Subscribe now

We just got back from a Clarify off-site, and every time we do one, I can’t help but think about how expensive they are (logistically and financially). Flights, hotels, food, and a week where everyone’s normal schedule gets rearranged add up quickly.

But Clarify is a remote-first company, and I’ve come to consider off-sites less as a perk and more as part of the cost of operating remotely. If we’re going to give up the office, we have to replace some of what happens there naturally.

Remote work changes how relationships get built

When people work together in an office, a lot of relationship building happens naturally. You grab lunch, have brief conversations after meetings, or grab a beer after work. Over time, you get to know people in ways that are hard to recreate through scheduled Zoom calls.

Remote teams lose a lot of those small moments, and I think that starts to matter most when people disagree.

If I’ve spent time with somebody and gotten to know them personally, I’m more likely to give them some grace when a conflict comes up. I have more context for how they think and where they’re coming from, so there’s usually more trust to fall back on when something gets difficult.

A remote company has to create those opportunities on purpose, instead of relying on the spontaneous interactions you get in person.

At Clarify, we try to get together a couple of times a year, sometimes a little more often. I’ve experimented with more or less than that here and at other companies, but this is what works for us.

If I were starting a smaller remote company today, I’d probably try to get everybody together for a week every quarter while the team was still small enough to make that practical. A week feels about right to me. Much shorter and it feels like you’re leaving just as you hit your stride. Much longer and people start wanting to get back to their normal lives.

I expect us to work

I have a pretty strong opinion about this: an off-site is a work trip.

We have great dinners, hang out, and try to pick places people actually want to visit. But the main reason we’re there is to work together, and I want us to take advantage of having everyone in the same place.

One of the biggest parts of our off-sites is the hackathon. It gives people time to explore builds they’re curious about, work with teammates they haven’t spent much time with, and take ideas much further than they could during a normal week.

At our last one, I worked on voice agents while other people explored things like memory and campaigns. Some projects make it into the product, while others might help us decide an idea isn’t as interesting or useful as we thought. Both are good outcomes.

The point isn’t that every hackathon project has to ship. It’s giving people room to stretch outside their normal day-to-day work and take an idea far enough that the rest of the company can react to it.

We also use these weeks to spend more time on company strategy, vision, and where we’re headed. Some of those conversations simply work better when everyone has a few days together and enough time to keep talking after the meeting ends. You can spend a few hours debating something, go to dinner, keep talking about it, and come back the next morning with a different perspective.

That’s hard to schedule into a 30-minute Zoom.

The ROI is hard to put in a spreadsheet

I can’t tell you exactly what an off-site will cost you. I can’t tell you exactly what you’ll get back from it.

There isn’t a clean calculation where I can measure flights, hotels, and dinners against some amount of additional revenue. But I can see the difference in how the team functions right when we get back from an offsite versus when we haven’t seen each other for a while.

I also don’t think there’s a universal answer here.

For example, if I were running an in-person company, I’d probably put more of this money into a great office, regular team dinners, and smaller activities throughout the month because you already have more opportunities to build relationships.

For Clarify, we chose to be remote. That gives us a lot of advantages, but it also means we have to be deliberate about replacing some of the things we gave up.

I’ve found that spending a week working, building, eating, and talking together makes the months afterward easier.

It’s expensive, but I think it’s worth it.

Until next week,

Patrick

Additional reads

How to run ridiculously good offsite

It was a great week, and I’m excited about this week’s newsletter. If you were forwarded this, join 412 founders getting their weekly dose of Founder Therapy by subscribing below:

Subscribe now

Let’s dive in.

—-

Before Elaine Zelby started Tofu (Agentic demand gen platform), she kept a spreadsheet with more than 450 business ideas.

Every week she’d write about a few of them in her newsletter. She’d think through why the timing made sense, how she’d build the company, and how she’d make money.

Then she’d ask herself one question:

“Do I actually want to build this company?”

Almost every time, the answer was no.

I think about that question a lot, and I believe it’s the hardest question to answer as an aspiring founder. Because you are going to bleed for that company.

Founder-market fit is personal

I ask every founder I talk to why they’re the right person to build what they’re building. The answer is different every time. But the best answers have something in common: the founder already knew the customer before they started. And they had a real reason—not a theoretical one—for wanting to keep working on the problem.

Building a company takes years. You’re going to spend thousands of hours with customers, make thousands of product decisions, and spend a lot of time thinking about the same problem.

I think it’s worth asking whether you’re the person who wants to keep coming back to that problem.

That is what I mean by founder-market fit: knowing the customer and the problem deeply, and having a reason you want to spend years working on it. Elaine’s story is a good example of that.

Godard Abel, Co-founder of G2, also talked about this in his episode of First Pitch, saying “Founder-market fit comes before product-market fit.”

Godard’s team had lived the pain of enterprise software discovery firsthand at BigMachines, and that personal credibility is what got Chicago Ventures to bet on an unproven model when Silicon Valley VCs wouldn’t.

She spent years working in go-to-market. She built marketing systems herself, invested in AI and SaaS companies, and talked with founders every day while she was a venture investor.

When she started thinking about Tofu, she already understood the customers, the pain they were dealing with, and the people she wanted to build for.

She also believed that AI had reached a point where it could unlock a much bigger opportunity in marketing.

In other words, for the first time it felt like she had founder-market fit.

She joked that she kept getting pitched companies where her reaction was, “I could go build this 10 times better.” Eventually she decided she had to stop saying it and go build the company herself.

When something you’ve built needs to change

A few years into building Tofu, Elaine’s team had already invested a lot into the product and its UI. Then she started seeing a change in how people wanted to work.

Her team made a pretty big decision: they took the product they’d built and, as Elaine described it, “cut the head off and put it in Slack.”

The functionality stayed. Customers could still run campaigns. But moving into Slack changed how people used it. Sales teams started joining marketing to select accounts. Others jumped in to review campaigns. The product became more collaborative.

That’s not an easy thing to do. You spend years building something, then change the thing you built. But Elaine talks to CMOs constantly. When the way marketing teams work starts shifting, she shifts Tofu with them.

I think founders naturally get attached to things we’ve spent a lot of time building. Elaine and her team had invested a lot in their UI, but they were willing to change it when they saw customer behavior changing.

To me, that’s part of knowing your market deeply. Elaine talks to CMOs constantly and pays attention to how marketing teams are working. When those behaviors change, she’s willing to change Tofu with them.

Founder-market fit doesn’t stop mattering after you start the company. It shows up years later - especially when something you’ve already built needs to change.

If you’re an aspiring founder…

Elaine’s spreadsheet wasn’t necessarily just a list of ideas. It was a way of figuring out which problems she wanted to spend years working on.

She didn’t force herself into an idea. She kept asking the question until she found one she couldn’t stop thinking about.

So if you’re considering starting a company, try framing the opportunity around founder-market fit.

Until next week,
Patrick

P.S. for the full podcast episode, watch below

Subscribe now

I’ve sold a company before, have been involved in the due diligence of acquisitions, and invested in more than 40 startups, but, until this year, I’d never been the exec sponsor acquiring a company.

Clarify’s acquisition of Seam AI was a first for me.

I wanted to share how it happened because the announcement only covered a small part of the story.

Nick Scavone and I had known each other for four years before the acquisition became a serious conversation, so there is some backstory here.

We’d known each other for four years

Nick and I met when Seam was still early.

The company changed a lot over those four years. Seam started as a product that let people ask questions about their company data using plain English.

The idea came from a problem he’d experienced working at Okta. Nick believed business teams should be able to work with their own data without waiting on an analyst or engineer.

It was a real problem to solve, but the market was pulling them elsewhere.

Nick admits he held onto the first product longer than he should have. He talks openly about this in our conversation.

He believed the market would catch up, but meanwhile customer conversations pushed Seam toward a more focused product. The team began helping sales and marketing teams find the right accounts, understand what was happening across them, and decide where to focus.

In other words, the market molded Seam into an AI-native ABM platform and that’s where they found signs of product-market fit. They generated $1 million in pipeline during a single month.

I watched a lot of the journey from a distance, but we stayed close, shared notes, and got to know how each other worked.

Why the fit made sense

Clarify and Seam were working on connected parts of the same customer problem.

We’re building an AI-native CRM that understands and structures a company’s relationships, emails, meetings, deals, and history.

Seam was helping teams understand which companies deserved their attention based on fit, buying interest, website activity, hiring, funding, and other changes happening around an account.

In other words, we had the internal context. Seam had market intelligence.

Bringing those products together meant the CRM could understand more of the customer story.

A sales team could see what had happened with an account, what was changing around it, and what to act on next. The information could live alongside the relationship instead of being spread across several tools.

As Nick and I spent more time talking through the products, we saw several places where the teams and technology fit together well.

We also knew each other well enough to talk honestly about the difficult parts. We could discuss the team, the product, the customers, and what joining Clarify would mean for everyone who had helped build Seam.

How Nick saw the choice

Nick had the option to pursue Series A and continue building Seam as an independent company.

He had customers, a team, and a product that was working. He’d spent years building the company alongside his co-founders and employees, so the decision affected a lot of people he cared about.

Joining Clarify gave their work a place inside the CRM itself. It also gave Nick the chance to keep working on the same customer problem with a team he already knew.

Nick decided to join Clarify, and I’m grateful he trusted us with the company he and his team had spent years building.

What I’m taking from it

The four years before the acquisition made the process much easier to navigate. We’d had time to watch each other build. We’d seen how each other handled setbacks, customers, and team decisions.

Acquiring a company also gave me a stronger sense of responsibility for all the work that came before the deal. Nick, his co-founders, and the Seam team spent years building the product, supporting customers, and learning the market.

I want to respect that work as we bring the teams together and have 1+1=6.

If you want to be the first to try what we build, join the waitlist here.

Until week,
Patrick

Additional reads:

1. The formal announcement

2. Basics of M&A for founders

And if you’re not subscribed to Founder Therapy yet, subscribe below to join 398 other founders getting this newsletter weekly.

Subscribe now

Let’s dive in.

Fundraising advice has a short shelf life right now. Depending on the month, valuations are up or down, capital feels available or suddenly tight, and advice that was right six weeks ago might already be wrong.

I keep coming back to the same thing: build a real business. Solve a problem people care about, make sure the business can support the product you’re selling, and spend most of your energy on execution because the market will shift before you can react to it.

But the mechanics of fundraising have genuinely changed in the AI era. I asked Sabrina Wu (Partner at Madrona) and Clayton Petty (Partner at Gradient) to share what they’re actually looking at when picking investments today. Their answers were more specific than the usual “the bar has gotten higher” takes.

The ARR problem

Sabrina’s point is one I think a lot of founders are missing: ARR doesn’t mean what it used to.

In the SaaS era, ARR implied durability. Real procurement cycles, genuine annual commitments, customers who had done a full evaluation. Today, a lot of what gets called ARR is MRR times twelve, often off trials or paid pilots. Consumption and outcomes-based pricing adds another layer. Revenue that can spike fast when a use case works and disappear just as fast when it doesn’t.

That same dynamic inflates growth rates. There’s a high willingness right now to experiment with AI products and pay small dollars to try them. Going from zero to $5M is easier than it’s ever been. The harder question is what’s actually behind the number. What’s live versus committed? What comes from customers who’ve integrated the product into how they work versus customers who are still running a test?

The margin compression trade-off

Clayton is seeing more demand for AI products than at any prior point. Usage-based pricing lets customers scale fast once a use case proves out. CEOs and IT teams are actively mandating AI adoption. The strongest companies are offering something genuinely multi-layered: revenue impact, labor reduction, real operational change, not just a new workflow tool.

That growth often comes with compressed margins. A lot of early AI companies run at negative or near-zero gross margins because compute and inference costs are real.

In general, margins are a lot tighter for AI product builders than the era of traditional SaaS.

Credit: The SaaS CFO

Clayton’s view is that this is fine when there’s a credible path to 70-plus percent over time. What he’s looking for is whether the team understands the cost structure and has a realistic plan to get there as they scale. What he’s not willing to underwrite is a founder who can’t explain it.

Sabrina takes the same position. Early margins can be rough. What matters is whether the team knows why and has a plan.

What both of them look for in founders

Sabrina is weighting AI nativeness more than she used to. Not a research background, though that helps. How deeply a founder actually uses AI in their own work. Are they building the company with AI at the core? How much are they actually pushing the models? This is less about credentials and more about how they operate day to day.

Clayton is focused on customer obsession in a more operational sense. He wants founders willing to go on-site, build forward-deployed engineering teams, and think about helping a customer change how their whole business operates, not just implement a single feature. That level of commitment is hard to fake.

Where I land

When it comes to choosing whether or not to raise, I like to use the red pill and blue pill analogy from The Matrix. You either bootstrap and grow from revenue, or you raise and use capital to move faster. Both are real choices. Both create different kinds of companies.

Clarify needed capital. CRM is a deeply competitive market with incumbents that have years of product behind them. The cash gave us time when the market shifted in ways we couldn’t have anticipated. We kept building through a stretch where our metrics alone wouldn’t have supported a new round.

For me the conviction question comes before the metrics question. Am I willing to spend the next 10 or 15 years building this? Once I have that, the capital decision gets simpler: does it get us somewhere meaningfully faster?

Having cash in the bank gives you options. It lets you build through a rough market and make decisions from a position of strength rather than pressure. What I take from Sabrina and Clayton is that the metrics investors use to evaluate that have gotten harder to read, not easier.

It’s the wild west out there right now. Like Y Combinator preaches, focus on building something people want. The fundraising conversation becomes simpler when you do that.

Until next week,
Patrick, with thanks to Sabrina and Clayton!

Additional Reads

1. Your AI feature is quietly destroying your gross margin

2. Aspiring for Intelligence - Sabrina’s Substack

Subscribe now

Some news on the Clarify side: we’ve welcomed Seam AI to our team and we’re building Clarify Signals - market intelligence in your CRM. We have a waitlist here.

Let’s dive in.

My experience founding Iteratively taught me an unbelievable amount of lessons. One that comes up a lot: I’d completely underinvested in investor relationships.

I wrote investor updates, but most of them never went out. We always hesitated to send them for some reason. There was always another milestone around the corner. Maybe we’ll have better numbers next month and the numbers will look better. Maybe we’d finish another launch first. Waiting felt easier than pressing send.

We missed a lot of chances to get investors involved when they could have actually helped.

Investors can only help when they know what’s happening

At Clarify, we send updates every month. They look like this:

We cover lows, highlights, what’s next, and where we’d appreciate help. We also have a kudos section to thank any standout investors for their help.

Before starting Clarify, I invested in startups myself. The companies that were easiest to support weren’t necessarily the ones growing fastest. They were the ones that made it easy to understand what was actually happening inside the business, and where they needed help.

Seeing it from that side helped me understand what is most helpful to investors, so I’m trying to live up to that standard. Even when an update feels ordinary, or even negative, I send it. Consistency matters more than waiting for the perfect story.

I don’t like success theater

Success theater is what happens when updates are written to impress instead of inform. I hate it.

So I’ve gotten more opinionated about how these updates are written.

I don’t like filling them with milestones that are just part of doing the job. Revenue should grow. We should keep shipping. But if those things aren’t happening, we need to be honest about it.

I’d also rather spend that space talking about where we’re making real progress, where we’re stuck, and where I’d genuinely appreciate help.

Those conversations have always led to better relationships than trying to make every month sound bigger than it was.

Investors shouldn’t run the company

I ask investors for feedback all the time. Many of them have experience I don’t, and I pay close attention when they share what they’ve learned, especially in areas where they’ve seen the same problems play out before.

But the responsibility for running the company sits with me and the leadership team. Investors should challenge our thinking and ask hard questions. The decisions are still ours.

Open company, no bullshit

One value I’ve always liked is “open company, no bullshit.”

I try to carry that into Clarify by sharing the same information with employees and investors whenever I can. People make better decisions when they’re working from the same facts.

Investors are in the boat with you. They deserve to know where we’re headed so they can help row.

Until next week,
Patrick

Additional reads

1. You aren’t getting the most out of your investors - First Round

Subscribe now

When we started Clarify, I already knew my role as a founder would change as the company grew. I’d done this before at Iteratively.

What I didn’t fully reckon with was the delta.

I love being deep in product details. It’s my strength. But Clarify grew faster than anything I’d built before.

And at each new stage, it handed me a different job - not instead of the product work, but stacked on top of it: how decisions get made when I’m not in every room, how to keep a growing team pointed at the same problem, whether the way we operate still fits where the business is going.

The product focus never went away. But the job got a lot bigger around it.

Rewiring how we build product

Early on, we kept falling back on the same instincts for product development: opportunity validation, design, development, launch and iteration.

Tried and true but it’s morphing in the era of agentic coding.

Design and development are getting closer, as almost anyone can spin up a prototype of a feature within a few minutes if they spec it out.

Or a big feature gap would come up and the default solution was usually to hire someone, or form a well-resourced project team. More resources, more people, more surface area.

I kept asking the same question: can we build a better system instead of always throwing more resources at the problem? Internally we call it a software factory… enabling our engineers to ship quality features quickly using AI.

That sounds simple. It isn’t. Everyone brings the habits of their last company with them. I do too. Changing how a team thinks about problems isn’t something that happens after one planning session. It’s something you work at, slowly, and revisit every time the business changes.

We’re still working through it. That’s the honest answer.

Growing the team

Around twelve people, something shifted. The two pizza rule strikes again.

When the team was small, information moved naturally. Everyone knew what everyone else was working on. Context was shared by default, not by design.

When we doubled the team, that stopped being true. Nobody changed. But information traveled differently. New groups formed. People had less visibility into work outside their own area. Decisions started feeling less connected.

We didn’t need more meetings. We needed better ways to keep everyone looking at the same picture.

Here’s what I’ve learned: communication becomes part of the product you’re building — the internal one. If people don’t have a shared sense of where the company is going, even good decisions start to feel off.

What actually changes

As the company matured, the job changed again. Now I spend most of my time thinking about where we’re investing our attention and whether we’re executing well enough. Those are conversations I come back to almost every week.

When something doesn’t go the way I hoped, I try to slow down before deciding what needs to change. Sometimes it’s the priorities. Sometimes a good idea just needs more time and more focus to actually work. The difference matters and it’s easy to confuse the two when you’re moving fast.

But the biggest thing I’ve learned is that there’s no point where you 100% figure out how to run the company.

The company changes. The market changes.

Your job changes with it.

And you have to adapt.

For that reason, even if this isn’t your first rodeo as a founder, you’ll always be a first-time founder.

The only thing I’ve found that helps is staying honest about where you actually are. Not where you thought you’d be by now, not where you want to be. Where you are.

Until next week,
Patrick

Additional reading

1. From founder to CEO

Subscribe now

Note: Can’t believe we’re approaching 400 subscribers and over 18,000 views on this thing. Thanks for being here. It means a lot.

Many founders hate sales, but the same founders love doing customer research.

They are closer than you think.

In research calls, you sit with a user, ask what broke, listen to what confused them, bring it back to the team. It feels like building the product.

But then you attach money to the same conversation, and everything changes.

Now it’s “sales.” You overthink how you sounded. Whether you were too pushy. The awkward pause. The demo that didn’t land the way you wanted.

Most founders didn’t start a company to become a salesperson. A lot of us come from product, design, or engineering. We like making things. We like solving problems. Some of us hate being sold to, which makes selling feel strange the first time we have to do it ourselves.

The mindset shift I’ve had to make is simple: founder-led sales isn’t some separate job that sits outside the company. It’s one of the ways you learn if you’re building the right thing.

The path to nailing founder-led sales starts with mindset shifts like these.

Here are a few from other founders that helped them.

Treat sales like a puzzle, not a performance

In the early days, your biggest risk isn’t a bad demo. It’s spending months building the wrong thing.

That’s why I don’t think sales and customer research should feel so far apart. You’re trying to understand the buyer: what hurts, what they’ve already tried, what they care about, what makes them hesitate. A sales call adds pressure because there’s a decision attached to it, but the learning is still there.

Sometimes the call tells you someone’s a great fit. Sometimes it tells you they’re not. Sometimes the pain is real but the urgency isn’t. Sometimes they care about a part of the product you barely mentioned.

All of it helps.

Amanda Zhu from Recall.ai has talked about being so nervous on early calls that she was visibly shaking. Now, when a prospect throws her a hard objection, she treats it like a puzzle to solve. I like that framing. It takes the drama out of the call. You’re not trying to win every moment. You’re trying to figure out which piece doesn’t fit yet.

Have a strong opinion, then let the calls change it

You need some kind of point of view when you start selling.

Maybe you think you should demo live. Maybe you think people should sign up on their own. Maybe you think you’re selling to one type of company, and the calls keep pulling you somewhere else. You won’t know until you put the idea in front of buyers.

Bobby Pinero from Equals sets a good example. They started with a freemium model, and new users weren’t engaging. So they moved to a sales-led model, having every new user take an onboarding sales call, no exceptions.

They talked through use cases, learned what people were actually trying to do, and got closer to the reason someone wanted the product in the first place.

They didn’t treat their first motion like a precious plan. They watched what wasn’t working, got closer to the customer, and changed how they sold.

That’s founder-led sales at its best.

“Good enough” is enough to start

Sales feels hard partly because founders are used to being great at something else.

You might be great at product or design or engineering. Then you show up to a sales call, and suddenly you’re a beginner again.

Steve Hind from Lorikeet rated himself around a six out of ten at sales when he started. You don’t need to be a ten to take your first call. You do need to be willing to practice: rewatch your calls, ask for direct feedback, notice where people get confused, get help when you need it.

The job takes grit

Some (ok, most) of founder-led sales just comes down to effort.

Esha Joshi from Yoodli boarded a cross-country flight with a few hours’ notice to get in front of the right VPs. That’s the kind of ask early sales will make of you.

Pylon’s Marty Kausas works so much that he was literally featured in the Wall Street Journal for it.

I don’t think every founder needs to become a road warrior or try to make the front page of WSJ for working so hard.

But you do have to accept that sales will ask more of you than you want to give at first. Buyers can feel it when a founder is hesitant. They can also feel it when a founder is curious, engaged, and genuinely trying to understand the problem.

I still take the calls

At Clarify, I still try to take as many sales calls as I can. In the first two years, that added up to more than 1,000 calls.

Part of that is because sales matters. A bigger part is because I don’t want to get too far away from the buyer.

The calls remind me what people are confused by, what they care about before they’re ready to buy, and where the product almost fits but needs to get better. You can hire sellers, and at some point you probably should. But in the early days, no one carries the same context you do as the founder. You know why the product exists, what still feels rough, and which parts of the company are still being figured out. That perspective shows up in how you run a call.

Founder-led sales gets easier the moment you stop treating it like a performance and start treating it like practice. Every conversation is a chance to listen a little closer, learn a little more, and do better on the next one.

Until next week,
Patrick

Additional reading

1. Understand the benefits of Founder-Led Sales

2. Harness continuous discovery

Subscribe now

We had (Co-founder and CEO of SuperMe) on the First Pitch podcast recently. A lot of things stuck with me, but the fundraising part most of all.

Casey and Ludo are building SuperMe — an AI-native professional network that grew out of years thinking about networks, expertise, and what AI makes possible now.

In the episode, he takes us back to the beginning of the journey. They’d started building. They had a pitch deck.

Casey now says the deck looks pretty embarrassing.

I liked that he said it. It’s the version of fundraising most founders live through but rarely admit out loud.

You spend a week trying to make the story make sense. You put slides together. You explain the market, the timing, the product, the team, the plan. Then people ask questions, and you realize the story still has holes.

Casey said Greylock thought a lot of the deck was dumb.

Luckily, they didn’t think Casey and Ludo were dumb.

You’re being evaluated, not just your idea

A lot of founders treat early fundraising like the whole job is making the idea sound more finished than it is. So the pitch gets polished until the rough parts are harder to see.

Casey’s story is a reminder that early rounds almost never feel that tidy while you’re living them.

At three months old, a company is still full of open questions. Greylock pushed on how the network would start, where demand would come from, how consumer the product should really be. Fair questions… and some hit places Casey and Ludo were still working through.

But Greylock wasn’t just evaluating the first version of the idea.

Casey and Ludo had spent years around network businesses and lived close to the problem. Casey had been advising founders and operators for years, which gave him real insight into how company-building knowledge gets shared.

At the early stage, investors are trying to understand the idea. But they’re also watching how you think, how you handle gaps, and how you respond when the story gets pushed.

That’s the part founders often forget.

Writing the deck forced clarity

Casey said something simple that I keep coming back to: writing is thinking.

A deck forces you to take the loose stuff in your head and put it somewhere people can actually react to. When the story only lives in conversation, it can feel cleaner than it is. Put it on slides and you’ll see fast which parts hold and which parts need more time.

Casey and Ludo had written memos on specific decisions, but the deck was the first time they tried to wrap the full story together. Looking back, Casey could see where the thinking was strong and where they needed to go deeper.

The deck didn’t have to be perfect, but it did have to make the thinking visible.

Know who’s asking the questions

One of the most useful things Casey shared was his prep approach.

At Pinterest, before product reviews, he’d think through what each person in the room cared about. Different people would push on different parts of the work, so he’d try to answer the obvious questions before they got asked.

For the Greylock pitch, he said he’d have done more of that.

Not in a fake way. More like: who will push on the entry point? Who will worry this is too enterprise? Who will ask how the network actually starts?

But this isn’t about slicing your story for every person in the room.

Andy Raskin has a good framework for this — originally about buyers, but it applies equally to investors. The goal isn’t multiple pitches. It’s one high-level story that’s strong enough to let different people find themselves in it. Each person will filter it through their own lens. Let them.

Your job is to know those lenses in advance and be ready when the questions come.

Raise for your company, not the market noise

Casey also talked about how much to raise.

Some founders around him were going big out of the gate. Casey and Ludo talked to consumer founders and kept hearing the same thing: raise less, keep the team small, don’t take dilution before you know what kind of network you’re actually building.

That’s not a universal rule. A B2B company with a long pre-sales checklist might need to hire fast. A consumer network might need a tight team and low coordination costs while the product finds its shape.

Casey looked at the company he was building and chose the round that fit.

The goal isn’t the cleanest story or the biggest round. It’s to know what investors are actually evaluating, own the gaps in your narrative, and raise for the company you’re really building.

The pitch will have holes. Raise anyway.

Additional reads

1. Why VCs “almost blindly” invest in founders with previous exits

2. The Greatest Sales Deck I’ve Ever Seen - a classic

Subscribe now

Thinking through how to budget AI spend has been on my mind for two reasons:

1. We just launched Agents in Clarify. One of the first questions we get from users is about how it’s priced. Pricing AI products is hard, and so is measuring your spend on them.

2. Someone sent me Marty’s tweet about Pylon’s AI spend.

Since then, a few founders have asked how I think about budgeting for AI spend at Clarify.

The honest answer I give them is that I don’t think there’s one clean category or number for it yet. I’ve seen it lumped into tooling, R&D, cost of goods sold (when tied directly to the product). But I’m less interested in the label right now and more interested in what the spend is doing for the team.

At Clarify, the main thing I care about is getting the right product to market and making sure customers love it. So when I look at AI spend, I ask one question: does this help us do that?

If it does, I’m pretty comfortable spending.

Budget for trying things

The models are changing fast. What was best two weeks ago might not be the best today. We encourage people to experiment and share what’s working. If someone finds a better workflow, I want them talking about it.

Claude, Codex, Cursor, Conductor, whatever someone is using, the tool changes how they work. The bottleneck is still the human using it. If one tool helps someone think better, ship better, or get unstuck faster, that matters more than standardizing the line item

Track usage without making it weird

We look at who’s spending tokens and who isn’t. It’s not a shame exercise; I’m curious whether usage maps to better work.

If an engineer is spending more on AI, are they shipping more useful products? Are they opening more PRs? Are customers getting value sooner?

If someone on the GTM side is using AI, are they getting more time back for customer work? Are their follow-ups more timely? Has their admin work been reduced?

There are probably ways we could save money. I’m sure there are. But right now, cost savings isn’t the main thing I’m optimizing for.

If someone can spend $100 and save 30 minutes, and that 30 minutes goes back into talking to customers or building something customers need, I’m probably fine with that trade.

This is also a question of how you value people’s time. People’s time is worth a lot to the business. If AI helps them get more of that time back, the cost can make sense pretty quickly.

Put AI where the work changes

Engineering is an obvious starting point. But one thing people still miss: AI changes the shape of the work, not just the speed. Instead of having an hour-long conversation about an idea, sometimes it’s better to just build a rough version and react to it.

The old software development process was more sequential. Research, design, prototype, engineering. Now an engineer can sometimes build the prototype first, show it to design, and get feedback earlier.

I like this line and diagram from Brian Balfour: AI prototyping is great for discovery and less so for delivery. It’s not about generating code that can later be shipped. It’s primary benefit is determining what to build, then handing a detailed spec to eng to actually build it.

It doesn’t always work. The engineer still needs to understand the customer and the problem before prototyping. But AI makes it easier to try things once the problem is defined.

I also think GTM is probably one of the highest-leverage places to use AI, especially for a company like ours. Marketing, sales, and post-sales all have a lot of work where AI can help.

Watch for AI junk

AI can also create a lot of junk.

It can create more emails, more drafts, more notes, more ideas, more everything. More isn’t automatically useful.

The goal is to use AI where it gives people time back or helps them do better work. Not just where it creates more stuff.

I also wouldn’t bet everything on one model too early.

I use different models. If I don’t like what Claude gives me, I might try Codex. Sometimes I want to see how two models approach the same thing.

There is some lock-in, mostly because of memory. The more I use one tool, the better it gets for me. Moving from one to another can be annoying because the context doesn’t move cleanly.

But I still think it’s worth keeping options open.

For now, my view is pretty simple:

• Budget for experimentation

• Track usage

• Watch whether the work is getting better

And I wouldn’t panic about the line item too early if the spend is helping the team get closer to customers and ship better products.

The bigger mistake, at least right now, is underusing AI because you’re trying to make the spreadsheet look cleaner.

Until next time,
Patrick

Additional reads:

1. Interesting Reddit thread: Is your company worried about how much AI is costing?

2. Agentic AI token costs - EY

Subscribe now

The first episode of our new podcast with Somrat Niyogi, The First Pitch, launched last week. You can listen on YouTube or Spotify. I’m in the hot seat for this one.

Now into today’s topic.

Every new function or pile of work forces the same question: do I hire someone, outsource it, or use AI?

My answer used to default to people. Now AI is always the first thing I test.

Not because AI can do everything. It can’t. But I don’t want to hire or outsource before I’ve figured out what’s actually possible without adding headcount or budget.

The way I think about it now: AI first, then look for the pain on your team, then decide what needs to stay inside the company. That order matters.

Start with AI

A small example: in the old world, having too many meetings to schedule, too many follow-ups to send, and too much CRM admin to do might have meant I’d hire a sales rep or even an EA.

In today’s world, I felt that pain, and built a workflow into Clarify that allows me to email rep@clarify.ai and it helps with sales admin work like scheduling, deal details, and follow-ups.

I wouldn’t have thought to build a rep@clarify.ai workflow if I wasn’t constantly forcing myself to look for places I could take work off my plate with AI.

When AI isn’t enough, hire for the pain

If AI can’t handle it, I don’t immediately think “who should I hire next”, but instead target where the pain is felt most among my team.

Who’s carrying the work right now? How much pain are they feeling? How long can they keep carrying it before it starts hurting their morale or their output?

For example, if someone is dealing with a ton of migrations and onboarding work, my first reaction might be to give them another person. But I also want to ask what’s causing the pain. Is it just too much volume? Is the system broken? Would another hire actually help, or would they get dropped into the same mess?

So before I hire, I want to understand what needs to change for that person to be successful.

Some things, you can’t outsource

When it comes to outsourcing, I am now more intentional about what we should own internally. Outsourcing is a great option, but there are specific instances where I think we should remain close.

Brand is one of them. It’s too important to treat it like a project someone else can fully own.

It needs to be felt inside the company. It needs to evolve with us. An outside partner can help, but I don’t think they can own it the same way.

Product design is another. I’m biased because of my background, but I wouldn’t outsource it.. It’s too close to how the product works and how customers learn to trust it.

Sales is similar. I wouldn’t spin up some outsourced BDR agency and expect them to figure out how to sell our product. If that worked, everyone would do it. I think you have to know how to sell your own software.

Messaging and positioning also need to be owned inside the company. You can get help, but the company has to own the thinking.

Engineering depends on the situation. I’ve outsourced engineering before, and I’d do it again. But usually I’d want it to be project-based. Clear scope, clear owner, clear finish line.

Think about it like beer!

The way I think about it is like running your own brewery. If I’m in charge of the whole operation, I’m not going to outsource the brewmaster. I’m bringing the best one I can find in-house. That person changes how the beer tastes.

But I might outsource or partner with people who make the bottles or the labels. That type of work doesn’t change the “taste” of the product. But it still adds up to the entire flavor of the company.

Did I take it too far with the beer example?

Anyway, to wrap it up: AI is a partner. Outside firms can be partners. But the things that require a close eye or a special touch? Keep them internal.

Your customers will know the difference.

Additional reading:

1. When to hire a FTE and when a contractor makes more sense - Mercury

Subscribe now

I keep reading that SaaS is dead.

I don’t think that’s right. Software isn’t dead. But some of the long standing mechanics of it is.

Pricing is getting disrupted. Seat-based pricing is bullshit in this era. And the bar for what makes a good software company is higher now.

Let’s dive in.

AI changed who can build software

That’s the obvious part. A founder can hack together a tool in a weekend. An engineer can build a prototype in an afternoon. Someone on the team can use AI to make a workflow that would’ve taken way longer a few years ago.

There’s always been a build vs. buy argument in SaaS. AI made the build side cheaper, though it didn’t make it free.

You can vibe-code a CRM, for example. But it only works if you’re the only person using it.

The second your team starts depending on it, the problems change. Now you need permissions, reliability, consistent updates, a real security model. You need everyone on the team using it the same way, the data staying clean, the thing not breaking when you’re trying to close a deal.

At that point you now need to maintain the software and respond to feature requests. And running that software becomes more of a job than you expected.

That’s why companies still buy tools. Not always because the tool is cheaper than building it. A lot of the time it’s because someone else is waking up every day making that thing better and you don’t have to.

AI also changed what makes a SaaS company worth building

The best founders I know are pretty ruthless about one question: what work actually makes our company better?

Most of the time the answer isn’t building your own CRM or maintaining your own analytics stack. It’s something specific to what you’re trying to do. The thing only you can do. Buying software is a focus decision as much as a resource one. The companies that win usually pick a lane and go deep on it.

That logic applies to SaaS founders too. If you’re building a software company, the question is the same: what do you own that nobody else can? AI made it easier to build. It didn’t make it easier to build something people actually need, trust, and keep paying for.

That bar is higher now. Not because software matters less. Because it’s easier than ever to build something mediocre.

So should you start a SaaS company?

Yeah. I think so.

I should say upfront I’m probably not the most neutral person to ask. I started a venture-backed CRM company. CRM is one of the most crowded categories in SaaS. So take this for what it is.

But I’d be honest about what kind of company you’re trying to build.

If you’re a first-time founder, this is probably a good time to build a solution to a niche problem. AI makes that more possible than it’s ever been. You don’t need the same size team or as much capital. You can get further with fewer people.

Venture-scale is a different question.

Starting is easier than ever, but standing out is harder because every other founder got the same leverage from AI. The 2021 world where public SaaS companies were trading at crazy revenue multiples is harder to count on. And the bar for funding is higher now too — some AI companies are setting expectations for revenue velocity that most SaaS businesses can’t match.

So taste matters more now than it used to.

Capital used to be one of the biggest bottlenecks. You needed money because you needed engineers, and engineers were expensive. AI made that more evenly distributed. A founder with $5 million can sometimes build something better than a founder with $100 million if they have better product judgment.

Despite my point that niche markets are interesting right now, I went into a crowded category anyway. I think about that a lot and wrote about it here.

The questions haven’t changed

Are you solving a problem people care about? Are people willing to pay you to solve it? Can you build something they trust enough to run their business on?

Those questions are the same as they’ve always been. What’s changed is how many people are competing to answer them.

It feels like a pretty good time to start, as long as you’re clear on what you’re actually trying to own.

Until next week,
Patrick

Additional reading:

1. SaaS is dead. Long live SaaS! - Crunchbase

2. Everyone says SaaS is dead. Here’s what’s actually happening to software - Shane Collins

Subscribe now

One quick note before we start, I’ve been working on a podcast with Somrat Niyogi called The First Pitch: we talk to venture-backed founders about the original pitch that secured them funding and the journey since. Trailer below, first episode on June 10th on YouTube and Spotify.

Enough notes - let’s dive in!

Two weeks ago, I was on a panel about finding product-market fit (PMF) in the AI era.

One of the speakers was Loren Alhadeff, former CRO and COO at DocuSign and current Venture Partner at Madrona. He said something I’ve been thinking about a lot since.

When he described PMF at DocuSign, he described it as a stage, a moment you arrived at. You deployed the product. If you found product-market fit, it was soon integrated into enterprise infrastructure and embedded everywhere.

Essentially, once you ‘had it,’ you were in.

Then he talked about PMF today, and it’s 100% aligned with what I see at Clarify and hear from fellow founders:

“It [PMF] doesn’t feel like a stage anymore. It feels like a state. There’s no such thing as ‘we now have it, so we’re good.’”

For me, hearing this was a timely reminder. As much as I know that PMF is now an ever-moving target that I have to stay ahead of, I still fall back into the old way of thinking about PMF as something you achieve once and revisit never.

And to be completely honest, sometimes I think about the old way and miss how simple it was.

You’d find PMF, you’d exhale, and you’d focus on scaling.

Today, you can’t afford to exhale. If you do, you lose PMF. That’s how quickly things are changing with AI.

PMF is no longer a final destination

I’m being reminded of this myself every day. It’s hard. As someone who founded a company before the AI era and another one during the AI era, I have to remind myself to stay in discovery mode all the time.

For example, I recently went through Clarify’s customer onboarding flow for the first time in six months. I went in thinking it was going to be nearly flawless and intuitive. Yet I found our activation funnel was broken in ways I never expected. Many things I thought were intuitive for customers were anything but.

That’s what happens when you treat PMF like a final destination, when in reality (and especially today), it’s more like a never-ending journey.

Reforge also reinforces this with the idea of the PMF Treadmill: Product Market Fit is a key milestone to reach, but it’s often misinterpreted as being a static moment in time. The reality is that your customer base is always changing and consumer expectations are always growing. Once you get initial product market fit, you not only have to keep it but also expand it.

AI is a large reason for this, which is great in some ways and challenging in others.

Nowadays, we can all build faster, ship more, and reach more people sooner. That’s great if you have PMF. It’s really bad if you don’t. Going a million miles an hour in the wrong direction is the fastest way to an ugly exit.

But if you have PMF, then AI makes rapid, exponential growth more possible. I like the way Collin Stewart, co-founder of Predictable Revenue, talked about this relationship between PMF and GTM:

“Product-market fit, that strength is a big multiplier of your go-to-market engine. The stronger your product-market fit is, the more you’re gonna get out of that engine.”

What maintaining PMF in 2026 actually looks like

For me it comes down to one thing: talking to customers. I’m still doing 10 to 12 customer calls a week.

Not because I love it more than building. But because the moment I stop talking to customers, I begin navigating by memory instead of feedback. And memory is a magician…it can play tricks on you and warp reality.

There’s also a difference between talking to customers and talking to the right customers.

I tend to weight about 80% of my calls toward prospects, 20% toward existing accounts. This mix helps me stay close to different needs of our target customers, including those already familiar with us and our product and those in the consideration phase. It’s the best way to pick up on any opportunities (or red flags) around our PMF.

What I’m looking for in those calls is an existential need for our product. So, I’m constantly thinking: If our product went away tomorrow, what happens to our customers’ businesses? If nothing changes, or if they can replace us with something else rather easily, then do we really have PMF?

Like Loren said during the panel, “If the reaction is, ‘Oh well, there’s like six other things that I kinda know about that I think sort of do some of the similar things, and it’s actually pretty easy to just swap over,’ that’s not [product-market] fit.”

Taste is still the difference-maker between ‘initial traction’ and PMF

It seems like everyone’s talking about it right now, but taste is another critical thing to keep in mind when thinking about PMF.

It’s easy to build something right now. Vibe code a product in a weekend, get it in front of people, see if it has traction. That’s a legitimate capability today. But just because something is easy to build doesn’t mean it’s easy to build well. You’re probably vibe coding something that’s a C product. You might get initial traction. But so many people doing this are mistaking that little bit of traction as product-market fit.

Before they find out they don’t have it, it’s too late.

Truth is, an A+ product still requires taste. You need to be a founder with conviction and human judgment about what matters and, just as importantly, what doesn’t matter. It also takes close proximity to a lot of customers to learn the difference between someone who’s excited and someone who would actually panic if your product went away.

There may be more software products than ever right now. That doesn’t mean they’re good. Most of them are C products, very few are A+ with true, ongoing PMF. Don’t trick yourself into thinking that the difference is speed. Everyone is fast now. The biggest difference-maker today is taste, hands down.

The innovator’s dilemma, reframed

Loren called maintaining PMF an innovator’s dilemma. You’re constantly iterating to keep users ingrained with the product. The moment you stop, someone else starts.

I think about it like fitness. Nobody finishes a marathon and stops training forever. The finish line is just a marker. The state is ongoing.

Stay close to the customer. Keep asking the hard questions. And whatever you do, don’t exhale.

Until next week,
Patrick

Additional Reads:

1. Product Market Fit Collapse - Reforge

2. Why $100M Doesn’t Mean You Have Product Market Fit - Elena Verna via Growth Talent

Subscribe now

Burnout is real. And if you burn yourself out, you’re not good for anybody. Not your team, not your investors, not your family.

At the same time, I do believe this:

So how should a founder push as hard as they can without burning out?

I wish I had a more exciting answer here, but the basics are still the basics: sleep, eat well, exercise, take time off when you need it.

Boring advice, but startups make that harder because there’s always a thousand things going on, and it all feels urgent.

I’ll be honest about the trap I fall into: late nights feel productive. This week I was up until 11:30pm most nights working. It’s easy to tell yourself you’re getting ahead, then the next day, you realize you’re tired, slower, and making worse decisions.

Short sprints work better for me than running at a million miles an hour all day.

Taking time off without the guilt spiral

I don’t take a ton of time off. I’m not going to pretend I’m the role model here. I’m definitely a “do as I say, not as I do” guy on this one 😅.

I try to take two solid weeks off a year, usually between Christmas and New Year’s. Even then, work tends to pull me back.

This last break, a new model launch changed how I thought about what we should build, and I was on my laptop during vacation messing with it.

I’ve found I need to be clear about what I’m protecting when I take time off. I keep one day on the weekend (usually Saturdays) 100% dedicated to my family.

Sundays sometimes turn into half a workday plus Sunday night. I don’t love it, but that’s where it lands right now.

Founding with two kids

My first company was pre-kids. Clarify has been with kids, and it’s harder. No question.

I want to show my kids what hard work can do, but I also don’t want to fast forward 15 years and have my kids feel like I wasn’t around.

So I protect a few rituals:

• 30 minutes with my kids in the morning before I go upstairs

• Dinner with my family

• Bedtime with my kids

There’s also a difference between quantity and quality time. My kids can tell the difference, and so can I. An hour of real presence beats a whole day of being physically there, but mentally at work.

Founder parents also need partner support. You’re asking your partner to pick up slack when you’re heads down on a launch or stuck in back-to-back calls. That’s a real ask, and it compounds.

This division is something you need to discuss with your partner early. My wife and I call it in the first team rule.

This balance is really, really, tough, and I’m sure other founder parents can relate.

Stress shows up everywhere

Work stress bleeds into home, home stress bleeds into work.

If I slept like shit because the kids were up, I might be grumpy in a 9am meeting and not even realize why. If a board call went sideways, I’m not the fun, goofy husband or dad that I want to be at dinnertime.

The best thing I’ve found is just being upfront about it. Telling my team, “Hey, rough night, I’m not at my best today,” goes a long way. Same at home. Pretending you’re fine when you’re not doesn’t fool anyone, it just makes everyone walk on eggshells trying to figure out what’s wrong.

This is the part I keep relearning. Staying sane isn’t a vibe or a productivity hack. It’s basic maintenance, plus a few protected routines, plus being honest about when you’re running on fumes.

Some weeks, I do it well. Some weeks, I don’t. The goal is noticing when things start to slip and course correct.

Until next week,
Patrick

Additional reads:

1. The 3 signs of founder burnout: Mercury

2. The way out of burnout: The Economist

Subscribe now

I stepped away from a company I co-founded out of college.

I was the technical founder, but I wasn’t the right person for the CTO job at that point. I also wasn’t in a financial position to take the risk. Both things were true, and either one alone would’ve been enough.

Role fit matters more than founders like to admit. So does knowing when to walk.

Brian Halligan did this at HubSpot. The company was moving from scale-up into the enterprise leagues, the job was changing, and he stepped into Executive Chairman so Yamini Rangan could run it as CEO. HubSpot wasn’t struggling. He just wasn’t the right person for the next chapter, and he knew it.

That’s the part most founders miss. The decision isn’t only forced on you when things are bad.

I still believe in giving it everything. I did that at my last company — I was willing to turn myself inside out to make it work. Same with Clarify. Effort matters. It just doesn’t guarantee anything.

Most founder writing is about how to hang on. This is about the other side of that.

Here’s how I think about letting go.

The two things founders run out of

Money is the obvious one. When it dries up, the options narrow fast. You can’t pay people, you can’t keep investing, and the call gets made for you.

Energy is the one founders don’t track. You can have money in the bank and still feel done. Tired, burned out, no drive to carry it forward. Energy hits zero quietly, and most founders don’t notice until they’re months past it.

They’re connected. When either one runs out, you’re forced to make a call.

Founder doesn’t always equal CEO

The job of a founder at an early-stage company isn’t the same as the job of a CEO as that company grows. The scale changes, it requires different skills, and an evolving type of leader. If the role you signed up for isn’t the role the company needs anymore, that’s a signal too. Not a failure. A signal.

Stepping aside when the company is doing well

Some situations have nothing to do with money or burnout, or even role fit.

If you don’t feel like you’re doing your life’s best work, don’t stay. If you don’t feel like you’re adding to the value of the company, don’t stay. Set the company up for the next chapter. Life is short.

I keep coming back to a simple frame: company > team > me.

It’s easy to apply that to other people… to spot when someone is putting themselves first. It’s harder to apply to yourself.

But the question is the same: is your presence helping the company, or are you staying in the seat because it’s your seat?

Some founders need help. Some founders need to step aside. The only way to tell the difference is to look at outcomes, team health, and whether the company is actually better with you in the role.

The hard part

Founders shouldn’t romanticize suffering. They also shouldn’t quit the first time it gets hard.

The hard part is telling the difference and paying attention before you waste years.

Until next week,
Patrick

Additional reads:

1. When should a founder step down? - The Economist

Subscribe now

Last week I had four conversations about the same decision.

An investor told us to go faster and ship across a wider surface area. A customer told us to slow down and add depth to what we have. A teammate told me we were thinking about it wrong. My co-founder told me to trust my gut.

All four were right, in their own way. None of them were going to make this specific decision for me.

This is most weeks.

Founders get input all day every day. Employees, customers, investors, board, spouse, friends. All of it is noise; not all of it is signal.

You can spend your whole day reacting if you let yourself. I try not to.

Most of the founder’s job is gathering input and making the call when it’s theirs to make. I try to make as few decisions as possible. Just the ones that are uniquely mine. If too many decisions hit my desk, it usually means we didn’t give someone enough context, or we put the wrong owner on it.

Your priorities determine how you listen

At this stage, my focus stays narrow: product-market fit and distribution. That’s it. You need to choose yours.

Inputs that help with those two things get my attention. Inputs that don’t become future Patrick’s problem.

The investor suggesting we hit more field events because they’ve seen it work well for that audience? That’s a distribution input. I take it seriously.

Multiple customers asking for a certain integration or MCP connector? That’s product-market fit. I take that seriously too.

Input that strays too far from those two categories gets less weight. Not because it’s wrong, but because it’s not load-bearing for what I’m trying to figure out right now.

What I do when someone sees it differently than me

I try to form my own view first. Then I weigh it against the data people (and the market) bring me.

Internally, if someone has the same information I have, knows our values, and is good at their job, I expect they’ll land close to the same decision I would.

When they land somewhere else, I get curious. I want to understand what they saw, what I missed, or how they think.

That’s what happened with my teammate last week. They had a different read than I did, and they had context I didn’t. I changed my mind.

What if you’re the input for someone else’s decision?

People sometimes come to me pre-decision, asking for information or my thoughts.

My first question back is simple: what would you do?

If their plan lines up with how I’d approach it, I tell them to go do it. If it doesn’t, I ask why and help them think through it.

Most of the time, they already know the answer. They just want someone to confirm it.

Where I go to learn

One of the hardest parts of being a founder is that everyone expects you to have answers. A lot of the time, I don’t. Not on the specific topic in front of me, anyway.

So I’m constantly trying to learn enough to be dangerous in the next conversation.

I like podcasts. But most of the time now, I just ask Claude.

If a podcast dropped and it isn’t in the model yet, I’ll link it, ask for a summary, then ask how it applies to our company.

I also go to people and communities I respect when I have low confidence on a decision. Those moments matter more; I’m reaching out because I’m missing something. I keep a short list of people I trust on specific topics: investors, advisors, teammates who have more context than I do.

The simplest filter

Advice sounds credible all the time. Credible doesn’t mean it applies to your situation.

Two questions I keep coming back to:

• Does this help with product-market fit or distribution?

• Does the person giving it understand the details of what we’re building?

If both are true, I dig in. If either is missing, I store it and move on.

Back to those four conversations. By the end of the week, I’d made the call. The investor input shaped the timing. The customer input shaped what we’d ship. The teammate changed my mind on the approach. My co-founder was right that I needed to trust my gut - but only after I’d done the work of weighing the rest of the inputs.

The trap is giving every input equal weight. The job is knowing which ones to pick up.

Until next week,
Patrick

Additional reads:

1. First Round - Advice is more important and more overwhelming than ever. Here’s how founders can cut out the noise.

Subscribe now

Positioning in the age of AI is getting messy.

Everyone’s racing to say the same things. AI-native. AI-first. Built for the AI era. The phrases are everywhere, and they’re starting to mean nothing.

I’ve been thinking about this a lot lately and talking to some marketers in my network about it, specifically through the lens of three audiences: customers, investors, and recruits.

And I keep coming back to the same hierarchy. Customers come first. Investors second. Recruiting last.

That order matters more than most founders act like it does.

The recruiting trap

The pressure to oversell in recruiting is real, especially when you’re an AI company competing for talent against bigger names with more brand recognition and bigger comp packages.

The temptation is to lean hard into the vision. We’re changing how the world works. This is a once-in-a-generation moment. Come build with us.

And the vision should be in the pitch. That part is true.

But we think about this a lot at Clarify. The pitch has to be exciting enough to get someone in the room, but honest enough that the work matches what you told them. Because if those two things are too far apart, you find out quickly. Someone joins, and a week in the gap between the story and the reality starts to show. That’s expensive for everyone.

The founders I’ve seen handle this well are specific about what the work actually looks like, not just where the company is going. Both things can be true. You can be building something that matters and also be honest that some days it’s a data model problem.

Investors are a different animal

They want a big vision. The TAM slide, the category narrative, the “we’re going to change how companies sell forever” framing. That’s fine, and it’s mostly a pitch deck thing.

But investors aren’t naive. They go talk to customers. They do reference calls. They check whether the story you told them in the boardroom matches what customers are actually experiencing. The gap between your investor narrative and your customer narrative is something good investors notice quickly.

Keep the vision ambitious, but keep the ground truth consistent. You can’t have two completely different versions of what you’re building.

How to talk about AI without overdoing it

When it comes to customers, I prefer a more subtle approach, and Apple is still the best example of how to do it right.

They don’t bash you over the head with endless AI / agentic feature talk, they bundle it all under “Apple Intelligence.” It’s elegant and reduces cognitive load.

They tend to lead with the problem you already understand or the outcome you want, and AI is just the thing that makes the solution work better. There’s no breathless announcement that the new iPhone has “AI-powered photo processing.” It just takes better photos.

The outcome is the message, and the technology is the delivery mechanism. We’ve tried to live this at Clarify. We just launched Prospecting. It’s AI-powered, but we didn’t use that term anywhere in the messaging. Here’s what we said, instead:

We lead with the outcomes and the use cases, and only nod directly to AI in the visuals.

There’s actually research backing this up, too. Irrational Labs surveyed 767 software users and found that explicitly labeling a product as “AI-powered” didn’t increase trust, didn’t raise willingness to pay, and in some cases actually lowered performance expectations. The researchers suggested this decline may reflect growing skepticism from tools that overpromise and underdeliver.

Worth sitting with that for a second. The thing most companies are leading with in their marketing is actively working against them with real buyers.

The study also found that companies are still overusing the AI label anyway, largely because of investor and analyst pressure to signal innovation, or because they’re optimizing for SEO, or just plain FOMO about falling behind competitors. In other words, AI messaging is often more about internal signaling than actual customer communication.

Leading with AI in your positioning also won’t age well. A few years from now, everything will have AI in it. HubSpot has AI. Salesforce has AI. The “AI-native” label stops being a differentiator the moment it becomes the default. You don’t want your core positioning to be something that turns generic as the category matures.

What survives is solving a real problem for a specific person. That’s always been true.

What I look for as a buyer

I’ll flip this around for a second.

When I’m the buyer, evaluating a new tool for the company, positioning still matters, but I’m skeptical by default. Most software companies have learned to write good copy. The bar for sounding credible is low.

What I actually care about is getting into the product and playing with it. If there’s a lot of friction to try it, a 20-minute demo before I can see anything or a “contact sales” button where a free trial should be, I’m already turned off. My assumption becomes that the product isn’t ready, or they’re afraid I’ll leave before I convert.

When I land on a website cold, three things matter to me right away:

Are they building for me? Not “for companies like me.” If I have to squint to see my problem in their messaging, I move on.

Do I understand what it does? Not in theory. Not “AI-powered revenue intelligence platform.” What does it actually do, in plain terms?

How much is it going to cost? This one has gotten worse in the AI era. A lot of companies are charging per seat and per AI usage and layering on credits on top of that. Three pricing variables before I’ve even signed up. I don’t trust it, and I don’t move forward.

Pricing clarity is table stakes. The research I mentioned above found that customers were unwilling to pay more for AI-labeled products unless the tool demonstrated clear, compelling benefits. Simply including “AI” in the description wasn’t enough to justify the price. That matches my own experience on both sides of the table.

Where I’ve landed

If you’re figuring out your positioning right now, I’d keep it simple.

Position for the customer. Lead with the problem you solve and who you solve it for. Be specific because vague positioning doesn’t stand out, it blends in. Keep the AI part tasteful. Let it be the thing that makes your product better, not the thing your product is.

As the Growth Unhinged piece put it: users won’t remember your tool as “AI-powered.” They’ll remember how it made their lives better. That’s the whole game.

Focus on the customer. Be thoughtful about AI. Make it easy to try. Be clear on pricing.

Until next week…

Patrick

Additional reads:

1. Customers don’t care about your AI feature

2. Y combinator thread: Using the term ‘AI’ in product descriptions reduces purchase intentions

3. Anthony Pierri LinkedIn post

Subscribe now

Quick news: we just launched the Founder-Led Sales Guide. 20,000+ words, 11 founders interviewed, free to read. Check it out and let us know what you think.

Now for the main event.

At my first startup, I made a mistake I see a lot of first-time founders make.

I assumed the hard part was building the product. If we built something good enough, people would come.

What I learned the hard way is that almost everything early on is a guess. The product, messaging, channels, pricing. Even what your company does can start as a guess.

Founder-led sales is the thing that tests all of it

What I mean by “all our assumptions were wrong”

At Iteratively, I assumed our customers would come from partnerships and outbound email. I spent real time on both. Neither worked.

What actually worked was Slack communities and content. I never would have predicted that going in.

That’s what I mean. Not that we built something bad. We just had wrong assumptions about channels, messaging, who would buy, and why. Most early-stage companies do. The question is whether you find out fast enough to do something about it.

If you stay close to customers, two things happen. You’re less likely to build the wrong thing. And if you do build the wrong thing, you can still find your way forward, because you have context. You know what people are trying to do, what they care about, what they keep tripping over.

Skip the discovery work and you keep restarting from zero. You run another experiment, but you don’t know what you’re even trying to learn.

The pattern I’ve seen

I’ve been doing this for about seven years across two companies. While putting together the Founder-Led Sales Guide, we talked to a bunch of founders and the pattern was consistent.

The founders who stayed close to customers changed their minds faster. Those changes were usually the reason the company worked.

The changes tend to show up in a few places.

Channels

At Iteratively, I tried a lot of things. LinkedIn, email outbound, webinars, partnerships. Some felt productive. Most produced nothing.

What actually worked was almost embarrassingly simple: hanging out in Slack communities where our customers already spent time, being genuinely helpful, writing content that answered questions people kept asking. Half our customers came from community. The other half came from content.

Other founders got there differently. Amanda Zhu at Recall.ai built her early pipeline through cold email, about 20 accounts a week, 35% response rate, because she came from the technical space she was selling into. Mark Tanner at Qwilr found his first customers through outbound calls and in-person demos. Enzo Avigo at June leaned on his personal network.

You have to try a few to know which one will catch.

GTM motion

Sometimes you find a channel that works and still struggle, because the motion itself is wrong.

Bobby Pinero at Equals pushed freemium and watched signups jump 4x. Looked great on paper. But people didn’t know what to do once they were in, and the upsell never came. The counterintuitive fix: put more friction in earlier. Qualify people. Spend time with them before letting them loose in the product.

Esha Joshi at Yoodli started B2C, then noticed customers would show up intensely for a single event, a speech, an interview, then churn once it passed. They moved to B2B, where the use case was ongoing and the contracts were annual. B2C is still the vision. Just not the move right now.

Both changes came from watching what customers actually did.

Messaging and positioning

You can build the right product and still tell the wrong story.

Marty Kausas at Pylon started as a Slack-to-Zendesk plugin, then kept hearing the same question: why do we need to buy Zendesk to use you? Can’t you just be the whole product? They resisted for a while. Felt too hard, felt like competing with Salesforce. Then they listened. Their tagline went from “building the next Zendesk” to “the Zendesk killer” because that’s the phrase customers kept using. They just ran with it.

Enzo at June started in product analytics, then noticed customers were using it for a completely different job. Tracking individual customers, taking action, not analyzing aggregate trends. “It’s scary,” he told me, “to let go of what you thought you were.”

The company itself

The hardest change is when none of the above is the problem. It’s the core thing you’re building.

Amanda and her co-founder were originally building Perfect Recall, a tool for user researchers. When they tried to sell it, they realized they were one of about 20 options in the market and the top three had already solved the core pain. So they pivoted to what they knew: the API layer, targeting developers. They understood the persona, could write cold emails that resonated, and started selling before the product was even finished. A lesson learned from spending too long building before validating demand.

Riya Grover at Sequence had a working product and early customers, but something felt off. Not broken. Just not different enough to build something generational. So they went looking for a better angle. That required a level of honesty most founders don’t want to have with themselves.

The only point of founder-led sales

Founder-led sales is a learning loop.

It gives you enough contact with customers to know what to change, and enough reps to tell the difference between “this is hard” and “this isn’t working.” Those feel similar. They’re not.

At Iteratively, I changed channels, messaging, and how we thought about the product based on what I heard in sales conversations. None of it was comfortable. Every time it felt like admitting I’d been wrong.

But that’s the point. When the market tells you something you didn’t want to hear, that’s the signal. The whole purpose of staying close to customers is to generate that signal early enough to do something about it.

What you do with it is what separates the companies that find product-market fit from the ones that keep building in the dark.

— Patrick

Additional Reading

For more stories and advice like this, check out our new Founder-Led Sales Guide.

Subscribe now

A few weeks ago I saw an internal Slack thread where engineers were talking about their legitimate fear that AI is going to take their jobs.

This conversation is happening in your company, too, even if it’s not out in the open.

I get it. People are staring at tools that can write code, design screens, draft copy, and spit out plans faster and sometimes better than they can. So they’re asking the obvious question: How secure is my job? Or, one level deeper, what is my job now?

Here’s where I’ve landed after thinking about it a lot.

Everyone’s job is getting disrupted

This isn’t just engineering. It’s design, product management, marketing, everything.

I said this in a pretty blunt way recently: if you’re not going to become AI pilled, you’re going to get left behind.

I don’t mean that as a threat. I mean it as a description of what’s happening.

If someone told me I had to do my job without Claude right now, I wouldn’t do it. Knowledge work has changed permanently.

The job is leverage now

The part that’s changing is what people think “doing the job” means.

For engineering, writing code has always been a big part of the work. AI is eating a lot of that. Pre-AI, a lot of hires were capacity plays. You knew what needed to get done, you just didn’t have enough people to do it.

Post-AI, I need people who can self-manage, figure things out, and handle context without a lot of oversight. The role is shifting from writing code to building product. The what and why matter more now, because the floor on the how keeps dropping.

Good ideas still matter. Knowing what to build or trash still matters. Taste absolutely still matters. Engineering isn’t going away - it’s getting more important. But the engineers who win will be the ones who shifted from writing code to building product.

Product and design already knew this

This is actually why there’s less anxiety in those roles. The thing product and design people were proud of was never the artifact they produced. It was what customers experienced. AI doesn’t threaten that. If anything, it gets them closer to it faster.

Junior talent has a weird tradeoff

That said, our team is quite senior, and junior talent has a weird tradeoff here. Read this note from the Stack Overflow blog:

As per the Stanford Digital Economy study, for jobs with the most AI-exposure—read: IT and software engineering jobs—employment has declined 6% for workers aged 22-25, while it’s increased 9% for workers aged 35-49. Hiring is happening, just not for people my age.

A lot of the soft skills I’ve mentioned above are built with years of experience. Junior talent won’t have that coming out of college.

But I’m optimistic. What they will have is AI literacy beyond what any current senior talent has. They’ve grown up with AI as a constant reality. They can push organizations to understand how it can be used to its full potential.

The downside is they might not know how to do basic stuff. Write PRDs, understand code, write emails. Because AI has done it for them. They know how to produce, but they might not understand how it all works under the hood.

My honest take: AI fluency without foundational skills is a fragile edge. It works until it doesn’t, and when something breaks, you need to understand why.

Founders have a responsibility here

This is the part that’s been sitting with me.

I’ve got two fears. One is that the company dies. Every founder can relate to that one.

The other is worse in a different way: we don’t push the team to adopt AI, and we end up with people who are behind where they need to be in AI literacy if they ever need to move on from Clarify.

That thought hit me hard a couple months ago. In January, I felt real resistance internally to adoption, and it stressed me out. The company might be fine, but if someone leaves and they can’t get hired because they didn’t adapt, that’s on us too.

So I do think founders have a responsibility to build the culture here. People can’t treat this like a nice-to-have. They have to use it, explore where it works, and keep pushing the edges.

The fear your team has about AI taking their jobs? It’s not irrational. Jobs are changing, some faster than people expect. What you can control is whether your team comes out of this more capable or less.

The hardest part is the rewiring. The dopamine hit has to shift. From writing code to managing agents, from doing the work to directing it. I went through a version of this when I started doing sales. I hated it at first. I had to force myself past the fear of putting myself out there and being bad at it.

That’s the job now. For them, and for you. Embrace the change.

Additional reads:

1. AI vs. Gen Z

2. State of the product job market in early 2026 -

Subscribe now

I used to think recruiting into an early-stage startup was mostly a comp problem.

Match what you can with cash, make up the rest with perks, upside, story.

The longer I’ve done this, the more I think that framing is backwards.

The best people I’ve worked with aren’t optimizing for free lunches and stipend spreadsheets. They want to do great work. They care about the craft. They care about who they’re working with. They care about whether the company is set up so they can actually ship work they’re proud of.

If you can create that environment, recruiting gets easier. If you can’t, comp won’t save you.

What’s actually important for senior talent

When you’re trying to pull someone into a pre-PMF startup , there’s a key question folks ask themselves that lives outside of comp and perks:

Will I be able to do my best work there?

Senior, high-caliber people tend to be intrinsically motivated. They don’t need a manager to push them. They don’t need a bunch of processes to tell them what to do. They want a clear problem, a high bar, and a team that takes the work seriously.

That’s why the environment you create matters so much. If things constantly get watered down or compromised, those people will leave.

It’s a double-edged sword.

The same thing that attracts them is the same thing that makes them intolerant of bullshit once they’re inside.

So when I’m hiring, I think a lot about motivations in a practical way.

If someone wants a certain kind of work, a certain kind of pace, a certain kind of culture, and we can’t offer that, we shouldn’t hire them. Even if they’re amazing.

The mistake founders make with “motivation”

A lot of founders try to hire someone and then “motivate” them.

I don’t really believe in that.

I want people who are already intrinsically motivated and have a high degree of locus of control. If someone needs the company to constantly manufacture motivation for them, it tends to break under stress, and startups are basically a stress machine of highs and lows.

Think this is illustrated nicely by Scott Belsky in the Messy Middle:

You shouldn’t ask yourself if you can motivate great people. They’re likely already equipped to do that themselves.

The better question is: are we set up to support how this person works at their best?

If the answer is no, it’s a fit problem.

Early-career hires want something different

This changes depending on where someone is in their career.

If I’m hiring someone in their 20s, I think the best thing for them to optimize for is learning and growth. Cash matters, obviously. But the real long-term comp is whether they’re going to level up here.

If we can’t provide growth, we shouldn’t hire them either because they’re going to leave to find it somewhere else.

The comp philosophy shift we made

We recently changed how we think about perks and comp, and it came from a simple observation: most people don’t care about perks the way founders think they do.

We used to offer a bunch of stipends and benefits. On paper, it looked great. In reality, what most people cared about was their salary. The bottom line number.

So we backed it into salary. Instead of giving people a few thousand dollars in “stuff” each year, we increased salaries and removed some of the complexity.

We also shifted how we split salary and equity.

We used to think about it like 50th percentile salary and 90th percentile equity. Now we’re closer to a 75/75 approach. Higher cash than before, and still meaningful equity, just not as extreme.

The implication is straightforward: newer hires will get less equity than earlier hires did, but their salary will be higher. Over time, we’ll also adjust the existing team where it makes sense, especially based on performance.

The part nobody likes talking about: attrition

Some natural attrition is expected. People change, their lives change, what they want changes. If someone feels like the right move for them is to leave, I’d rather support that than try to guilt them into staying.

That’s hard as a CEO because you’re always balancing “people first” with “what does the company need,” and those don’t always align cleanly.

However, you do need to prepare yourself for it. According to data from Carta, it’s more or less a coin flip that an employee will leave your startup within three years of starting.

But I care about relationships. I want people to leave Clarify and do amazing things, and still want to work with us again someday, or refer someone, or just be part of each other’s lives in the startup community.

Perks are nice. Expectations are necessary.

One more thing we learned the slightly annoying way: perks and policies don’t matter if expectations aren’t clear.

Unlimited PTO is a good example. If you don’t set expectations, you end up with a weird imbalance. Some people take a lot. Some people take almost none. Then you get resentment, or guilt, or confusion, and the team performance gets uneven.

We didn’t do a good job setting clear expectations there, and it sparked internal conversations we should have had earlier.

That’s one of those founder lessons you only learn by stepping on it.

What I’d tell an early-stage founder about perks and comp

If you’re trying to recruit great people into an early-stage company, don’t start with perks. Start with the environment.

• Can someone do their best work here?

• Will they work with people they respect?

• Will the company protect the quality bar, or constantly trade it away?

• Is it clear what “good” looks like?

Then pay people fairly, simplify where you can, and be honest about what you can’t offer yet.

Perks are easy to copy. A real environment for great work is not.

Additional reads

1. Employee attrition - Carta

Subscribe now

I’m not into April Fools stunts. But founders being fools? That I can get behind.

Mine: when I started my first company, I closed a $20k contract that should’ve been $200k. I had no idea how to feel out willingness to pay, no negotiation instincts, nothing. Just said yes and moved on.

Here are five more stories like that - from founders I genuinely respect. Not “lessons learned.” Just the stuff you only figure out by stepping on the rake.

1) Winging discovery calls

Sachin Gupta, Breakout:
“I spent a decade building multi-million dollar startups by moving fast and breaking things. So when I started Breakout, I figured I already knew how to run discovery calls. I didn’t prepare, I just rambled through questions and trusted my instincts.

Our team would start building, only to uncover critical details later that completely changed the solution. I was so obsessed with speed that I neglected the prep required to run good discovery calls, which are one of the most vital parts of a startup’s foundation. Turns out, slow really is smooth, and smooth really is fast.”

This one hurts because it’s so common. You don’t “save time” by skipping prep. You just move the cost downstream and make it more expensive.

2) Breaking the thing that just got built

Jesse Eisenberg, StreetTalk:
“After weeks of trying to adhere our website form to best practices, which included multiple agencies, technical employees, and endless configuration, we finally got it working perfectly. Everyone breathed a sigh of relief.

A few days later, I went in to delete some test submissions and accidentally deleted the entire form. No restore option. Hundreds of real submissions…gone. Live form…gone. Ad campaigns…paused.

I had to personally call every stakeholder to tell them I’d woken the sleeping giant. We rebuilt it in a couple of hours, but there’s a specific kind of humbling that comes with being the one who breaks the thing everyone else worked so hard to fix.”

We’ll give Jesse a pass here as he was trying to make something happen quickly. As they say, “Make mistakes of ambition, not of sloth.”

3) Taking losses personally

Mirko Novakovic, Dash0:
“I’ve founded 3 companies over 25 years, and early on, losing was personal. Lost a deal? I got frustrated. Someone quit? I took it to heart. That emotional attachment to winning cost me a lot…in energy, in relationships, in perspective. Spending years treating every loss like a personal affront took its toll.

Now I know that almost nothing in business is actually personal. You can’t win every deal. You can’t retain every A player. Winning still fuels me, but it no longer defines me. That shift took way too long to make.”

This is the one nobody warns you about. Because your company takes up so much of your life, it’s hard to separate yourself from the emotion of it. But you have to try.

4) Doing everything, delegating nothing

Pranav Piyush, Paramark:
“I’m not great at delegating, and when we started Paramark, that became obvious fast. My instinct was always to jump in and fix things myself. Felt like the fastest path forward. But every time I grabbed the wheel, I became the bottleneck.

The company could only move as fast as I could. Learning to step back, trust your team, and let them iterate – even when you could do it faster yourself – that’s the real work of being a CEO. I’m still figuring it out.”

If you ever catch yourself thinking “I’ll just do it myself,” congratulations, you’ve found your next constraint.

5) Selling to the wrong buyer

Peter Conforti, Good Content:
“For 18 months, we spun our wheels trying to find product-market fit. We were burning through time, money, and morale. We kept trying new offers, new services, new angles, getting just enough traction to keep going.

What I didn’t realize was that we were rowing as hard as we could in the wrong current. The problem wasn’t with the product or the offer. The problem was that we were selling to the wrong buyer. They had no budget, no incentive to choose us over what they already had. Quality didn’t matter. Effort didn’t matter. There was no amount of product development that could save us because we were talking to the wrong people.”

This is the most brutal one because it looks like “keep iterating” until you realize you’re iterating on the wrong person.

If you’re reading this and thinking “cool, I’ve done at least three of these,” welcome. You’re one of us.

And if you haven’t done any of these yet, don’t worry. There’s still time.

–Patrick

P.S. Thank you to all the founders in this piece for sharing their stories!