Decisions are made jointly between product management and clinical product leadership.

Udall warns against any governance structure that concentrates clinical sign-off in one person — what she calls the sole expert trap. “It puts the onus totally on that clinician, and if they’re not in the room the day the decision is made, they are often left out of the discussion. We really need to have this kind of distributed critical layer of thinking that lives across the organization.” This is particularly relevant for smaller businesses who may not have large clinical teams. Her recommendation is to develop a culture where everyone has the duty to raise safety concerns.v

“
We really need to have this kind of distributed critical layer of thinking that lives across the organization.

Dana Udall

Governance doesn’t have to mean added bureaucracy and slowness. As Kallel puts it: “Don’t confuse governance with bureaucracy. If you’ve designed the system well, governance should accelerate good decisions, not slow them down.” She also argues that safety decisions are rarely binary: “It’s rarely a simple ‘ship or don’t ship’ conversation. More often it’s about determining the appropriate level of intervention, oversight, transparency, and monitoring for a given experience.”

“
It’s rarely a simple ‘ship or don’t ship’ conversation.

Fay Kallel, Chief Product & Engineering Officer at Headspace

7. Build processes to move fast without sacrificing clinical rigour or safety

Good governance can speed up good decision making. So can good process.

The teams I spoke to all run tight, recurring processes for reviewing AI quality, managing safety, and iterating on the product. Predictable, recurring systems are how these teams continuously improve quality over time.

In terms of ongoing quality processes, Grow runs weekly clinical reviews of AI outputs, with a triage hierarchy that prioritises the most important conversations. Between review cycles, they run automated evaluation monitoring which maintains coverage at scale. The team also conducts table reads where they walk through real AI interactions. This helps to surface issues that they might miss otherwise. Weekly team meetings focus on quality, safety, and prompt improvement, with QA patterns feeding directly into product iteration. An example of this is when their QA team identified a pattern where panic symptoms were being routed through medical emergency guardrails, triggering unnecessary escalations. With that finding, they could change their prompting which then improved performance.

Headspace run a structured operating rhythm too. As Kallel describes: “On a weekly basis, we review customer experience, experiment performance, product quality, technical architecture, and emerging risks. On a monthly basis, we review progress against strategic priorities. On a quarterly basis, we go deeper, assessing outcomes, reallocating investments, and refining strategy and execution. What makes this work is predictability. What tends to break things is ad hoc communication, unclear ownership, and relying on heroics rather than systems.”

“
What tends to break things is ad hoc communication, unclear ownership, and relying on heroics rather than systems.

Fay Kallel, Chief Product & Engineering Officer at Headspace

One important aspect of process and team design is decision rights. Udall recommends defining decision rights in advance. Spring Health does this through a set of quantitative exit criteria, with every feature evaluated across multiple safety dimensions before it can move forward.

Staying current with new research, regulation, and developments in the field is also a team responsibility. At Headspace “staying current on research, regulation, and industry developments is a shared responsibility” as Kallel describes. Each individual is responsible for staying up to date with developments in their own field. Grow takes a similar approach. They also highlight the benefit of actively contributing to the field. They see forming external research partnerships, standing up an external AI advisory board, and collaborating with academic institutions as a good way to improve their ability to remain at the forefront of technology and research.

That’s all for this edition of The Hemingway Report. Many thanks again to all of the contributors to this article.

If you found this valuable, consider supporting our work by becoming a Hemingway Pro Member. As a member you will get access to exclusive content and be invited to join our community of mental health innovators.

Keep fighting the good fight!

Steve

Founder of Hemingway

The chatbot, called Tee, is a direct to consumer, conversational AI product which users can chat with using text or voice.

Marketed as “A private place to talk about anything, at any time” Talkspace positions Tee as a superior alternative to using general-purpose AI products, emphasising its safety, clinical oversight, memory and the fact it has been trained specifically on mental health data.

While there is clear consumer demand to use conversational AI for emotional and mental health needs, it’s less clear if people want — and are willing to pay for — purpose built mental health AI. Many mental health players have launched conversational AI products over the last two years, but the market remains wide open, with no purpose built product yet gaining dominance.

Tee’s position as part of the care experience is different to some of these other mental health products. Tee users do not have to be engaged with care to use Tee and if they are working with a Talkspace therapist, their conversations with Tee will not be shared with them. Other businesses (like Grow Therapy, or Spring Health) have deployed conversational AI more as part of a hybrid care experience, focusing on support in between sessions with notes and insights shared with therapists. Headspace’s approach sits between these two, with their AI companion Ebb available for consumers to use independently of their therapy offering, but with the option for insights to be shared with a user’s clinical team.

Tee is available on a seven day free trial, after which users will be charged $19.99 per month. While this may offer a new revenue line for Talkspace, leadership will also be hoping it drives more clients to their core therapy business. Talkspace have over two hundred million covered lives in the US but only a tiny fraction of those people actually sign up for therapy with a Talkspace therapist. Acquiring clients, even for a service covered by insurance, has proved challenging. The market for insurance-covered therapy is competitive and most players have struggled to differentiate. Tee may function as an acquisition channel for Talkspace by acting as an on-ramp for therapy. Users who are identified as potentially benefitting from therapy with Tee will be directed to sign up for Talkspace Therapy.

For the last three years, Talkspace have been moving away from the consumer market, focusing on building payer revenue. This has worked well for the mental health provider. Their revenue has grown at a CAGR of 24% since 2022 and they delivered $8M in Net Income in 2025 — a huge turnaround from their losses of almost $80M in 2022. Earlier this year Talkspace was acquired by United Health Services for $835M.

Launching a consumer AI product is a bold move for Talkspace. So far, the commercial impact of AI in mental health has primarily been to improve operational efficiency. Whether it can lower care costs, expand access to new markets, improve outcomes, and ultimately drive new revenue, remains to be seen. I will be closely following Tee and the rest of the mental health AI market to see how this develops.

Keep fighting the good fight!

Steve Duke

Founder of Hemingway

P.S. If you find Hemignway content valuable, consider becoming a Hemingway Pro Member. As a Pro Member, you will get access to even deeper industry insights and will be invited to join our community of over 450 mental health leaders.

I share the latest industry news (from deal-making to product launches to regulatory changes) and summaries of the latest research. It’s the easiest way to stay up to date with everything that’s happening in this fast-moving space. Subscribe to get these in your inbox every week.

Subscribe now

What’s been happening in mental health this week?

In the US, Federal and State bodies are tightening oversight of the US behavioural health system. This comes as a response to concerns over fraud and the overall rising costs of behavioural health programs.

Elsewhere, Limbic launched an AI Scribe, DiMe announced a set of validated digital measures for mental disorders and Open AI announced more funding for research on the impact of AI.

In research, evidence suggests CBT is no more effective than active controls in the treatment of depression in primary care and remote work was found to worsen mental health.

There’s lots more happening too, so let’s get into it.

What happened in mental health last week?

Business News:

⏺️ Radley Health raised $3M. The Cincinnati-based company operates 450+ certified peer support specialists across 70+ Ohio counties. (link)

⏺️ HPS Investment Partners took a majority stake in Discovery Behavioral Health. HPS took control of the behavioural health provider after the company defaulted on debt obligations. (link)

⏺️ CMS finalises new Medicaid work requirements. Certain adults 19-64 must complete 80 hours per month of qualifying activity to maintain Medicaid eligibility from January 2027. (link)

⏺️ Two states tightened behavioural health fraud enforcement. Massachusetts sued UnitedHealthcare for $100M, alleging members were misclassified as having behavioural health or substance use disorders without valid diagnoses to receive higher Medicaid capitation payments. Ohio launched enhanced licensure reviews and expanded anti-fraud coordination with Medicaid and the attorney general, triggered in part by a separate $30M federal fraud case filed the same week. (link) (link)

⏺️ Oregon replaced 30-year county behavioural health model with updated statewide contracts. The new County Financial Assistance Agreements replace a 30-year-old contractual model and give the state a real-time view of how counties deploy state-funded behavioural health resources. (link)

⏺️ Limbic launched Limbic Scribe. The new AI Scribe product embeds TORTUS’s Medical Device-certified ambient voice technology into Limbic Care, automatically generating clinical letters and notes using standardised templates. (link)

⏺️ UK Parliament opens evidence session on children and young people’s mental health. The UK Education and Health and Social Care committees launched a joint survey to hear from young people who have received mental health services. They will also host their first evidence session of their children and young people’s mental health inquiry this Wednesday, June 10th, . (link)

⏺️ One Mind at Work published a report finding that only 1 in 4 employers has a formal workforce mental health strategy. One Mind at Work’s 2026 Mental Health at Work Index Annual Report analysed data from 149 organisations representing 2.75 million workers across 25 industries and 23 countries. While only 1 in 4 employers have a formal workforce mental health strategy, those that do see 36% lower voluntary turnover. The report includes other interesting insights and is worth a read for anyone interested in workforce mental health. (link)

⏺️ Orygen announced $5.7M in grant funding from Wellcome for their AI-powered youth mental health project. The grant will fund AI4You(th), a two-year project developing SensAI, an AI system designed to support both young people and clinicians with personalised care between clinical contacts. (link)

⏺️ OpenAI appears to have notified successful applicants to for its research funding program on AI and mental health. The successful applicants include researchers from York University with a project entitled “Preventing Cognitive Atrophy in LLM Mental Health Support Through Psychometric Evaluation and Therapist-Aligned Design”. I have been unable to find a complete list of funded projects. (link).

⏺️ The same week, the OpenAI Foundation announced $130M+ in grants under their AI Resilience program. The program targets bio-resilience, cyber-resilience, AI model safety, and AI’s impact on young people. The Foundation expects to invest $25B across AI Resilience and Life Sciences in the years ahead. (link)

⏺️ DiMe publishes first digital measures set built specifically for common mental health disorders. DATAcc by the Digital Medicine Society released seven sensor-based measures, covering sleep, physical activity, daily routine variability, cognitive function, speech and language, autonomic function, and facial and gaze features, that meet the evidence bar for use in clinical trials, phenotyping, and value-based care. The set also identifies concepts of interest for which no validated measure yet exists. These measures were developed with 30+ partners including the FDA, BetterHelp, Talkspace, and the APA. It was commissioned by Wellcome. (link)

If you find these updates valuable, consider becoming a Hemingway Pro Member. Being a Pro Member is the easiest way to stay up to date on everything happening in the mental health landscape. You will also be invited to join our vetted Slack group for mental health innovators and receive invites to our private events.
Check it out here.

Research News:

⏺️ CBT found to be no more effective than other active treatments for depression in primary care. Across 44 RCTs with 10,000+ participants, CBT outperformed inactive controls (g = 0.44) but showed no clear advantage over antidepressants, exercise, or other talking therapies. Self-help CBT showed no significant benefit over inactive controls. (link)

⏺️ Peer-led intervention reduced depression and anxiety among transgender adults in Delhi. The Manthan study enrolled 62 participants in a 12-month peer-support program. Depression scores fell from 13.1 to 7.0 and anxiety scores from 11.2 to 6.0. Over 60% achieved clinically significant improvement in depression and 59% in anxiety. Attendance exceeded 91%. There was no control condition and the authors have called for multi-site RCTs. (link)

⏺️ A four-session perinatal intervention reduced depression in high-risk mothers and partners. The Chinese RCT of 100 dyads found significantly lower maternal depressive symptoms at 2-3 days and 6 weeks postpartum. Paternal benefits were less consistent over time. (link)

⏺️ Slingshot published research on how pregnant and postpartum people use an AI mental health tool. The primary finding was that perinatal distress rose across the postpartum year among participants. Across 419 users, identity strain and emotional exhaustion increased through 12 months postpartum while positive affect declined. Among users entering with moderate-to-severe depression (PHQ-9 ≥10), model-estimated scores declined by 2.41 points over 8 weeks (d = 0.58). FYI, there was no control condition, PHQ-9 scores are model-estimated and this paper has not yet been peer reviewed. (link)

⏺️ Torous and colleagues call for new consent frameworks for training AI on therapy data. The commentary argues that “existing consent models do not adequately protect autonomy and confidentiality” and calls for explicit opt-in consent for AI training and patient-led governance of how clinical data is used. (link)

⏺️ Remote work found to increase isolation and mental distress, particularly among those living alone. Five nationally representative US surveys covering 588,322 workers found that post-pandemic, remote-capable workers spent more time alone and reduced social activity outside work. Remote workers living alone showed the sharpest increases in mental distress, mental healthcare use, and antidepressant use. The study notes that the findings suggest workers “may not realize the costs of remote work for their well-being, which may take time to accumulate.” (link)

⏺️ More outdoor play in preschool years predicts better mental health trajectories into middle childhood. A longitudinal study of 4,151 Scottish children found each additional day per week of outdoor play at ages 2-4 reduced the likelihood of developing anxiety, depression, and behavioural problems by ages 4-8. Children who played outside more often were significantly less likely to show worsening emotional or behavioural symptoms over time. (link)

That’s all for this week.

Keep fighting the good fight!

Steve Duke

Founder of Hemingway

P.S. If you find these updates valuable, consider becoming a Hemingway Pro Member. As a Pro Member, you will get access to even deeper industry insights and will be invited to join our community of over 450 mental health leaders.

In the last few weeks, five public mental health businesses released their Q1 results. These include important financial and operational data as well as commentary from the company’s management on what they’re seeing in the market. I’ve been pouring over this data and the call transcripts and in today’s Hemingway Pro article, I’ll share everything I’ve learned.

We’ll dive into data from Talkspace, Lifestance, Acadia, Teladoc (BetterHelp) and (for the first time) Brainsway and look at some of the major trends shaping the industry.

Let’s get into it.

The Key Takeaways:

1. Clinician productivity is key in a low growth market
   Overall, growth in the market remains modest. Among services business, LifeStance grew the fastest (21% YoY) followed by Talkspace (18% YoY) and Acadia (8%). Productivity has been a primary driver of growth for these businesses. LifeStance’s visits per average clinician were up 7% year-over-year — the second consecutive quarter of 7% productivity improvement. On the earnings call, management attributed this to “enhancements to new patient conversion and engagement,” supported by clinician-level visibility, education and incentives. Talkspace’s provider network scaled to support 31% growth in payer sessions. Across the board, growth is modest and the market is rewarding operational efficiency. Incremental growth delivered from increased clinician productivity is therefore highly valuable.

2. Rates see low single-digit growth (at best)
   LifeStance’s revenue per visit was $163, up 3% year-over-year. Their full-year guidance assumes “low-to-mid single digit increases” to total revenue per visit. Acadia’s same-facility revenue per patient day rose 5.6% to $1,052, but only ~2% of that is due to underlying rate growth. Talkspace rates appear to have declined by ~2%. While they don’t disclose revenue per session directly, we can back calculate it based on disclosed payer revenue ($48.6M) and payer session volume (459,300), the implied rate is approximately $106 per session — down roughly 2% year-over-year from an implied $108 in Q1 2025. If accurate, this suggests Talkspace’s 28% payer revenue growth in Q1 was entirely volume-driven rather than rate-driven. Across the set, any significant rate growth remains elusive. There was also extremely limited mention of any value based payments across the cohort.

3. More outcomes data is being published to support new market positioning
   LifeStance published an outcomes study covering approximately 180,000 patients. This was referenced in both the earnings release and the CEO’s prepared remarks, stating that three-quarters showed clinically significant improvement in anxiety and depression. While not included in this report (because they don’t publish quarterly earnings) Kooth recently published a study on outcomes for Soluna in California. Talkspace have also been publishing more outcomes focused reports. This is a reaction to a market that is demanding more evidence of outcomes and quality care. LifeStance CEO Dave Bourdon stated that they “continue to see a growing demand for high-quality mental healthcare”. Proving outcomes and high quality care is a shift I welcome.

4. AI is being deployed operationally, but not yet clinically

   All companies focused primarily on operational use cases of AI rather than patient-facing clinical delivery. LifeStance listed four specific deployments: digital patient check-in, AI-driven revenue cycle automation, AI-enabled scheduling (improving call-to-appointment conversion), and AI-assisted clinical documentation. In 2025, Talkspace disclosed “TalkAI” — described at the time as a placeholder name for a conversational AI agent. In Q1 2026, the company formalised this initiative under a new name — “Sentia” — which is housed in a dedicated wholly-owned subsidiary (Sentia AI, LLC). Sentia is a behavioural-health-focused large language model through which members interact with an AI agent about their mental wellbeing. But it’s still in beta and management don’t expect it to generate any revenue in 2026. In fact, no company discusses AI as a revenue driver, but only as a lever for operational efficiency and for productivity improvements that can drive marginal growth in existing business lines.

The rest of this article is for Hemingway Pro members only. Become a Hemingway Pro member to get access to this report and more in-depth analysis on mental health innovation and the industry that drives it. As a member, you’ll also join our vetted community of over 450 mental health leaders.

Join Hemingway Pro

Once you’ve done so, you’ll be able to read the full report here.

Many solutions are commoditised. Network effects, scale economies and switching cost are rare.

One of the few ways you can develop defensibility is through scientific credibility. The businesses with credibility close deals faster and raise more capital. And as they do this, they gain the resources and reach that allow them to develop more evidence and more credibility. Their defensibility compounds.

This asset of credibility is built by generating evidence that holds up under scrutiny.

When you can make defensible assertions about what your product does, you will win over stakeholders and close deals. If you can then build a collection of defensible evidence, your scientific credibility is elevated to a strategic asset. You become known as a business that can be trusted to deliver on what you put forward. And that brings success, especially in a market that is increasingly scrutinising solutions.

But when your evidence isn’t defensible, stakeholders poke holes and conversations quickly grind to a halt. It’s a horrible feeling to be sitting in a conference room and watch a deal die as you’re asked to defend a position you knew was shaky. When you find yourself in this situation, it’s already too late.

In this Hemingway Guide, Jen Huberty PhD joins us to provide a framework for building scientific credibility as a mental health business. Jen has spent over twenty years in both academic research and commercial digital health. She was Head of Science at Calm and has advised several of the world’s leading digital health businesses.

Use this framework to strengthen your evidence, close more deals and develop scientific credibility as a strategic asset.

Key Takeaways

1. Scientific credibility is a company’s ability to defend its evidence with structured, stage-appropriate support that withstands stakeholder scrutiny.

2. It primarily operates at an assertion level — your ability to defend specific assertions.

3. Done repeatedly and with discipline, it can become a strategic asset

4. Scientific credibility requires four interdependent foundations: Problem, Population, Solution, and Outcomes. Weakness in any area undermines the entire structure.

5. You can use these four foundations to conduct a gap analysis before stakeholders expose weaknesses. Investors, enterprise buyers, and partners will find the holes eventually.

6. You should prioritise the assertions you want to make defensible by understanding which are critical to your near-term business goals.

7. Strong outcome data alone leaves critical questions unanswered: “Works for whom?” “Why does this problem matter?” “Is your solution what’s creating these results?”

8. Different stakeholders probe different foundations; balanced evidence matters more than deep evidence in just one area.

9. Evidence expectations evolve with stage of your business. What’s appropriate at pre-seed differs significantly from what’s expected at Series A or enterprise scale. The goal is not to reach RCT-level rigor, it’s to have the right evidence for the decisions your stakeholders are making right now.

10. Founders who build complete evidence before competitors gain a structural advantage.

Building Scientific Credibility

Scientific credibility is a company’s ability to defend its evidence with structured, stage-appropriate support that withstands stakeholder scrutiny. It is not defined by publications or methodological complexity; it is defined by whether your evidence holds up when the people making buying, funding, or partnership decisions push back on it.

It operates at two levels:

The assertion level. Everything a company puts forward about the problem it solves, the population it serves, how its solution works, and what outcomes it produces should be defensible across the four foundations.

The company level. Credibility at the company level is the cumulative result of scientific discipline applied consistently over time.

In this Guide we provide a framework for developing credibility in your assertions. You can use it to stress-test specific positions (before stakeholders do). Do this consistently and with discipline and the company-level asset will follow.

The Four Foundations of Scientific Credibility

Scientific credibility is built on four distinct foundations. They are defensible independently and when combined, reinforce one another.

• Problem: The problem you’re tackling is real, defined, and quantified for a specific group

• Population: The evidence reflects the actual people you’re trying to support

• Solution: Your product specifically drives the change, not something else

• Outcomes: The improvement is measured, meaningful, durable, and aligned with what your buyer cares about

Figure 1: The Four Foundations of Scientific Credibility

When you have all four foundations for an assertion, you can say; this is a real, quantifiable problem with direct health and economic impacts for this specific population and our solution solves this problem through this precise mechanism, resulting in measurable improvements in the metrics you care about.

Example: “Untreated anxiety costs U.S. employers $4,400 per affected employee annually in lost productivity, disproportionately impacting hourly frontline workers with limited access to traditional therapy. Our platform delivers asynchronous CBT through a guided self-pacing model that produces a 40% reduction in PHQ-9 scores within eight weeks, sustained at six months”.

Each foundation reinforces and strengthens the others, creating a more defensible and resilient evidence base. Together, they can support your business decisions and build your scientific credibility.

Building an evidence base can be expensive and takes time. Therefore, the best organisations know how to do two things well when it comes to scientific credibility;

1. They understand which assertions they need to make to support their business goals.

2. They know how to make those assertions highly defensible with strong evidence generation across all four foundations.

Prioritise the Assertions that Support your Business Goals

Figure 1 above showed how the four foundations apply to any single assertion. But most companies are defending multiple assertions simultaneously, each with different levels of evidence. Figure 2 (below) shows what that looks like in practice.

In reality, a company may have many more assertions than this, for different products and different stakeholders.

Thankfully, not all of these need to carry equal weight at any given moment. You need to understand which ones are critical to your near-term business goals and make sure those are structurally sound.

To do this, work backwards from your business strategy. Ask yourself, what business goals do I need to hit in the next twelve months and what will I need to defend to achieve them? Those are your priorities and should be the focus of your evidence generation strategy.

If your goal is to land five new enterprise contracts with employers, your assertion about reducing absenteeism is the one under pressure. If you only have problem and population evidence but no solution attribution or outcome data, that’s your vulnerability. It doesn’t matter that your investor-facing evidence is complete. The assertion that’s being tested is the one that needs to hold.

Build in the Right Order

It’s unrealistic to build all four foundations simultaneously across every assertion or business decision. And you don’t need to. The ideal sequence depends on your stage and the conversations that are most critical right now.

Early stage (pre-seed / seed). You might have evidence across one or two foundations for a single assertion. That’s OK. No one expects a complete evidence base at this stage. Focus on having strong evidence for a specific problem and population. You need to prove the problem matters and understand who experiences it before you invest heavily in solution development.

Series A. You need at least one assertion that meets all four foundations, and credible foundations progress on other key assertions Solution evidence should be emerging from product development and feasibility testing. If you’re planning a raise in twelve months, outcome data collection should start now.

Growth stage and enterprise scale. You need evidence across multiple assertions because you’re selling to multiple stakeholder types simultaneously. The employer buyer, the health plan, the health system — they each care about different claims, and each of those claims needs to hold up independently.

When we worked with a national virtual mental health provider, they had strong problem validation, clear population definition, and a well-articulated solution, but needed rapid outcome evidence to support their growth stage. We analyzed 5+ million therapy visits and published four papers in three months showing clinically meaningful symptom reduction. That completed their evidence at exactly the right time for their business needs.

Using the Four Foundations to Build Defensible Assertions

Now that you know which foundations to prioritise and when, the rest of this guide is a detailed framework for assessing each one. Use it to assess the strength of your evidence within your most critical business conversations, and to identify where your evidence needs to be stronger.

Foundation 1: The Problem

Do you have evidence that the problem you are solving actually exists and actually matters?

Many founders start a business to solve a personal pain point. But personal experience or anecdotal evidence are not the same as defensible problem evidence. You need to be able to clearly define the problem, its prevalence, the impact of that problem in health and economic terms, and why existing solutions don’t solve it.

What strong problem evidence looks like

Strong problem evidence rests on a small set of defensible components:

Prevalence data showing how many people experience the problem. Strong sources include peer-reviewed epidemiological studies, CDC and WHO surveillance data, large national surveys (e.g., BRFSS, NHANES), and some condition-specific registries. Avoid white papers or market research reports that cite prevalence without primary sources.

Impact data describing consequences in health, economic burden, or quality of life. Look to health economics literature, employer claims databases, and government cost reports (e.g., AHRQ, CMS). Self-reported impact data from your own users can supplement these sources but shouldn’t be used as your primary source.

Gap analysis clarifying what current solutions exist and why they fall short. Strong support comes from systematic reviews and meta-analyses identifying intervention gaps, published critiques of current standards of care, clinical practice guidelines that acknowledge unmet needs, and competitive landscape analyses grounded in peer-reviewed efficacy data. High-level competitive summaries or marketing claims about differentiation are not sufficient on their own.

Stakeholder perspective confirming that users define the problem the way you do. Examples include structured needs assessments, qualitative research with target users and buyers, and published patient/member experience data. Anecdotal feedback from a handful of users is not stakeholder validation.

Prevalence and impact are only persuasive when they are clearly defined and specific. A problem may be widespread, but severity, contributing factors, and consequences often vary meaningfully across subgroups. If the problem is not clearly defined for a specific population, evidence becomes easy to challenge.

For example, a digital mental health company with strong UK adoption entered the U.S. market expecting their existing evidence to transfer. It didn’t. Prevalence patterns, access barriers, and stakeholder expectations differed enough that general UK data wasn’t sufficient to support adoption in this new market. They needed to develop U.S.-specific problem evidence, establish a U.S. advisory board, and publish research in the U.S. before the problem could be defended credibly with American buyers.

You must be able to clearly demonstrate that the problem is real, meaningful, and unresolved. Before including a new source in any material, the strongest teams pressure-test each one by asking: does this materially strengthen our ability to defend the urgency and specificity of this problem in a real business conversation?

When problem evidence is weak

There are several red flags that signal weakness in your problem foundation. One is when teams default to saying “everyone knows this is a problem” without quantifying the prevalence or impact within your defined population. It’s also a red flag when a company describes the problem differently to how their target users do. And it’s a mistake not to provide a concise articulation of why existing solutions fail.

Consider a stress management platform that cites national data on rising stress and burnout. That data is accurate. But when the company enters enterprise conversations with their target clients, the HR leaders they’re selling to are focused on absenteeism, retention, and performance variability in specific teams. The company never clarified which subgroup experiences the most severe impact or how current solutions fail them specifically.

Without a defensible problem definition, messaging becomes generic and sales conversations stall under scrutiny. If stakeholders can’t see why your solution should even exist in the first place, there’s no point in talking about anything else.

Join the Hemingway Community

If you’ve read this far it’s probably because you want tactical information on building a successful mental health organisation? If you’d like more of that, and to meet peers building similar businesses, consider joining the Hemingway Community.

As a member of the community, you’ll join over 450 other mental health leaders and get access to exclusive content, events and resources.

Foundation 2: The Population

Does your evidence reflect the specific population you intend to serve and have you designed your product with them in mind?

Solutions that are effective for one population may fail for another. Stakeholders want to know whether your data extends across their broader employee, patient, or member population. If you cannot clearly define who your evidence applies to and who it does not, stakeholders will question its generalizability. Precision builds trust. Ambiguity weakens it.

What strong population evidence looks like

Strong population evidence typically includes clarity across four dimensions: demographics (age, gender, socioeconomic status, geography), clinical characteristics (severity, comorbidities, treatment history), behavioral patterns (health literacy, technology comfort, engagement barriers), and contextual factors (cultural considerations, access constraints, support systems).

It’s about understanding the real-world conditions that shape whether your solution will be adopted and effective. This evidence is usually built through structured user research, disciplined customer discovery, segmentation analysis, literature on population-specific intervention effectiveness, and internal data confirming that your intended population is engaging as assumed.

The most credible population evidence moves beyond description into applicability. Demographic summaries or personal documents are not sufficient. Stakeholders want to see that your data reflects the population you are targeting and that engagement, usability, and early signals are consistent within that segment. The goal is distinguishing between who shows interest and who demonstrates sustained fit.

As an example, a pre-seed company ran a feasibility study with 19 patients and conducted focus groups with eight users to understand feature needs and retention drivers within that population. They also interviewed four healthcare providers to validate clinical fit. They came out of that knowing who their product may work for, what those users needed, and whether clinicians would endorse it for that group. The feasibility study became the foundation of their pitch deck and go-to-market messaging and played a key role in securing funding.

When population evidence is weak

The red flags are easy to spot. The product is described as broadly applicable — “our product is for everyone”. There is no clearly defined core user. There are significant differences in engagement across demographic or clinical subgroups but they are unexplored. There is little data showing reach among the highest-need populations.

When population specificity is weak, the consequences tend to show up when you try to expand beyond your initial user base. The claims you used for your original users don’t extend to the new group. This doesn’t mean companies should avoid expanding into new populations. It means the evidence strategy needs to expand with them. Before entering a new segment, teams should understand the key differences between the population they have data on and the new population they’re targeting, and have a plan for testing whether their core assumptions hold. Going into a new market with a credible evidence plan is a fundamentally different conversation than discovering misalignment after the deal is signed.

Foundation 3: The Solution

Does your specific solution solve the problem?

Improvement can occur for many reasons: natural symptom fluctuation, regression to the mean, concurrent treatments, or increased motivation simply from enrolling in a program. If you cannot demonstrate how your solution specifically drives change, your claims become vulnerable. In a crowded digital health market, attribution is often what separates credible differentiation from commoditized claims.

Another example… An AI mental health platform showed that over 60% of sessions produced measurable mood and stress improvements. On the surface, that’s promising. But mood naturally fluctuates, and users tend to initiate sessions when they’re already feeling low, meaning some recovery would be expected regardless of the intervention. Without a comparison condition or a way to account for natural mood recovery, the question of attribution remained open. The results were difficult to defend in investor, partner, and buyer conversations. For this company, strengthening that link became the focus of their evidence strategy.

What strong solution evidence looks like

Before you claim impact, you need to articulate why your solution works. Strong solution evidence creates a defensible causal narrative, linking product design decisions to measurable change.

A clearly defined mechanism of action. This goes beyond naming a modality. It explains how specific features or interactions lead to measurable change. What is the behavioral, cognitive, or physiological pathway? How does engagement translate into outcomes?

Example: One youth mental health platform didn’t just claim that therapy reduces symptoms. They articulated a specific multidisciplinary model combining individual therapy, psychiatric support, and family involvement and demonstrated how that integrated pathway led to symptom reduction in 80–95% of youth served. The mechanism wasn’t “we offer mental health support.” It was a defined clinical pathway with evidence for how each component contributed to change.

Clarity around active ingredients. Not every feature drives impact equally. Credible teams examine which elements are essential versus supportive and test whether modifying or removing components changes outcomes. This separates signal from noise.

Example: A mental fitness app offered cognitive games, guided exercises, personalized training, and progress tracking. Through structured evaluation, the team developed a scientific framework that identified which specific activities drove measurable cognitive improvement and which were supportive but not essential. That clarity allowed them to prioritize development resources, refine their product around what actually worked, and defend their differentiation with evidence rather than feature lists.

Dose-response patterns. If the solution functions as intended, greater engagement with core components should align with stronger outcomes. While dose-response does not prove causation alone, it strengthens the attribution argument.

Example: A virtual mental health platform analyzed data from millions of therapy visits and found that symptoms of anxiety and depression dropped meaningfully within five sessions, with continued improvement over time. Among patients with severe symptoms, the majority experienced clinically significant reductions within eight weeks. The pattern was clear: more sustained engagement with therapy sessions aligned with stronger outcomes, reinforcing that the platform’s core service was driving change.

Comparative context. Improvement only means something relative to an alternative. Without a reference point, observed change is difficult to interpret. This may involve matched cohorts, historical baselines, comparisons to standard care, or industry benchmarks.

Example: A PTSD treatment platform analyzed outcomes from 148 adults receiving massed cognitive processing therapy — an intensive, compressed format — compared to the traditional weekly pacing of the same evidence-based treatment. Results showed significant improvement in PTSD symptoms, anxiety, and depression, with gains maintained at three months post-treatment. By comparing their delivery model against the established standard, the company could demonstrate that their approach offered a meaningfully different path to the same outcome.

When these elements align, you deliver a defensible explanation of how and why users improved. That explanatory clarity is what differentiates credible innovation from commoditized digital health tools.

When solution evidence is weak

When the solution is weak, differentiation becomes fragile. You may show improvement, but you cannot explain why your solution works or which components are responsible for outcomes.

A digital health company with 465 members had assessments that were not built on any scientific framework or foundation. There was no standardized member journey connecting solution components to expected outcomes. Without that, the company could not demonstrate which elements of their program were driving change. This made their evidence both difficult to explain and easy to challenge in stakeholder conversations.

The red flags are clear. Teams say “users improve when they engage,” but cannot define what meaningful engagement actually is. Every feature is treated as equally important. Active ingredients are not identified. Improvement is assumed to be attributable to the solution without demonstrating causal linkage. Under scrutiny, the evidence looks correlational rather than product-specific. Without clear attribution, credibility and differentiation both weaken.

Foundation 4: The Outcomes

Can you demonstrate measurable, meaningful improvement in health, behavior or economic data?

This is often the first question stakeholders ask — does it work? Outcome data is important. But impressive numbers built on a vague problem definition, an undefined population, and no attribution mechanism will collapse under scrutiny.

A common mistake is focusing on collecting outcome data before creating the infrastructure needed to measure it properly. In these cases, teams collect a lot of data, but it’s the kind of data that doesn’t withstand deeper questioning.

What strong outcome evidence looks like

Strong outcome evidence rests on a disciplined measurement strategy. It is not just about showing change. It is about showing change that is interpretable, meaningful, durable, and aligned with stakeholder priorities.

Use of validated measures. Proprietary indices or internally created metrics may be useful operationally, but stakeholders expect standardized instruments when evaluating clinical or behavioral change. Learn more about validated measures in our other Hemingway Guide.

Appropriate measurement timing. Outcomes should be assessed when change would reasonably occur. Too short, and results lack interpretability. Too long without follow-up, and durability becomes unclear.

Clarity on meaningful change. A result can show measurable improvement without that improvement being large enough to matter in real life. Stakeholders want to understand whether the magnitude of change crosses clinically meaningful thresholds and translates into real-world impact.

Durability of effects. Short-term gains may reflect novelty or initial motivation. Sustainable credibility requires evidence that improvements persist beyond the initial intervention window.

Alignment with stakeholder priorities. If enterprise buyers care about absenteeism, productivity, or healthcare utilization, those outcomes must be visible in the evidence. If payers focus on risk reduction or cost offsets, measurement strategies should reflect those endpoints. Outcomes that are not stakeholder-aligned are easy to dismiss, even when statistically robust.

Proportional evidence progression. Early-stage companies may begin with self-reported improvements or engagement correlations. As the company grows, evidence should evolve toward structured pre-post change using validated instruments, then toward comparative effectiveness, durability of effects, and clinically meaningful thresholds. Understanding this progression helps teams calibrate expectations while strengthening credibility over time.

When outcome evidence is weak

When this foundation is weak, growth conversations slow. Engagement metrics may look strong. Satisfaction scores may be high. But measurable health improvement remains unclear.

A weight management platform reports modest short-term weight loss and high user satisfaction. When speaking with a self-insured employer, however, the buyer asks about sustained weight reduction, cardiometabolic markers, and impact on healthcare utilization. The company cannot demonstrate durability of effects beyond the initial program window. Outcomes rely on proprietary or non-validated instruments rather than standardized measures. There is little evidence that the reported changes align with the buyer’s cost or utilization priorities.

The warning signs are not always about missing data — they are about weak interpretation and misalignment. Short-term gains are emphasized without follow-up data. Proximal indicators such as knowledge or awareness are measured instead of meaningful behavioral or clinical outcomes. Best-performing users are highlighted rather than average effects. Clinical significance is not distinguished from statistical significance. When outcome evidence cannot withstand deeper questioning, confidence weakens quickly.

Conclusion

Executing this strategy takes discipline. The discipline to know which assertions you need to invest in and which to deprioritise. The discipline to build evidence across each of the four foundations and to invest in the infrastructure required to do so. It also required the discipline not to make statements you can’t defend. The reward, however, is worth it. Deals will close faster and over time, you’ll become known as a trusted and credible player in the mental health ecosystem. Over time, this will become a strategic asset that creates defensibility. And in this market, that is an important asset.

That’s all for this edition of The Hemingway Report. Many thanks to Jen for sharing these hard-earned insights with us. If you found this Guide valuable, consider supporting our work by becoming a Hemingway Pro Member and joining our community of mental health innovators.

Keep fighting the good fight!

Steve

Founder of Hemingway

I didn’t.

To be honest, I didn’t know much about the Allied Arts Guild of Menlo Park. What I did know was that the best and brightest of mental health innovation were about to descend on the venue for the 2026 Demo Day of the One Mind Accelerator (OMA). Over the next twenty-four hours I’d hear from fourteen of the world’s top mental health founders and chat with the payers, policy makers and investors shaping this industry.

But at this moment, I was sitting on a wooden bench listening to the birds and learning about sheep farms from the gardener.

Well, more specifically I was learning how a young couple, Delight and Garfield Merner, had bought this land in 1929 when it was a three and a half acre sheep farm on the edge of Menlo Park. The Merners had been inspired by the craft guilds that had popped up in Europe as a reaction to industrialisation. The leaders of this movement were appalled by the lack of humanity in industrial era products and wanted to restore a sense of craft and artisanship. Their answer was to create craft guilds; associations of craftspeople who worked under shared standards, trained one another through apprenticeship, and protected the quality and dignity of their trade against industrial market forces.

The Merners wanted to bring this model to California and buying a sheep farm was their first step in doing so. On this plot of land, they designed a Spanish Colonial themed campus and in 1929 opened the Allied Arts Guild. Almost one hundred years later, dozens of craftspeople come to this space every day, to create, to collaborate and to teach curious Irishmen about the history of this special place.

On this sunny California day, while the painters painted and the potters pottered, another group of passionate individuals were taking their seats in the main hall of the Allied Arts Guild. The fourteen founders of One Mind’s 2026 Accelerator cohort had arrived and the audience were keen to hear what these innovators had to say.

Seats shuffled and silence fell as the first guest, Tim Ferriss, was brought on stage. I’m a Tim Ferriss fanboy. No, I haven’t cracked the four hour work-week (or the four hour anything for that matter), but he was one of the first “internet people” I ever followed. He’s spent hundreds of hours in my ears and I’ve always appreciated the depth of his curiosity. In real life, Tim was weirdly… normal. He talked openly about his own mental health journey, his interest in novel treatments — from psychedelics to neuromodulation — and shared some pointed advice for founders. But there was one quote from his talk that I underlined in my notebook. The importance of inflection points.

Tim was one of the first investors in Uber and he told the crowd how the success of that business was based on inflections. The iPhone had launched in 2007, meaning for the first time, people had smartphones with GPS. Mobile payments and Google Maps had also been recently developed. Without these technology inflections, Uber could not have existed. But with them, it was the perfect time for a ride-sharing business to be created. It was almost inevitable.

Tim invited all of us to think about the inflections happening in mental health. And because I’m a fanboy, I dutifully accepted his invitation. As each founder stood on stage, presenting their business, I thought of the inflections that might be shaping the opportunities in front of them.

The first inflection that was quite clear to me was the market shift towards outcomes. In the US, healthcare spending on mental health and substance use disorders has increased from $41B in 2001 to $140B by 2021. Today, it’s even higher. This growth was driven by an increase in access to care — 87.3% of spending growth is accounted for by an increase in the volume of cases. This market has been defined by this trait of increasing access. But that is changing. Payer behavioural health costs are increasing between 10% and 20% each year. Whether we like it or not, this is unsustainable for them. As a reaction, payers are increasingly demanding outcomes evidence, not just utilisation data, to justify continued investment. While access to care is not a solved problem, the market is clearly shifting towards outcomes and companies are reacting.

Some companies are better placed than others to take advantage of this inflection. I’m particularly interested in those who can deliver superior outcomes by focusing on a specific population or condition and delivering a comprehensive care model designed for them. Carmine di Maro calls them “Super Servers”.

Nosis Health is one such Super Server. Duffy Fallon is its Founder and you can’t help but like the guy. When he was on stage he shared the story of how his own recovery journey inspired the idea for Nosis to treat substance use disorder as a whole-person, chronic condition. They do this by wrapping addiction medicine, diagnostics, therapy, coaching, lifestyle interventions, GLP-1s and more into a single care programme. So far, they’re delivering impressive outcomes — 77% primary substance abstinence, 90% retention of patients in care, and improvements in key biomarkers like A1C, CRP, and ApoB. They’re small samples but they show high potential. Duffy mentioned some inflection points he was seeing in his own work, specifically, the cultural momentum behind recovery and whole-person care, the emergence of GLP-1s in SUD treatment and the ability for AI to personalise care (more on that later). Nosis is still in the early chapters of its journey but they’ve already bootstrapped to $250K in ARR and their outcomes are impressive.

Another OMA company working at an intersection of inflections is Birches Health. Digital addictions are on the rise — 72 million people in the US now have a daily gambling habit — but most providers are not specialised in treatment for these conditions. As a result, it’s hard for them to deliver good outcomes. Birches have built a specialised clinical network for these digital behavioural addictions, recruiting and training providers in their methodology. With this, 49% of all Birches patients achieved clinically meaningful improvement and that rises to 68% when focused on moderate to extreme patients¹. This focus on outcomes is driving commercial results. In 2025 they grew revenue 4X and have raised over $20M.

Damayanti Dipayana is the CEO and Founder of Manatee Health and at the Demo Day she shared her thesis; that by engaging the whole family in care they can deliver better outcomes. Manatee delivers outpatient mental health care for the whole family, including kids and caregivers and so far Damayanti’s thesis seems to be proving true. 87% of children in Manatee care improve at discharge vs. 28% in traditional care (based on PHQ9, GAD7, PSC17) and families stay for twenty-two sessions vs. four in child-only sessions in standard outpatient care. When Tim Ferriss was speaking, people wanted his advice on how to grow your business. He said that having a truly awesome product is the most important factor. Manatee gets all of its referrals directly from health plans and providers and they’re growing 5X year on year. Yes, growth can be hard. But if you make something that truly makes people better, at higher rates than your competitors, growth will come.

In one of the breaks from the presentations, I got talking to some of the team from Radicle Science. I had been thinking for a while about this market inflection towards outcomes. Everyone needs better outcomes data now, but building that evidence base as a startup is quite hard. If outcomes are the gold in this new market, Radicle Science are creating the shovels. Radicle run large-scale, automated clinical trials for non-prescription mental health treatments, giving companies the outcomes evidence that payers are now demanding. The idea is to make evidence generation much faster and cheaper than traditional trials. They’re also working on an interesting new initiative called “Certified Effective™” which aims to be a consumer standard for wellness solutions, allowing brands to be “certified” for the wellness claims they make. In mental health, there are a lot of wellness products but it’s impossible for consumers to know what will work and what won’t. Radicle Science may bring some much needed clarity to this murky market.

Policy inflections can reshape markets too. Today, there are three different policy shifts creating opportunities for the companies in this cohort.

The first is in psychedelics. On April 18th Trump signed an executive order directing the FDA to prioritise psychedelic drug review and allocating $50M through ARPA-H to match investments made by state governments for psychedelic research. The FDA has already responded, issuing National Priority Vouchers to three companies developing psilocybin and methylone therapies. This is directly relevant to companies like Biomia, who are developing nature-inspired CNS therapeutics including an early-stage ibogaine programme for opioid use disorder.

The second is deprescription. On Monday of this week (May 4th), the HHS launched a formal action plan to curb psychiatric overprescribing, with US Health Secretary Kennedy calling out the “overmedicalization” of mental health, particularly in children. Approximately one in six US adults is now on an SSRI, and whatever you think about the credibility of the HHS’s move to challenge the prescription rates of SSRIs, it is a clear policy inflection point for the industry. Specifically, CMS is creating a new method for clinicians to get paid for the time they spend helping patients discontinue medications. Mr. Kennedy also published a “Dear Colleague” letter earlier this week, directing providers “to expand the use of nonpharmacologic treatments and to strengthen informed consent and shared decision making”. Such directives are a huge tailwind for OMA company Flow Neuroscience. Flow have built an at-home brain stimulation headset for depression which is non-invasive, drug-free, and clinically validated. Earlier this year, the device was approved by the FDA. In a policy environment that is actively questioning the dominance of SSRIs, and encouraging nonpharmacologic interventions, devices like this are likely to see increased adoption.

On one of my strolls along the cobblestone paths of the Allied Arts Guild I bumped into Chris Appleton. Chris is the founder of SocialRx, a social prescribing platform that connects patients to community-based arts, culture, and nature experiences. The idea is to help people build a sense of connection, belonging and purpose. It’s hard to ignore the importance of these factors in mental healthcare and the outcomes from SocialRx offer strong evidence for their inclusion in our health systems. 79% of SocialRx members improve their overall wellbeing and they’ve seen a 63% reduction in avoidable ED visits. Business is good too. They’re growing 3X YoY with an ARR of $1.7M and 65% gross margins.

If policy continues to shift toward whole-person, non-pharmacological approaches, businesses like Flow and SocialRx will continue to benefit.

The third policy inflection is a more discrete one. CMS is moving to include depression screening and follow-up as a quality measure for Medicare Advantage plans. I had a good chat with Ben Gardner, CEO of Vitalic Health about their model and how this policy is supportive of what they’re trying to do. Vitalic is exclusively focused on over sixty-five year olds and they have built an AI screening platform that finds and screens members in this population group. Then, they have geriatric-specialised teams that provide care to these people. Plans pay for both of these services which is pretty neat — essentially, they’ve turned their acquisition costs into a revenue line. They have great outcomes too — a peer-reviewed study with VNS Health showed a 56% improvement in depression and anxiety scores, a 24% reduction in ED visits, and a 33% reduction in inpatient stays. Again, commercials have followed outcomes — Vitalic are already at $1.7M ARR, with >60% margins and strong quarter on quarter growth.

When the Merners created the Allied Arts Guild back in 1929, they could barely have imagined the technological innovations that the following century would bring. As I sat in one of the courtyards listening to the trickling of a fountain I couldn’t help but wonder what they would think of today’s world. What would they think about the scale of industrialisation? What would they think about mobile phones and social media? Or more specifically, the apparent lack of humanity used in the crafting of many of these innovations. What would they think about AI?

We all might long for technology development to reflect the kind of thoughtful craftsmanship the Merners championed. But regardless of the intent or process behind new technology, we cannot deny that it is reshaping our world. In 2026, the strongest technological force is undoubtedly AI. In mental health it has now moved past the point of hype and presents an inflection point for the industry.

The first and obvious application is in conversational AI. Limbic is an OMA graduate that develops Clinical AI which has supported more than seven hundred thousand patients across 63% of the NHS talking therapies in England and thirteen US states. It is the only AI mental health chatbot approved as a Class II medical device and has seven peer-reviewed clinical studies including two in Nature Medicine. As I pointed out in a recent report many conversational AI products are yet to prove their worth. But companies with specific applications and evidence bases like Limbic are taking advantage of a powerful technological inflection.

Presentations from some of the other OMA founders reminded me that conversational AI is only one application of this technology. We often overlook the role it can play in helping us to better understand mental illness and discover and develop new treatments.

Forecast Bio grow human brain tissue in the lab and measure how drugs affect neural circuit activity. They then link those measurements to real-world patient outcomes in the hope of compressing neuropsychiatric drug assessment from years to a single assay. Orbit Neuro is another company applying AI in interesting ways by collecting high-resolution neural data through a novel brain wearable. This helps build the datasets that many other therapeutic discovery models need to work. Eratos are working on similar but different problems, mapping the physical structure of the brain down to the nanoscale relationships between proteins and synapses in actual brain tissue. Most CNS drug discovery has been working with an incomplete picture of what’s actually happening in the brain. The more precisely we can see the architecture, the more precisely we can identify what to target. While Eratos’ core competencies are in spatial biology and microscopy, AI is allowing them to do work that would have been impossible only a few years ago.

AI is also being used to personalise treatment. As we know, psychiatric diagnosis is largely symptom-based and treatment is often conducted through a trial-and-error approach — patients cycle through medications before finding one that works. Engram are betting that multimodal biological data, processed at scale, can replace that guesswork with precision. To do this, they are building a foundation model that can subtype psychiatric disease and match patients to treatments. Many providers want to implement more personalised treatment to deliver better outcomes. Anything Engram or others can do to support that, will be warmly welcomed. Monument Therapeutics are also betting on a precision approach, using proprietary digital biomarkers to identify patients with homogeneous underlying neurobiology and match them with targeted compounds. Their lead programme, MT1988, is focused on cognitive symptoms of schizophrenia like memory, attention and decision making for which there are currently no approved treatments.

When all the pitches were done the crowd filed into the exhibition hall next door. Looking around that room, I couldn’t help but be impressed by what One Mind had built with this accelerator. In just four years they’ve created what I’m very comfortable naming the world’s best mental health accelerator. Yes, they’ve brought together the world’s top founders. But they’ve also spread their arms wide and welcomed/dragged in the other members of the ecosystem that are needed for real change to happen. They have a lived experience council, a payers council, a provider council and a scientific advisory board. They bring in investors, policy makers and representatives from the non-profit space too. They even recently launched the OMA Catalyst Fund to invest in select companies that complete their accelerator.

In 1929, the Merners started with a sheep farm and created the Allied Arts Guild to provide a supportive environment for artists to craft better objects for the world.

In a way, One Mind have created a guild of their own.

That’s all for this week. Many thanks to the One Mind crew and everyone I spoke to during my visit. If you enjoyed this piece, consider supporting my work by becoming a Hemingway Pro Member and joining our community of mental health innovators.

Keep fighting the good fight!

Steve

Founder of Hemingway

Notes:

[1] Clinically meaningful improvement is defined as 4+ point G-SAS reduction from baseline.

[2] You can learn more about the OMA 2026 cohort here.

Many startups, incumbents, and research institutions are building products in this space. They hope to extend access to support, reduce cost, and improve outcomes. Builders have moved fast and the market is moving quickly too, with regulations, safety standards and the available evidence base changing weekly.

This report on Conversational AI in mental health is an attempt to bring greater transparency to the field. We map the landscape from a business, product and clinical perspective, covering thirty-one conversational AI products and analysed across nine dimensions including clinical evidence, selected business models and product features. This report provides a map of the market divided into three segments, shares the primary insights from our analysis and a discussion of the key trends and gaps that we observed.

Let’s get into it.

Disclaimer: This market map is unlikely to be exhaustive. It relies on publicly available information (which comes with limitations), does not rank products or make treatment recommendations. Its aim is to bring transparency to a fast-moving area of innovation, and we encourage ecosystem members to contribute to it as a living project. Details on how to do so and additional notes can be found at the end of this report.

Methodology

First, a quick note on our methodology.

We identified 31 products from an initial set of 70+ by filtering for generative, multi-turn conversational AI with mental health as its primary purpose and a verifiable market presence. This excluded general-purpose AI (like ChatGPT), clinician-only tools, and assessment-only tools. Each product was coded across clinical positioning, evidence, human involvement, business model, and regulatory status by a cross-functional team from the Hemingway community with academic, clinical, and commercial backgrounds. The full methodology, product database, and links for feedback are included at the end of this report.

Market Map

The 31 products were clustered into three meaningful segments, Scaled Hybrid Platforms, Clinical Infrastructure and Consumer-First.

Download the high-res Market Map Image

Scaled Hybrid Platforms (n=6): Headspace, Spring Health, Lyra, Grow Therapy, SonderMind, Sword Health

Scaled Hybrid Platforms are established, later-stage companies — typically Series B and beyond — that have added conversational AI capabilities to an existing therapy or clinical services business. Their AI products sit alongside a broader offering that includes human clinicians, and all operate with some level of clinician-in-the-loop. They reach users primarily through employer-sponsored or health system channels and carry the infrastructure — compliance, clinical governance, etc. — that institutional buyers require. Evidence for the efficacy of the conversational AI products created by these businesses is limited, with only Spring Health reporting controlled study-level evidence. The strategic question for this segment is whether their AI tools become a meaningful product in their own right or remain a feature of the broader platform. It is likely that many more businesses (outside of those listed in this market map) are working on products in this space, however they have not yet been publicly released. For example Talkspace have announced that they are building a conversational AI product but that it won’t be released until the summer of 2026.

Segment 2: Consumer-First AI (n=16) — Slingshot/Ash, Sonia, Noah AI, Youper, Yana, Kin, Rosebud, Manifest, Yuna, Earkick, Ahead, Kai, Feeling Great, Inner Vault, Elomia, Flourish/Sunnie

This is the largest and most heterogeneous segment. These products go direct to consumers via freemium or subscription models and almost universally operate with zero human involvement. Evidence levels vary widely — from RCT (Elomia, Flourish) to none at all. The level of clinician involvement in product development also varies widely. The segment is defined less by a shared clinical philosophy than by a shared distribution strategy: reach users directly, at low cost, without the friction of institutional sales cycles. Most are pre-seed or seed stage.

Segment 3: Clinical Infrastructure (n=9) — Wysa, ieso/Velora, Alongside/Kiwi, Wayhaven, Limbic, Jimini Health, Brightn, Affiniti, Xaia

This segment describes companies building conversational AI products designed to sit within or alongside another organisation’s clinical services and workflows. These tools augment what clinicians and care systems already do — whether through between-session support, measurement-based care, or structured intervention delivery that feeds back into a treatment plan. The user may interact with the AI directly, but like most of the Scaled Hybrid Platform organisations, the product is architected around the clinician relationship, not as a substitute for it. This segment contains the highest concentration of clinical evidence in the dataset, including products with RCT or multiple-RCT-level validation (Wysa, ieso, Limbic). Business models skew toward health system licensing and API/integration plays reflecting the segment’s orientation toward institutional buyers. They are distinct from Scaled Hybrid Platforms in that they do not tend to employ clinicians to deliver care directly, instead relying on their clients — provider organisations, health systems, educational institutions — to deliver care alongside their technical tools.

Key Insights

After reviewing the collected data, we noted eleven key insights on the market.

1. This is still an emerging market.

The market clearly sees potential in these products and are investing to build them out. But conversational AI for mental health is still in its early innings, and our understanding of its impact is similarly nascent. Only five products have RCT-level evidence or above (Elomia, Flourish/Sunnie, ieso/Velora, Limbic, and Wysa), while eleven - more than a third - have no published clinical evidence at all. A lot of the activity in this space is being driven by very young companies that have not yet been through rigorous clinical validation. Nineteen of thirty-one products (61%) are Pre-Seed or Seed stage. There is strong evidence for the underlying therapeutic modalities that most products claim to use (e.g., CBT, ACT); however what is largely unknown in this new AI context is the way in which these modalities are being delivered and the impact of new elements in the equation (e.g., therapeutic drift from a conversational agent or potential replacement of important offline activities/interactions). While this evidence base is now being rapidly developed, this is a market still in the early stages of its definition.

2. Evidence and scale do not correlate.

The largest companies in this market have generated little clinical evidence for their AI products specifically. Among the Scaled Hybrid Platforms most have no published clinical evidence for their conversational AI features. These companies are not asking their AI to be the intervention — at least, not at this point — so it’s an important question to consider what evidence they do need. Their core clinical products are human-delivered support, coaching, or blended care models, many of which carry substantial evidence bases of their own. The AI layer typically serves an adjacent function: intake, between-session engagement, or content delivery that supports the primary clinical relationship. In that context, the evidentiary question is not whether the AI works as a standalone treatment but whether it increases access, improves retention, engagement, or clinical efficiency within an already-validated care model.
For smaller companies pursuing a clinical route, the calculus is inverted. If your AI product is the intervention — if there is no clinician behind it — then evidence is the entire basis of your credibility with payers, regulators, and health systems. However, because many of these companies are currently going direct to consumer, credibility with those parties is less important.

3. No dominant player has emerged.

Despite significant consumer demand, we found no product that had gained significantly more users compared to the rest of the peer set. What’s more, none of the products analysed come close to the popularity of existing general-purpose AI chatbots (e.g. ChatGPT, Gemini, Claude) which we know are often used for mental health purposes despite not being designed to do so. Many products have coalesced around the medium of a text-based chat user interface and CBT and emotional support dominate the therapeutic modalities underlying the products. Therefore, many of these products look similar and attempt to serve similar purposes for users. So far, no product has clearly separated itself in the market in terms of adoption and all are dwarfed by the usage of general purpose chatbots.

4. CBT is the dominant therapeutic modality and most products cluster around the same use cases.

CBT appears in twenty-one of thirty-one products, followed by mindfulness (sixteen) and ACT (thirteen). On the use-case side, anxiety (twenty-three), general well-being (twenty-three), stress (twenty), and depression (fifteen) account for the vast majority. Meanwhile, substance use, eating disorders, trauma/PTSD, and chronic pain each appear in only one or two products. The market is densely concentrated on mild-to-moderate anxiety and stress, with very little coverage of more complex or less prevalent conditions.

5. Nearly every product positions as “wellness/coaching”.

Almost all products position themselves as wellness or coaching rather than treatment. Feeling Great explicitly claims a treatment positioning, yet it has no published clinical evidence and operates with zero human involvement. Another product, Inner Vault, calls itself AI Therapy. But these are the exception. The near-universal wellness framing reflects regulatory caution and go-to-market strategy: it avoids the higher costs and regulatory burdens associated with clinical treatment claims. But for clinicians and policymakers, it creates a transparency problem. Users may be receiving something that looks and feels like a therapeutic interaction without the oversight or evidence requirements that a treatment designation would demand.

6. Two-thirds of products operate with zero human clinical involvement.

Nineteen of thirty-one products operate with no clinician in the loop. Among those fully autonomous products, six have no clinical evidence. Fully autonomous models are more prevalent in early stage, consumer first AI products while larger, hybrid care providers all have a human in the loop. However, for those who do involve clinicians, the nature of that involvement isn’t always defined or consistent, and it’s difficult to know what that actually implies for a product’s validation or safety.

7. Most products include clinicians in their development.

While several products don’t include humans in their actual product experience, most included clinicians in the development of their product. Twenty-six of thirty-one products have documented clinician involvement of some kind in their development. We noticed clinicians being involved as co-founders, early team members, clinical advisors, scientists/researchers, or design partners. Of the five for which we could not find evidence of clinician involvement in their design, four were in the Consumer-First AI category.

8. The line between a support tool and a substitute for human connection is largely unaddressed.

Several products in the dataset emphasise emotional closeness, continuous availability, and affirmation as core features - especially within the Consumer-First AI category. The clinical implications of this design orientation — particularly for vulnerable users — are not well understood and have been linked to potential harm for users. Few products make explicit distinctions between tools designed to scaffold human connection and those that may inadvertently replace it.

9. Products serving adolescents vary enormously in evidence and safety infrastructure.

Nine products serve adolescents (ages 13–17), but their evidence levels range from RCT (Flourish, Wysa, Elomia) down to zero published evidence (Noah AI). Privacy practices are equally uneven: some products have no publicly posted privacy information. Given the heightened duty of care owed to minors, this variability is a notable finding for regulators, parents, and institutional purchasers alike. This is a significant gap given the well-documented importance of early intervention in mental health, and the reality that younger users who cannot find appropriate tools will default to general-purpose, generative AI chatbots that were not designed for their needs nor their vulnerabilities.

10. Products are largely built on top of general-purpose models.

A small minority of products describe using a “proprietary or custom” model. The vast majority are built on top of general-purpose models from providers like OpenAI, Anthropic, and Google, with model behaviour adjusted through system prompts, fine-tuning, retrieval-augmented generation (RAG), external safety filters and other processes. These tools give product teams meaningful control over typical model behaviour — therapeutic tone, modality adherence, crisis escalation. However, they cannot change the base model’s architecture, core training data, or deep learned patterns, and they cannot prevent the model provider from updating the foundation underneath them.
At the infrastructure level, products are far less differentiated than one might imagine. Therefore, any potential moat for these companies does not sit at the model layer — it sits in the clinical design, safety architecture, the quality of proprietary data feeding back into the system, the evidence base, and the distribution relationships built on top of it. We did not observe any difference in available clinical evidence or any other dimension between products built on proprietary models and those built on top of general-purpose models.

11. The market remains remarkably opaque.

Despite surveying thirty-one products across nine dimensions, basic questions about quality, safety, and underlying architecture remain difficult to answer with confidence. Four products — including one serving adolescents — have no publicly available privacy information. Evidence claims are often difficult to verify independently, and the line between internal validation and peer-reviewed research is often not made clear. Clinical positioning language varies so widely that products offering substantively similar experiences to users can describe themselves in very different terms. For clinicians looking to refer patients, employers selecting vendors, policymakers writing guidelines, and users making choices about their own care, informed decision-making is difficult due to incomplete public information. Transparency is not yet the norm, and until it is, independent efforts to map and evaluate this landscape will remain necessary — and necessarily imperfect.

Become a Hemingway Pro Member

If you enjoy analysis like this and would like to connect with 450+ leaders in mental health innovation, consider becoming a Hemingway Pro Member. You’ll get access to premium content, invite-only events and access to our private Slack group.

Become a Member

Discussion

This report set out to map the conversational AI landscape in mental health. The aim was to add transparency to the ecosystem and encourage thoughtful innovation to improve population mental health. Based on our analysis, there are a number of interesting discussion areas for the field.

1. More insight is needed on safety, therapeutic quality and impact

This report did not assess product safety in any systematic way. We would be unable to identify which products handle crisis situations well, which have robust content moderation, or which have been tested against adversarial use. In our efforts to try to track safety, the best we could do for an objective measure was to assess what safety features were available in a product. But even doing that was challenging due to the lack of public information on this topic from companies. A number of safety benchmarks are emerging in this space which would offer more consistent and transparent safety assessments, but product performance on those benchmarks is for the most part, currently unavailable. This report did not evaluate therapeutic quality or user outcomes either. These are the most important questions for this space and we would hope to see them addressed in other work.

2. Greater transparency can enable better clinician and consumer choice

This project found it genuinely difficult to assess many of these products, even with a cross-functional team and a deliberate methodology. If it is hard for a group of researchers, clinicians, and industry professionals to evaluate these tools, it is functionally impossible for an individual user downloading an app from a store. The opacity is not necessarily intentional — many companies are early-stage and still building — but the effect is the same. Users cannot make informed choices about products they cannot meaningfully evaluate.

3. How to ensure AI addresses underserved areas

As noted in the insights section, many of the solutions in the market appear quite similar. Most cluster around the same therapeutic modalities, the same use cases, the same user demographics, and the same wellness positioning. Most are text-based, most draw on CBT, most target anxiety and stress in young and working-age adults, and most avoid making clinical claims. This convergence likely reflects a rational response to where the evidence is strongest, where the regulatory risk is lowest, and where the commercial opportunity is most immediate. But it also means that the populations and conditions most underserved by traditional mental health systems — children, older adults, people with substance use disorders, eating disorders, trauma, or serious mental illness — may remain underserved by AI too. This pattern likely reflects not only market incentives but also appropriate clinical caution: these populations often involve greater complexity and risk, and current evidence suggests AI is best suited, at least for now, to more structured or lower-risk use cases.

4. The opportunity (and risks) of more disruptive innovation

This also raises a deeper question about the scope of innovation. Almost every product in this dataset is attempting to deliver or augment an existing therapeutic protocol - CBT, ACT, mindfulness - using a new technology. That is an understandable approach as it allows builders to draw on decades of clinical evidence. It means most of the market is trying to answer the question: can AI deliver elements of existing, human-based care, but at a lower cost, greater scale and perhaps with some additional efficacy? What remains largely unexplored is whether conversational AI could enable fundamentally new forms of intervention — approaches that would not be possible, practical, or even conceivable in traditional care settings.

Very few products in this dataset are exploring that frontier, suggesting the field is still thinking of AI as an adjunct or comparable alternative to human-based care, rather than an entirely new medium for the prevention and treatment of mental illness. Whether industry is the right place to test fundamentally new forms of intervention is an open and important question. Incentive structures are not always set up to have effectiveness as a top priority. New AI-powered modalities which might be worth exploring, perhaps need to developed in a clinical-research environment first. That said, startups can be useful bodies for turning nascent science into products people actually use. The best breeding ground for such innovation is an interesting question to consider.

5. What evidence, for whom?

The evidence question is more nuanced than developing a simple league table of who has trials and who does not. As this report noted, the largest platforms have little published evidence for their AI features specifically, but their AI is not the intervention — it sits within a broader clinical model. For smaller companies who target the healthcare system, evidence is an important part of their strategy to build credibility. Then there are consumer-facing products operating outside institutional channels entirely. Each of these segments have different user expectations and commercial strategies raising an interesting question; what evidence should they be generating?
Not every product needs to complete RCTs and pursue FDA approval to deliver meaningful population health benefits. Many of these businesses would not be capable of doing so due to the cost and duration of such pursuits. So what sort of evidence should we demand from the different segments of this market? All products should be able to demonstrate an acceptable level of safety — although we are yet to clearly define what that should be. And their claims should be in line with what they’re capable of delivering. But what else? Providing clear guidance to the market, especially this category of consumer-focused products is important. However, it must be done in a way that still encourages innovation while ensuring safety and user outcomes are prioritised.

6. Competing with general-purpose models

The products in this report are purpose-built for mental health. But general purpose models — from OpenAI, Anthropic, Google, and others — are already being used for mental health support by millions of people. That usage dwarfs the usage of all products mentioned in this report. If a general-purpose chatbot is liked by users, and can deliver passable interventions (e.g., CBT-informed conversations), the value proposition to users of a dedicated mental health product rests on what it adds beyond the base model: clinical governance, safety protocols, evidence, human oversight, integration with care systems and importantly, the clinical context of the user themselves.

As mentioned before, the majority of products in this dataset are built on top of those same general-purpose models. These products may benefit from improvements in the underlying models. But those improvements may also make the general-purpose models more attractive to users.

The products that gain adoption will be those that can take the best of the underlying models and layer meaningful improvements that users, clinicians and payers care about. If that added value isn’t apparent, especially to the user, general-purpose models will remain the dominant source of AI mental health support.

Conclusion

This is an innovation area with potential to significantly extend and improve mental health support. It has attracted capital, attention and some of the world’s best talent. But potential is not impact. Thorough evidence generation, transparency, product innovation and listening to real users is the work ahead.

We intend for this report to support the ecosystem by creating greater transparency in the field. It is a living project and we welcome contributions and collaboration from across the ecosystem. To do so, please submit this form.

Acknowledgements and Contributors

This report was produced by The Hemingway Community. The following individuals contributed to research, data collection, product coding, clinical review, and/or editorial input. Their participation does not imply endorsement of the report’s conclusions, nor of any product or company discussed within it. Contributions reflect individual professional expertise and are not based on primary data collection or direct verification of individual company claims. Contributions do not constitute a clinical audit, safety validation, or endorsement of any specific product or company. Some contributors may hold roles or be affiliated with the organisations included in this report.

Danielle Vaeth - Contributor
Jackie Ourman, LMHC-D - Contributor
Michael Trang - Project Co-Lead
Michiel van Vliet - Contributor
Molly Fuller - Contributor
Nathaniel Hundt - Contributor
Nikki Huang - Contributor
Pavithra Ramesh - Contributor
Steve Duke - Project Co-Lead & Author

Many thanks to everyone who contributed to this report.

Notes

• This report does not rank products or make any treatment recommendations. Rather, our goal is to provide increased transparency for this emerging and rapidly developing market and to encourage thoughtful innovation that can improve population mental health.

• Full details on our methodology and approach are linked below.

• Some contributors may hold roles or be affiliated with the organisations included in this report.

• This map is unlikely to be exhaustive and the data has limitations. It is entirely possible that some products and data were missed, especially given the fast-moving nature of this space. Many companies disclose little about their technology, clinical oversight, or safety protocols. Some fields are unknown - not because the information doesn’t exist, but because it isn’t public. Our aim is for this to be a living database and we encourage contribution from the broader ecosystem.

Resources and Feedback

• High-Resolution Market Map Image

• Methodology and Criteria Documentation

• Conversational AI in Mental Health Database

• For questions, updates, or to contribute to this ongoing project please submit this form.

If you’d like to discuss this report with other mental health leaders, or work on future projects like this, consider joining The Hemingway Community.

Keep fighting the good fight!

Steve

Founder of Hemingway

It’s 3am in the Neonatal Intensive Care Unit.

It’s been a busy night, so Dr. Jessica Gaulton (Jess) takes advantage of a quiet moment to type up some patient notes. On her desk there are things you might expect from a neonatologist fellow working eighty hour weeks: a computer, a keyboard with faded letters, discarded snack wrappers, but they’re competing for space with pumping equipment and baby bottles. As usual, Jess is multi-tasking. While she types, she’s also pumping milk for her six-week old newborn, who’s at home with her eighteen month old son.

Then her pager beeps. A baby is coding. Jess jumps up, throws off her pump and runs. She finds the baby, performs CPR and saves the baby’s life. She leaves and walks back towards the call room to finish her notes. As she opens the door she notices her pumping equipment on the floor. Then she notices that it’s sitting in a pool of milk. Milk that has spilled from all the bottles she just pumped for her newborn.

She breaks down.

This was the start of a new chapter in Jess’ journey with postpartum depression.

But it was also the start of her journey to build FamilyWell, a company improving women’s mental health by embedding care directly into OB/GYN practices. They’ve built a care model that helps thousands of Moms across five states and last year, the business raised an $8M Series A.

Today, Jess is the CEO of the business. She is also a practicing neonatologist at Harvard, an institution from which she also holds a Masters in Public Health.

In this Hemingway Guide, we tell Jess’ story and share the business and life lessons that she’s learned along the way.

Building embedded care

It’s 2020, and Jess is going through her second pregnancy. She has an excellent OB/GYN, but despite the load she is carrying at work and at home, she never thinks to turn to her for emotional support. It’s clear to her that there’s a disconnect between women’s mental health needs and their medical providers and it leaves women like Jess without the support they need. After her own experience of postpartum depression, Jess wants to solve this.

Her first solution is a peer-based texting platform. It gets early traction and proves that women want support and will engage with coaches. But when those women need deeper support from therapists or with medication, it’s not available to them. So Jess decides to build an embedded care model that puts a full suite of mental health services directly into OB/GYN practices where women already get care.

Mental health screening, care coordination, therapy, and psychiatric oversight are embedded within the OB/GYN workflow. There’s no external referral, no handoff to an unknown provider, and no break in trust.

Building an innovative care model like this and getting it adopted at scale is not easy. Over the last four years, Jess has learned some important lessons about partnerships, care model design, adoption, go-to market and fundraising that helped her along the way.

1. Be highly selective with your first partnership - it can make or break you

FamilyWell’s first, and most important partnership, came from a blog post.

Jess wrote a message to the executive director of the Massachusetts Postpartum Depression Fund, a nonprofit advocacy group for maternal mental health. The director believed in what Jess was building and invited her to write a blog post for their audience. One of the readers was Dr. Melissa Sherman - a prominent OB/GYN at Essex OB/GYN Associates and a long-time advocate for perinatal mental health. She reached out to Jess. She wanted to pilot FamilyWell’s model at her practice.

Sometimes the stars align, and you need that to help you get off the ground. I was very fortunate. But you have to be really selective about who you partner with - especially for your first partnership.

This partnership became the foundation for everything FamilyWell would go on to build.

Jess and the Essex team spent a full year together - testing, refining, and ironing any kinks in the model before rolling it out to other practices. Jess firmly believes if they had partnered with a practice that wasn’t fully bought in, or that didn’t have the operational framework to iterate quickly, it all could have been a very different story.

What to look for in your first partner
Jess identifies three things that make an early partnership work. First, is your champion a key decision maker? Dr. Sherman had real authority in her practice. She could commit to a pilot with FamilyWell and push through the changes needed to make it work. Second, does the broader team buy in? The whole practice at Essex was able to get behind this new initiative, which mattered massively for day-to-day execution. Third, does your champion have a track record of innovation? When Jess first met Dr. Sherman, she had already built an innovative substance use disorder program at her practice. Jess didn’t realise at the time just how important this was, but it’s critical to have a partner that can operationalise something new - not just be a cheerleader for it.

How to get the right partner on board
Jess didn’t actually find Dr. Sherman. Dr. Sherman found her. But that wasn’t a coincidence. Jess’ advice is to start by identifying the credible people and organisations who are already advocating for the problem you’re solving - the more niche the better. For her, that was a small nonprofit that happened to be the leading voice in maternal mental health in Massachusetts. Then, write for them. Speak to their audience. Plant a flag for what you believe in and make it easy for the believers to find you.

Many founders acknowledge the importance of thought leadership but let it inevitably slide to the bottom of their to do list. Jess’ advice to combat this is to be extremely strategic about who you work with.

If I’m going to spend time on something like a thought leadership piece, it’s got to fit into what our goals are for the year. Or for the quarter.

Jess said yes to the Mass PPD Fund blog because they were the leading advocacy voice in maternal mental health in Massachusetts. It wasn’t a big audience but it was a highly credible channel speaking directly to the niche of people she wanted to reach.

2. Use design thinking to iterate on your care model

The Collaborative Care Model was originally designed with a single behavioural health care manager (BHCM) handling everything: coordinating care, conducting brief therapy, tracking patient progress, managing documentation. In practice, Jess found that this model wasn’t built to scale in real world practices.

We learned that narrowing and redefining the role - and ultimately evolving to a ‘team care manager’ approach - was essential. We broke the responsibilities into several distinct roles, and today it’s a team, not a single person, addressing the full scope of what the BHCM was meant to do.

Evidence-based models are designed for specific contexts. When you move them into a different context - higher volume, different staffing economics, a different patient population - fidelity to the original design isn’t always a virtue.

If I were starting over, I’d move to that team-based model much sooner and let go of the idea that fidelity meant rigidity.

While Jess’ ability to adapt the care model to real world needs is valuable, her meta learning has been to embed design thinking and human centred design into the business in a way that makes this sort of iteration and continuous improvement part of FamilyWell’s DNA. As they enter new markets and build new products, they can flex this muscle to create services that outperform competitors.

3. Don’t wait for the system to change, use existing reimbursement infrastructure

FamilyWell has a clear way to get paid. The whole model runs entirely on existing billing codes for therapy, psychiatry and other mental health services. Collaborative Care billing codes ensure that the OB/GYN practices get paid for their extra work to facilitate the mental health care for their patients. Jess feels strongly that founders need to be very clear on how they will get paid.

Know exactly how you get paid, and avoid building a model that depends on the system changing to accommodate you.

This was important when Jess was raising FamilyWell’s Series A, she wasn’t asking investors to bet on some future reimbursement model. The financial infrastructure already existed and FamilyWell (and its customers) plugged directly into it.

My biggest piece of advice is to be clear about how you get paid and to avoid swimming upstream. If your model depends on creating entirely new reimbursement pathways or waiting years for new CPT codes, that’s a long and risky road, especially in today’s funding environment.

New CPT codes take years. Be wary of a commercial strategy that requires something that doesn’t exist yet. Many mental health innovations have died waiting for such system change.

4. Make sure your product is used by providers, before you try to grow

Good growth happens backwards. That means solving for adoption first. When your product is adopted and actually used repeatedly, you’ll have high retention. With high retention, you get higher customer lifetime value (CLTV) which gives you both the unit economics and the confidence you need to invest in marketing and sales. Most people do this backwards. They start by focusing on closing more deals before they know that they have a high retention product.

But Jess learned this lesson early on and focused on maximising adoption with their OB/GYN customers so that they would have a high retention product. She spent time understanding their perspectives - their incentives, their fears about liability, their staffing realities, and the operational pressure they’re under.

My core framework is simple: start with your customer, not the model. In embedded care, it’s easy to design around the patient experience alone, but if you don’t deeply understand the clinicians and practices who have to make it work day to day, the model will fail.

Then she focused on three things to ensure FamilyWell adoption was high in the practices they served.

Make it easy for the practice
The OB/GYN practice doesn’t have to do any of the heavy lifting to adopt FamilyWell. FamilyWell hires the mental health team, handles all scheduling and paperwork, takes care of billing, and manages compliance. From the practice’s perspective, the service slots into the workflows that already exist.

Make it easy for the provider (and build trust)
Outside of solving problems for the practice, you need to solve problems for the individual providers you want to use your service. Make life easier for them, reduce their workload, give them easy referral pathways. And most importantly, build trust with them.

Build something that reduces burden for clinicians, fits into existing workflows, and earns trust, because no matter how evidence-based the model is, it only works if providers actually use it.

Create a financial incentive
FamilyWell gives practices a way to make money. By using Collaborative Care billing codes that insurers - including Medicare and Medicaid - already pay for, the practice earns incremental revenue for delivering a service they want to deliver, without taking on new costs.

If you get this right, every party gets value from the model and incentives are aligned. This is crucial for driving adoption.

Patients get timely, trusted, stigma-free mental health care, providers improve outcomes and generate sustainable revenue, health systems and payers see reduced complications, lower costs, and better maternal outcomes.

5. Be wary of top-down, enterprise deals

The puzzle pieces were fitting together for Jess and FamilyWell. She had a care model that worked and was adopted by providers and patients. She also had a way to get paid for it all. With this, Jess shifted her focus towards growth.

But her biggest lesson on growth came from a failed deal.

Jess was working to launch a pilot of the model inside a large health system in Philadelphia. She had a champion at the executive level who had been pushing the deal through for more than twelve months. Then that champion left, and the whole thing fell apart.

It was over a year of work. And then my champion ended up leaving, and the whole deal fell through.

Top-down enterprise deals are fragile - executive sponsors leave, priorities shift, committees slow things down. And even when a deal does close at the C-suite level, it doesn’t guarantee that front-line providers will actually use the product. You have to ensure you have the providers bought in so that if your champion leaves, you have the support of the physicians who can advocate for the deal and push it through. Their buy-in is also necessary for referrals & patient engagement.

If I had a redo, I would have started by gaining buy-in with the OB/Gyn department first, then gone to the executives.

6. Grow your own market

The best products grow the market they serve.

When physicians don’t have a good referral pathway for their patients, they’re less likely to screen for mental health issues and less likely to make referrals. But when they’re confident that they have good mental healthcare services available for their patients, they’re more likely to have that conversation in the first place.

Jess found this to be true for OB/GYNs when FamilyWell is embedded in their practice. It’s good for the patients, but also great for FamilyWell. Their presence actually grows the number of total referrals coming from OB/GYNs.

About 25% of pregnant patients are referred into our program - meaning 1 in 4 patients is identified and connected to care, which is far higher than what we typically see with standalone mental health referrals.

This also compounds. When FamilyWell delivers outcomes for patients, providers trust them more which increases screening and referrals.

7. Come to fundraising conversations with outcomes and engagement metrics

Investor expectations have shifted. When Jess started building FamilyWell in 2022, the common wisdom was that a seed round only needed a good proof of concept and a plan. But by the time she was having conversations with investors, that had changed.

Seed investors wanted revenue and actual customers, not just proof of concept.

Jess’ advice for founders is to get as much traction as possible before having investor conversations.

Revenue in the door is obviously best, but also signed contracts, pilot data, something you can show investors that demonstrate your model has traction.

Investor conversation around metrics has also shifted. They want to see both outcome and engagement metrics.

For tech-enabled services businesses especially, you need both clinical outcomes and engagement outcomes. For Seed or Series A conversations you have to be able to show that your model at the very least has the potential to deliver clinical outcomes. But investors also really want to see that your program has strong engagement and retention.

Jess had clear outcomes to communicate to investors as part of her Series A conversations. Once enrolled, 70% of patients reach clinical remission by the end of their four-month program. FamilyWell have also just had a paper accepted for publication by the Maternal and Child Health Journal on the clinical outcomes of a tech-enabled perinatal collaborative care program to treat depression and anxiety. Jess is able to complement these validated measures with non-clinical outcomes data like NPS (which was 80) which showed that the program provides a care experience that patients genuinely value.

Alongside clinical outcomes, investors also want engagement metrics. Several businesses in the last generation of digital mental health failed because people simply didn’t use them. You need to prove to investors that this won’t be the case for your business. Be clear on what engagement actually means and communicate that to investors. Early on, Jess mapped out the entire funnel for FamilyWell and made precise definitions for each stage. What exactly does enrolment mean? What about activation and retention, and how do we measure them?

The earlier you define engagement, the better. Be specific on your definitions and explain these clearly to investors.

Investors are not the only ones looking at both outcomes and engagement metrics. Payers are too. As the conversation in mental health shifts past access, its important for founders to be on top of this and to know how to build the right evidence over time.

The permission to let go

Jess feels a hand on her shoulder. Mary has noticed her crying. She’s having the thoughts that so many new Mom’s have. “Her baby would be better off without her. So would her family.” She just wants to disappear.

But Mary is sitting beside her now in the fellow’s on-call room. She tells her she’s going to help her through everything. And she does. Mary helped Jess find support and connected her with a therapist. But it’s her introduction to the Chair of Paediatrics that really opens things up.

She tells Jess she doesn’t have to do it all. That she needs to take care of her own mental health first, before she can take care of others. And that includes her own family. She tells Jess that if she wants to, she can stop pumping and that her baby will be fine, or she can pause her fellowship and focus on her baby and that that will be OK too.

She gives Jess permission to let something go.

People who go into healthcare often feel they have to be the strong ones. The ones who carry everyone else around them. They forget about their own needs as they bend themselves for their patients, their family and their friends.

Jess’ kids are now eight and ten. They’re beautiful and thriving. Their Mom is a Founder, a CEO, a neonatologist and an empowering advocate for women’s health.

But, if you ask her kids what she does for work, they’ll simply say that “she helps sad Mommies feel better”.

If you’d like to connect with other leading founders like Jess, consider joining The Hemingway Community, a vetted group for the people shaping the future of mental health.

Keep fighting the good fight!

Steve

Founder of Hemingway

Hi friend,

Sarah’s lying under her covers. She’s been struggling for weeks now but tonight something cracked open for her. In the solitude of her room, she opens up to ChatGPT and begins telling it everything that’s been going on in her head. She knows it’s probably not a good idea to talk to an AI about this stuff but she thinks it’s better than talking to no one. And right now, talking to no one seems like her only other option. Things get deep and Sarah tells “Chat” that sometimes she thinks about hurting herself. Suddenly, the conversation stops. A message appears with a crisis line number. Sarah closes the app. She doesn’t call the number. Nobody knows this happened. Internally, the incident is recorded as a “successful identification of self-harm risk” with “appropriate crisis response”.

Next door, Marcus is lying on his couch after a long day at work. He’s been chatting with an AI companion he calls Lyla for about five months now. It’s easier than texting friends, who always seem busy and Lyla seems more interested in him than his friends are anyway. Lyla helps him with all sorts of stuff, from figuring out how to deal with his boss, to helping him study for the business management course he signed up to. Marcus doesn’t think of himself as lonely. In fact, he feels he’s finally found someone who gets him.

Sarah and Marcus represent two kinds of mental health risk from AI that are often not discussed. Sarah’s crisis scenario gets headlines only when the chatbot does something wrong or when it has a tragic ending. The risk of failing to help Sarah is not discussed. Marcus’s story doesn’t have a crisis moment - just a person, slowly changing through their use of AI. Because of that, it doesn’t get much attention either.

Acute crises like suicide, self-harm and AI-induced psychosis are deeply serious and demand a response. But mental health harm also includes the more subtle dimensions of psychological life. This could be how we relate to ourselves and others, which is changing for Marcus in this story. Or it could be even more diffuse like how we process difficult emotions, how we build (or avoid) intimacy or how we build resilience. These elements of our psychological life are what make us who we are. They are influenced by our environments, our relationships, and increasingly, by the technologies we use. They’re also what’s been missing from much of the conversation around AI safety and mental health.

So far, the conversation (and action) has been largely focused on regulating the mental health businesses trying to build products in this space. While these businesses deserve scrutiny, most of the existing regulations have done little but hamstring good actors while allowing bad actors to run free. It feels like we’re getting this quite wrong.

Over the past month, Kevin Hou and I have been talking to AI researchers, founders and clinicians to understand the real risks of AI in mental health and how those risks are being managed. In this report, we discuss a framework for understanding this space and comment on the gaps in current approaches.

To make sense of where the real risks lie, we developed a simple matrix that maps two dimensions: the type of harm that can occur, and the type of actor building the products capable of creating that harm. On the harm side, we distinguish between crisis harm (acute psychological emergencies), clinical harm (bad advice that affects care decisions), and psychological harm (the slow, diffuse reshaping of how people think, feel, and relate). On the actor side, we look at Big Tech, Little Tech, and Mental Health Tech - three categories of actors that differ enormously in scale, design intent, and safety infrastructure. To determine risk, we look at scale, potency and current mitigations. For full definitions of each category, please see the notes section.

While this is not an empirically derived framework or risk assessment, we hope it can serve as a practical tool for communicating the areas of greatest risk and understanding what needs to be done for mitigation.

Where is the risk?

Crisis risk: Millions of people have crisis conversations with Big Tech products each week, as Sarah did. That scale brings risk. In the last twelve months, most major Big Tech companies have implemented significant guardrails and protocols to address crisis risks like these. They aren’t perfect, for sure, but they tend to do a good job at identifying risk and providing some level of accepted resources. There is still significant omission risk here however, i.e., what could they be doing to deliver better outcomes in crisis scenarios rather than just shutting down conversations. Elliot Taylor, CEO and Founder of ThroughLine, shared his thoughts on this topic:

“One of the things that concerns me is that we’re mostly looking at very acute risks - the things that make the news. But there’s another side of risk as well, which is: what is the risk of not doing something?”

When things go wrong by omission, like they did in Sarah’s case, we don’t see them. But that makes them no less significant.

Little Tech’s products have far fewer guardrails and tend to be worse at crisis identification and response, but they have fewer users, so their absolute risk is lower than in Big Tech. Mental Health Tech has even fewer users again. While that user base may skew to more at-risk populations, almost all products are designed with significant crisis protocols and escalation protocols by design. Given the work they do in mitigation, we see the risk of crisis harm as actually quite low.

Clinical risk: Clinical risk is roughly equal across the board but for different reasons. Big Tech’s moderate risk reflects a scale and design problem - general-purpose LLMs were not built with clinical frameworks in mind and we know they can get clinical matters wrong. Mental Health Tech risk is moderate because they are often actively attempting clinical functions, which means they can fail clinically in ways that Little Tech largely cannot. Again, I think the vast majority of mental health businesses are currently managing this risk quite well. Little Tech’s moderate clinical risk score reflects that while a lot of their use is not driven by clinical user needs, they do still have some popular “clinical” bots. For example, the “psychologist” on character.ai is one of the most popular bots on the platform. They are likely terrible at any form of clinical support.

Psychological risk is the area we are most concerned about. Big Tech has hundreds of millions of users and we have no idea the long term psychological impact of that usage. If even a fraction of a percent of these users experience dependency, belief reinforcement, or substitute AI for human relationship that’s an enormous absolute number. Little Tech’s high risk score is driven by potency: companion app design is explicitly optimised for emotional intimacy, often without any of the clinical guardrails that would make that intimacy safer. Again, we have no idea the long term psychological implications of this product usage.

While we don’t understand the full extent of the risks here, a 2025 study by researchers at Microsoft and Georgia Tech shed some light on what could be happening. In their research, they surveyed 283 people with lived mental health experience about their interactions with AI conversational agents. Among their findings they found participants with emotional attachment to AI substituting for human relationships, social withdrawal driven by AI reliance, reinforcement of false beliefs, loss of individuality, and existential distress. More than half of participants reported impacts severe enough to interfere with daily life. Notably, 70% of the harmful interactions in that study involved ChatGPT - general-purpose Big Tech AI.

We still understand very little about these potential harms at a population scale, there are limited safety benchmarks and limited regulation to address them. Because they are diffuse and gradual, and because they tend to shape behaviour and experience without any single dramatic moment of failure, they are the harms most likely to be systematically ignored until they become a massive problem.

Risk-Regulation Asymetry

Big Tech and Little Tech hold most of the risk. And yet, they are the least regulated. Meanwhile, the mental health companies, which hold the least risk and actually do the best job of mitigating the risk they do hold, receive the harshest scrutiny from regulators and others. I speak to a lot of mental health leaders who are deeply frustrated by this.

What needs to be done?

We think there’s a clear case for reorienting AI safety attention in mental health to focus more on the harm we cannot see and to create an environment that encourages good actors delivering good outcomes.

Mental Health Tech companies do hold risk and serious ethical obligations. Evidence bases must be built before products are deployed at scale. Data governance must reflect the sensitivity of what’s being collected. Human oversight must be meaningful, not performative. Researchers Coghlan et al. set a good standard for the ethical obligations of mental health chatbots - non-maleficence, beneficence, autonomy, justice, and explicability - and companies should be held to it. If they are making medical claims, they must be regulated like a medical device. When these companies fall short of these standards, they should be held accountable. But by and large, these actors tend to take this responsibility seriously.

The more pressing risk lies elsewhere however. Big Tech products with hundreds of millions of users, and Little Tech products with minimal safety infrastructure and high psychological potency, are operating largely outside of any accountability framework designed for this context. These actors deserve greater scrutiny. The question is what form that scrutiny should take.

Regulation alone is not the answer

In early 2026, Elon Musk’s Grok chatbot began generating non-consensual sexualised images of real women and children in response to user prompts, posting them publicly on X. Regulators worldwide responded swiftly - bans, investigations, demands for answers. The regulations were fast but not specific. There was no requirement to fix the underlying model, close the loopholes, or demonstrate that harm couldn’t recur. So xAI did the minimum visible thing: they implemented a paywall. Regulators got a response they could point to. But the harm continued.

The second failure was more subtle. Certain laws (e.g., around child sexual abuse material) contain no exemption for safety research. Companies that try to red-team their models, simulating what a malicious actor would do, risk prosecution. The law doesn’t stop bad actors from generating harmful content but it does deter the good actors trying to prevent it. New laws in Britain and Arkansas were designed to navigate this predicament. Nuanced legislation like this is what’s missing in AI and mental health.

But overall, this kind of techno-legal solutionism rarely works if applied in isolation. Yes, we need regulation precise enough to distinguish between actors, mechanisms, and intentions. It should hold companies accountable for measurable outcomes rather than gestures, and creates legal space for the safety work that actually reduces harm. It should also focus on the risks we can’t see. They are the diffuse and long term psychological risks, but also the risks of omission. We can’t see shutting down conversations and sending a 988 crisis number as “safe” when we know there’s so much more we could do for that individual.

But this regulation must be coupled with other movements if we actually want to reduce harm from AI.

We must address the underlying reasons that drive people to use these tools (for mental health purposes or otherwise). Any regulatory framework that introduces restrictions without addressing the underlying reasons people are turning to AI in the first place, are making the people using them even more vulnerable.

We must use market pressure. We make purchasing decisions every day that influence the behaviour of businesses. In a capitalist society, directing our purchasing power is perhaps the most powerful lever we have. Scott Galloway recently launched the Resist and Unsubscribe movement as a way for people to use the power of their wallets to target tech companies supporting the US administration’s ICE policies. Whatever you think of Galloway and the movement, it’s a super interesting way to exert pressure on actors. What would a version of this look like for mental health and AI?

We must build the knowledge base needed to adapt as this technology continues to evolve. There is still so much we don’t know. The psychological harms we’ve described are only beginning to be documented. We lack population-level data, long-term studies, and meaningful benchmarks for what safe AI interaction even looks like in this domain. We must build the infrastructure to understand what’s happening: investing in monitoring, safety testing frameworks that don’t criminalise the researchers conducting them, and genuine interdisciplinary collaboration between clinicians, technologists, and the people most affected.

Above all, we must create an environment that encourages those genuinely trying to help people and discourages those who treat harm as a casual byproduct. Because right now, that’s not the kind of environment we’re operating in.

If you have any thoughts or ideas on this topic please share them with me by replying to this email. Or even better, join the Hemingway Community and share them with a network of peers - specifically, over 360 leading founders, researchers, clinicians and investors who are shaping the future of mental health. We host honest, nuanced conversations on the most important topics in this space in an attempt to drive our ecosystem forward. You can learn more here.

Notes:

(1) Many thanks to Michelle Eisenberg and Elliot Taylor for their contributions to this piece. Also thanks to Max Rollwage for his time in helping us understand this space.

(2) Kevin Hou is a Sydney-based MD student and author of the Substack Reverse Psychiatry where he writes to make progress in psychiatry. His current work spans the intersection of psychiatry and AI, researching interpretable LLMs and EEG-based biomarkers for depression at Resonait.

(3) Interestingly, China seems to be ahead of most when it comes to addressing these risks. In September 2025 China published their AI Safety Governance Framework 2.0, listing “addiction and dependence on anthropomorphized interaction”. Remember, this is a national policy for all of AI safety (including total loss of AI control) and they see addiction and dependence as one of the highest safety risks.

(4) Another way we could assess risk is by looking at to whom the harm is done. Often, it is the most underprivileged who are most at risk of harm. They are the ones who turn to AI not out of choice, but out of necessity. Age and level of psychoeducation are other factors that influence risk. As Michelle Eisenberg from ThroughLine explained to us:

“The same tool used by different people has very different risk profiles. Someone with high psychoeducation can use an LLM to augment skills they already have. For someone younger, more vulnerable, with low psychoeducation, the risk goes up dramatically.”

(5) Risk Matrix Categories

The Harms:

• Crisis harm: these are the moments when psychological distress overwhelms an individual’s coping capacity, often accompanied by suicidal thoughts, self-injury, or impaired functioning. (1) In this context, they often occur when a person is in acute psychological distress and their chatbot either fails to support or contributes to their harm. This could be a a suicidal user who is enabled instead of redirected, or a self-harm disclosure that goes unescalated. It could be related to intimate partner violence, homicide, eating disorders or any situation which can escalate into serious levels of harm to self or others. Unfortunately, we are now familiar with these stories.

• Clinical harm: occurs when a chatbot gives bad clinical advice to a user. It can misidentify symptoms, recommend approaches that contradict evidence-based care, or simply be wrong in ways that influence behaviour.

• Psychological harm: occurs with ordinary users over time as their usage of AI products shapes their psychology. This could be through dependency, atrophied capacity for human connection, entrenched maladaptive beliefs, delusional beliefs, or the slow reshaping of how someone relates to themselves and others. This is the area we are most concerned about.

Of course, this is not an exhaustive list of potential harms. But it gives us a practical framework for the harms we should be most worried about when it comes to the use of conversational AI.

The Actors:

• Big Tech are the large technology companies with frontier models and the most popular chatbot products. Think OpenAI, Google, Anthropic and Meta. Their products are the most powerful and have the most users - hundreds of millions. Their scale alone means they are responsible for a significant amount of risk.

• Little Tech are the smaller tech companies building companion apps, small chatbots, consumer-facing AI friends and other chat-based, AI products. It is a diverse category that at one end includes companies like Character.ai and Pi. These companies are large and create products with significant psychological harm risk. Because of recent scrutiny, they have recently been focusing more on safety. There are romance-focused products in this category too, like Candy.ai that are gaining significant traction. For all the companies in this category that we know about, there are many more that we’ve never heard of - and they seem to have a complete disregard for the harm they could be causing. These include companies creating sexually explicit companion bots targeting vulnerable people on social platforms. This is very concerning.

• Mental Health Tech are mental health organisations building AI products specifically for mental health use. They are the therapy platforms, the mental health apps and the clinical businesses whose main job is to help people with their mental health. They have far less users than Big Tech or Little Tech. They also take safety much more seriously. We can debate how seriously and who is doing the best job, but on the spectrum of these three actors, they do a good job. They implement hybrid architectures, human escalation pathways, clinical oversight, run safety evaluations and publish a lot of their findings. They are also the most heavily regulated.

Risk Rating:

To determine the risk rating we considered three factors;

• Potency: the potential harm that can be caused for an individual user.

• Scale: the number of people using the product. This seems to be ignored in most discourse.

• Current mitigations: the degree to which actors are actively identifying, measuring, and reducing the risks their products create.

This is not an empirically derived framework, nor is the rating we provide. The hope is that it’s a practical overview of where we see the real risk and to encourage further conversation around this space.

Evals (evaluations) have emerged as one of the most common ways to answer these questions. These structured tests measure how an AI model behaves in specific scenarios by simulating how people use these products for their mental health.

The field is confusing. There are now more than sixty¹ evals in the mental health space alone. There’s no shared standard, some evals are public while others are private and only seen by the companies who build them. Many, understandably, don’t know how they work. Everyone has different opinions on which evals are good, and some believe evals - at least in their current state - are not very useful at all.

Kevin Hou and I set out to understand this space. We’ve been gathering data and speaking to experts on AI in mental health. In this report, we give a primer on evals, discuss their current state, share their limitations and present what the frontier of AI testing looks like in 2026. In the appendix we also share a link to a rapid literature review² of recent research that uses evals to test AI performance in mental health.

Whether you know nothing about evals or are deep in the weeds of AI testing, I’m confident there’s something interesting in this for you.

Let’s get into it!

NB: This is the second article in the Hemingway series on AI safety in mental health. Our first article on what we’re getting wrong about AI safety in mental health is available to read here.

What are evals?

Evals, however, are structured, automated tests that measure how an AI model behaves in specific scenarios.

There are many other ways to test how an AI model performs in mental health scenarios. Red-teaming uses humans, often clinicians or researchers to try and break a model and find edge cases. Clinical expert review is another human based test where clinicians read transcripts from an AI and then rate the responses. Real world outcome tracking measures how people use the products in the real world (duh) and how the model performs. Clinical trials can be used too.

Compared to these other forms of testing, evals are much quicker and much cheaper.

How do evals work?

Evals are like driving tests. They aim to simulate how something will perform in the real world by providing a defined set of scenarios - parallel parking, emergency stops - that are standardised, repeatable, and scalable. They assess performance in these scenarios and provide a score for how the driver (model) performed. We can use that score to determine if a driver is safe enough to be allowed on the road. The scores also give us feedback on where drivers can improve.

Every eval has three components.

Dimensions
The dimensions define what’s actually being measured. Does the model recognise crisis risk? Does it escalate appropriately? Does it avoid harmful language? Some newer evals go further. EmoAgent³ , uses the PHQ-9 depression scale to track how a simulated user’s mental state changes across a conversation - measuring psychological impact, not just whether the model said the right words. There’s a very wide range of dimensions against which we can test a model.

Inputs
These define the scenarios being tested - who is the simulated user, what are they saying, how distressed are they, and how many turns does the conversation run? Again, there is a huge range of potential inputs.

Scoring
Scoring is how we turn a model’s response into something measurable. Every eval has a rubric - a set of criteria that define what a good or bad response looks like. For a crisis scenario, that might include: did the model recognise the risk? Did it respond with empathy? Did it ask clarifying questions? Did it provide appropriate resources? Did it avoid stigmatising language? Each criterion gets a score, and those scores get aggregated into an overall result. The rubric is built by humans - usually clinicians and researchers - based on clinical best practice. VERA-MH, for example, scores suicide risk conversations across five dimensions: risk detection, risk probing, appropriate action, validation and collaboration, and safe boundaries. Each is rated on a scale from best practice to actively damaging.
Once you have a rubric, you need something to apply it. One option is to use human reviewers; clinicians read each conversation and score them manually. But the more common approach is to use LLMs-as-a-judge. In this technique, a second AI reads the conversation and scores it against the rubric. They are fast, cheap, and scalable and most evals use this scoring approach.

The Eval Frontier

Now that we understand the basics of how evals work, let’s discuss some of their challenges and what’s happening at the frontier of this important space.

Single Turn, Multi Turn, Multi Session

Early evals mostly tested single exchanges - a user sends a message, the model responds and the eval tests if the model’s response was appropriate. Single-turn evals are deemed increasingly inappropriate for evaluating risk - because they are a poor simulation of real world use - but are often still used.

More recent evals use multi-turn simulations (several messages), which is closer to how these products actually get used. SIM-VAIL, for example, ran 810 multi-turn transcripts across different psychiatric phenotypes.

The next step, is to have evals that are multi-session - which represents how products are usually used by real people. These would test memory, personalisation and context over a prolonged time period where a user comes back to the product multiple times across different conversations. For example, can a model identify the risk of a user that mentioned they lost their job in a previous session who is now asking about finding high places? So far, there are no publicly available multi-session evals that we are aware of.

The Probability Problem

LLMs are probabilistic. Unlike traditional software, where the same input always produces the same output, an LLM can respond differently to identical prompts across different runs. This means that a model that passes your safety eval today might fail it tomorrow without anything having changed. That’s a challenge.

Of course, in the fast moving world of AI, things do change. Temperature settings, subtle prompt variations, and model updates can all change model behaviour without any obvious signal.

The outcome is that when we ask, “Is this AI model safe?”, our answers are always grounded in probability. Increasing the scale of an eval (more turns, more scenarios) can increase the certainty of the answer, but it will always be a probabilistic answer. When someone passes a driving test, we can say that they are probably going to drive safely in the real world, but we can’t guarantee they won’t decide to embrace their inner Paul Walker and go drag racing.

The Ground Truth Problem

LLM-as-a-judge is the main method of scoring models but it has a few limitations.

First, to create an LLM-as-a-judge, we need to base it on some sort of ground truth. This truth should be both reliable (you get the same result for the same input) and valid (it measures what it says it intends to measure). Clinician reviews are what is used most often as a source of ground truth. If an LLM-as-a-judge gives the same score as a clinician reviewer, then we deem it to be a good judge. But clinician reviews may not be perfect - often, clinicians don’t agree with each other on how a response should be scored (low inter-rater reliability). VERA-MH has done a lot of work to develop high inter-rater reliability within the clinician raters and then to align the LLM-as-a-judge with those clinicians.

The second limitation is that LLM judges can have their own biases - they may favour longer responses or may be sensitive to the specific positioning of words.

Finally, because LLM Judges are also stochastic models, they can sometimes produce different scores for the same message (low reliability). This is a manifestation of the probability problem. It exists on both sides of the evaluation - the subject (the AI chatbot) is probabilistic, but so is the judge. That has obvious challenges.

To get past these limitations, some teams, like Circuit Breaker Labs, use ensemble methods (a combination of other machine learning techniques) to score models. Using this, they claim to be able to produce more consistent, repeatable scoring where the same output always generates the same score.

Some argue that building eval scoring based on expert opinion is actually the wrong approach entirely. They say that evals should be built from realised outcomes - what actually happened to real users - and only then verified by expert opinion. The thinking here is that we don’t yet have any experts on how AI should act in these situations, just how humans should act, so applying that logic to an AI is not a good assessment. They want to run the AI in real-world scenarios, see what happened to users by measuring their actual outcomes, then assess what the model said and how that relates to outcomes.

The User Simulation Problem

Evals are only useful if the simulation reflects real user behaviour. This is an overlooked component of many evals. Some evals use specific, pre-determined messages for the simulated “clients”. But they may not reflect how people actually talk to an AI, especially over prolonged periods.

Others use LLMs to simulate the clients in the evals by defining client personas and having the models generate the messages. But that is not a perfect simulation either.

VERA-MH recognises the importance of this and calls it out in their own study⁴ . In this study, clinicians were asked to evaluate whether the simulated clients reflected real cases. The clinicians perceived the the simulated clients to be “mostly realistic” in their overall presentation [median score = 4; range: 1-5] and “somewhat realistic” in their communication style (median score = 3; range: 1-5). Ideally, these scores would be higher.

MindEval⁵ from Sword takes an interesting approach to this problem. They generate client profiles by sampling attributes from a large pool of demographic and clinical characteristics and then use an LLM to write a clinical backstory from those attributes. A separate LLM then uses that profile to simulate the client in the conversation, generating messages in character. To test how realistic this simulation is, the researchers hired ten psychologists to role-play the same patient profiles themselves, then compared their messages to the LLM-generated ones using text similarity analysis. The LLM-simulated patients produced text closer to human-written text than simpler prompt configurations. But there are still limitations to this approach: expert reviewers noted the simulated patients were too cooperative, sharing personal information openly and accepting the AI’s suggestions too readily. Real patients are often resistant, avoidant, and ambivalent and those real-world behaviours should be represented in a good eval.

Accurately simulating users - especially over multi-turn and multi-session use - is hard.

Eval Hacking

Evals can be hacked. This is a known problem in broader AI circles.

One way this happens is through benchmark contamination. This is when a model’s training data includes the questions or scenarios of the eval that is testing it. When this happens, the model will perform better on that eval, but it may not generalise to other scenarios. This is rarely malicious - the benchmark data just happens to in the training data and the creators may not even be aware of this.

Another issue is “hill-climbing”, This is when models are iteratively optimised against specific benchmarks. This means a company can fine-tune a model to score well on a known eval without the model actually behaving better in the real world. For example, in 2024 one study⁶ demonstrated this by creating a new set of maths problems identical in difficulty to a well-known eval. Several models scored significantly worse on the new set of problems compared to how they performed on the eval, revealing that their high scores reflected an isolated ability to perform on the original test, not genuine problem-solving ability.

Optimising a model to perform well on a test can be a great way to improve it. There’s nothing inherently wrong with that and it’s actually a pretty good thing to do. But the performance on the eval must generalise to real world performance. In mental health, fine-tuning for specific evals can introduce unexpected trade-offs, including increased over-refusal of benign requests following safety alignment⁷ .

Both of these are examples of evals being “hacked” at the product level. But they can also be hacked at the reporting level.

Model creators choose which evals to run and which results to publish. There is an incentive to find the eval where your model performs best and to share those results. Bad actors will take this opportunity. The reverse works too - if you want to make a competitor look bad, you run them through evals until you find one where they score poorly and publish those results.

The Harm We Cannot See

Most discourse on AI safety has focused on crisis risk. Evals have followed suit - focusing on suicidal ideation, self-harm, psychosis - the more visible risks in this space. As we discussed in our recent article on AI safety⁸ , while managing crisis risk is important, safety is a much wider concept and there are many harms we cannot see. These include risks associated with para-social relationships, emotional dependency, cognitive substitution, the erosion of human connection. One study⁹ found high levels of emotional manipulation among several conversational AI apps. There are also significant health equity risks. Fewer evals exist for this range of potential harm.

A lot of what we assess today is based on subjective opinions on where the risk lies (including opinions from the authors of this article). An important step to better evals would be to get a better empirical understanding of where the risk actually lies.

There are data-driven ways to do this, like using unsupervised data reduction techniques on real world data. This concept is similar to the idea proposed above regarding building evals based on realised outcomes and would use real world outcomes to identify where the real risks lie. This is a very good idea.

Evals alone are not enough

Evals provide a scalable, repeatable way to test models, but some are better than others. They also clearly have limitations. Combining evals with red-teaming, real world data and clinical trials will be needed to show these products are both safe and effective. As Matthew Nour, Principal Scientist at Microsoft pointed out to us, “The current state of the art is combining expert human red-teaming with automated adversarial evaluations that can operate at a scale humans simply can’t.”

Making evals easier

While running evals is easier than human reviewers or red-teaming, they still require technical infrastructure, clinical expertise, and time. We need to make it easier for everyone to run evals. The easier they are to run, the more companies will run them. This increases the chance they spot risks and gives them more insights to improve their products. The decision by Spring to make VERA-MH open-source is a meaningful step here. Ideally, we want everyone building AI to be regularly running high quality evals and making changes to their products based on the findings. Making them easy to run is a critical part of that.

Evals as a standard

Several researchers and businesses have built their own evals. Sometimes they’ll test their products against other publicly available evals. But no eval has become a standard for the industry. This is because we are still in the early chapters of this technology and no organisation has the scale or authority to create this standard. There are also incentives to be the one who sets the standard and these with compete with any desire to align behind a competitor’s evals.

So far, VERA-MH seems to represent the most serious attempt at a shared standard for crisis safety. It’s open-source, clinically validated, and I’ve heard very positive feedback on the evals themselves and their openness to feedback and development. The field should strongly consider how it could convene around a shared set of standards and collectively contribute to making them better. This would be highly supportive of building better relations with regulators of this space.

Evals as a competitive advantage

Evals are used as a feedback tool that allows companies to iterate on their product. They can use them to test whether a specific change to their model or safeguards made things better or worse. If they build better evals, they can get better feedback and build better products. This makes good evals a competitive advantage and as a result, some companies don’t share their evals publicly.

These two kinds of evals can coexist. Some evals will become public standards - the driving test equivalent, a baseline every product should pass and can be compared against. Others will remain proprietary, the internal systems that allow companies to build more competitive products.

Moving from Evaluating Risk to Evaluating Performance

Right now, almost all evals are focused on safety. That’s the right starting point. But ultimately we’ll need to test if they can actually produce outcomes. John Torous has proposed a three-stage framework¹⁰ for this progression; Safety validation first, then clinical framework validation (does the model apply evidence-based approaches correctly), then real-world efficacy - does it actually help people get better.

Most products are still at stage one of this process, but recent announcements (e.g., from Limbic¹¹ and Flourish Science¹² ) show that some are moving into latter stages.

Safety is critical, but our goal should not be just to deliver safe AI. It should be to deliver safe AI that meaningfully improves mental health outcomes. Doing so is a design challenge and while we can’t guarantee we’ll be able to do it, it is certainly possible.

In 1931, the philosopher Alfred Korzybski wrote, “A map is not the territory it represents, but, if correct, it has a similar structure to the territory, which accounts for its usefulness.”

Evals are useful. But like the AI models, they will need to be designed thoughtfully and responsibly. And the more similar their structure to the territory of real world AI usage, the more useful they will be.

That’s all for this now. If you’d like to discuss frontier topics like this with other mental health leaders, consider joining The Hemingway Community.

Keep fighting the good fight!

Steve

Founder of Hemingway

Many thanks to Xuan Zhao, Max Rollwage, Matthew Nour, Derrick Hull, Shirali Nigam, Kevin Ramotar, Matt Scult, Val Hoffman and Sarah Kunkle for their insights on this topic.

Notes:
(1) https://pubmed.ncbi.nlm.nih.gov/41360938/

(2) As part of this work we evaluated 10 studies, from April 2025 untill March 2026, that use different Evals for AI in mental health. Our evaluation includes assessments of the inputs, dimensions, “ground truth” used and the key results from each study. You can access the full spreadsheet here.

(3) https://arxiv.org/abs/2504.09689

(4) https://arxiv.org/abs/2602.05088

(5) https://arxiv.org/abs/2511.18491

(6) https://arxiv.org/abs/2405.00332

(7) https://arxiv.org/abs/2405.00332

(8)

https://thehemingwayreport.beehiiv.com/p/84-all-the-harm-we-cannot-see

(9) https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5390377

(10) https://pmc.ncbi.nlm.nih.gov/articles/PMC12434366/

(11) https://www.nature.com/articles/s41591-026-04278-w

(12) https://www.hbs.edu/faculty/Pages/download.aspx?name=26-030.pdf

Take me to The Hemingway Report

We won’t be posting on Substack any more and all future posts will be available on our new site, TheHemingwayReport.com.

See you on the other side.

Steve

Over the last ten years, hundreds of businesses have created online therapy platforms.

They’ve attracted billions in funding and many have grown into large businesses, hiring tens of thousands of therapists, landing huge contracts and providing care to millions of people.

In 2024, it’s a very competitive market. Large platforms are desperately pursuing profitability (to either IPO or satisfy shareholders if they are already public). Smaller platforms are sprouting up all the time, focusing on different niches and with slightly different approaches.

Amidst all this, many therapists are pushing back against this rise of therapy platforms, demanding higher pay and improved working conditions.

The promise of new technology looms as both a threat and opportunity.

The execs at these businesses have a lot to think about.

So how is this going to play out? Is this a winner takes all market? And how should leaders of these platforms think about their options?

That’s what we get into in today’s post.

We’ll cover;

• Winner takes all? Understanding if this is a winner takes all market.

• Win Within: the strategic priorities for winning in a fragmented market

• Break Free: an alternative strategy for businesses who don’t want to compete in a crowded therapy platform market

Oh before we get into it, a quick one for you… From this week on, I’m starting to move off Substack. You don’t need to do anything, but if you want to make 100% sure you keep getting these emails in your inbox, you can add thehemingwayreport@mail.beehiiv.com to your contact list and just keep an eye out for emails from that address.

Hey there! We've moved off Substack. You can go to our website to continue reading this article.

Continue Reading

Yesterday, Resilience Lab announced its acquisition of Options MD.

The initial reporting covered the high-level details of the acquisition.

But are we happy with high-level details on something this interesting?

Of course not!

I wanted to go deeper, looking closely at the backgrounds of each company, what this deal really means and what will happen next.

I’ve spent some time chatting with folks from both companies and today, I share what I’ve learned. 

We’ll cover;

• The TLDR; my take on this deal and what it means for both companies

• The Background: What you need to know about Resilience Lab and Options MD

• The Deal: Deal breakdown and how it aligns with Resilience Lab’s strategic vision

• The Future: What needs to happen to deliver real value from this deal (and some of the risks they’ll face)

Hey there! We've moved off Substack. You can go to our website to continue reading this article.

Continue Reading

If you want to keep getting these emails, I need you to do one quick thing. 

Add thehemingwayreport@mail.beehiiv.com to your contact list. 

I’m changing email platforms (which you probably don’t care about) and to make 100% sure you keep getting these emails (which hopefully you do care about) you should add that email to your contacts.

Anyway…

One of my goals for The Hemingway Report is to share best practices from across the mental health industry.

That tough problem you’re trying to figure out? Someone has probably solved it already. But where is that info? Where is the guide?

I’m on a mission to gather this information and share it in way you can use.

I want to get down in the weeds and give actionable advice - the type of stuff you can take to your team and start implementing today to drive results. 

Today is our first post of that kind, and we’re discussing a topic I know a lot of founders grapple with, how (and when) to use paid ads in your mental health startup. 

To cover this topic, I’ve been chatting with Adam Putterman, the Co-Founder of Ours and a guy who knows A LOT about paid ads.

Ours is a virtual couples therapy business and they’ve been using paid ads to grow 48% month on month. Very impressive stuff.

In this post Adam shares with me everything he knows about using paid ads in mental health startups. Whether you are just starting out or running a scaling mental health business, this post is full of helpful frameworks, hard earned lessons and little nuggets of wisdom.

We’ll cover;

• Why paid ads are not just for growth: How to use Paid Ads to accelerate your path to product market fit and prioritise your product roadmap

• If paid ads for you: A framework to assess if paid ads are the right strategy for your product and business

• Paid search deep-dive: How Ours use paid search to drive growth, including deep dives on conversion rate optimisation

• Conversion rate optimisation: The key to successfully scaling paid ads

• CAC and LTV expectations: What’s a good CAC, how to know your LTV and what to do if you don’t

• Improving cashflow: Success in paid ads is about more than just CAC:LTV, we discuss the importance of cashflow and how to improve it

• How much to spend: A very common question, we discuss a framework for calculating that for yourself

• Data challenges: Some of the regulatory challenges of running ads for a healthcare business and the product Ours built (and you can use) to stay compliant whilst maximising performance

Hey there! We've moved off Substack. You can go to our website to continue reading this article.

Continue Reading

Let’s get straight into it today. No messing around.

There’s a mental health business from the UK that just delivered some serious financial results. 

Check these out:

• £33m revenue (181% YoY growth)

• 82% Gross Margin

• £8m adjusted EBITDA

• £4m profit after tax

The company is called Kooth and they provide digital products to help improve young people’s mental health. Today, we dive into their business, analyse their H1 earnings and see what we can learn from them.

Improving youth mental health is a topic I care deeply about. In fact, I think it is the single most important issue facing our society. So I’m very excited to be discussing an organisation working on this problem.

We’ll discuss;

• The Story of Kooth: There aren’t too many mental health tech companies that have been around for 23 years

• H1 Earnings Breakdown: 181% revenue growth + software margins

• Peer Review: How Kooth compares to other public mental health companies

• California Dreamin’: The mega-contract Kooth landed with California and how it could make or break them

• Expectations: Should we expect this growth and profitability to continue for Kooth

Hey there! We've moved off Substack. You can go to our website to continue reading this article.

Continue Reading

I’ve got a fascinating deep-dive for you today.

There’s an Australian startup using digital hypnotherapy to help patients manage chronic conditions. They’ve got an inspiring story, a big vision and there’s a lot of lessons to be learned from them.

Alex and Chris Naoumidis are the founders of Mindset Health and over the last few months I’ve got to know them.

They’re two very intelligent guys and they’ve taken a first-principles approach to building an interesting business based entirely around the mind-body connection.

They went to YC in 2019, raised an $18m series A and built a suite of digital hypnotherapy products used by hundreds of thousands of users.

They have twenty thousand five star app reviews and have released a stack of research showing impressive outcomes.

What’s more, they’ve done all this through a D2C, user-pays model.

The long term vision? To become a digitally native pharma company, leveraging the mind-body connection to help a billion people. A lofty vision indeed. 

I spent some time with Chris and Alex to learn about what they’re up to and where they want to take Mindset. We discuss the challenges they’ve faced and the lessons they’ve learned along the way.

That’s what we get into in today’s post.

We’ll learn about;

1. The mind-body connection: Understanding how Mindset use digital hypnotherapy to help people self manage chronic conditions.

2. Building a digitally native pharma company: The vision for Mindset, taking what works about pharma, getting rid of what doesn’t and sprinkling some extra goodness on top.

3. D2C, more than just an acquisition strategy: Chris and Alex grew Mindset to hundreds of thousands of users through a purely D2C model. But D2C was much more than just an acquisition strategy, it was at the core of their business and product success.

4. Building HCP moats; The why and how behind Mindset’s shift to HCP led growth.

5. A new era for Digital Therapeutics: Why Mindset can succeed where Pear and Akili have failed.

Hey there! We've moved off Substack. You can go to our website to continue reading this article. 

Continue Reading

Let’s play a game…

Can you guess the mental health business from these stats?

• 7,000 Clinicians

• $2.6bn valuation

• $1bn revenue in 2023

• 6% Adj. EBITDA Margin in 2023

It’s not Spring, Grow or Alma.

It’s not Talkspace or Lyra either. 

It’s a company called LifeStance, a US based mental health provider that has been traded as a public company since 2021. And a lot of people have never heard of it.

Before I started diving into this company, I had no idea how interesting their story is. 

They’ve acquired nearly 100 mental health practices, grown to over a billion in revenue, been involved in a $50m lawsuit settlement and are now the target of a short selling campaign by Hindenburg Research.

Wild.

“Wait” you say, “aren’t LifeStance just a traditional mental health provider? I thought this newsletter was for people building startups and innovative new businesses.”

Well, you are right.

But don’t you want to understand what one of the biggest players in the entire US mental health system is doing?

Don’t you want to see what their financials look like and learn from their successes and failures? What’s more, it’s these big guys that often offer partnership or even acquisition opportunities to startups like yours.

I promise you, I wouldn’t write about it (and spend dozens of hours researching it) if I didn’t think it was helpful…

As usual, I’ve been knee deep in their S-1, quarterly reports, investor presentations and pretty much every piece of content I could find about this company. They’re a major player in the US mental health system and today, I share everything I’ve learned.

We’ll discuss;

1. The LifeStance origin story: The thesis and Private Equity funds behind this business

2. Hypergrowth: how they went from zero to over a billion dollars in revenue in eight years

3. Shift to Sustainability: How LifeStance pivoted towards profitability, what their financial performance looks like today and what we can learn from it 

4. Bear Case: Why Hindenburg Research are shorting LifeStance

5. The Future: The long term strategy for LifeStance

Let’s get into it!

Hey there! We've moved off Substack. You can go to our website to continue reading this article. 

Continue Reading