For years, research was expensive. Whether it was customer interviews, competitive analysis, surveys, or usability testing, every question came with a cost—and time. Teams had to prioritize only the questions with the clearest business case, leaving countless curiosities unexplored.

As the speed and cost of research drops dramatically with AI, researchers no longer have to reserve their time for only the highest-priority questions. They can afford to follow hunches, investigate odd behaviors, and explore around the corners of their product—often uncovering insights that prove far more valuable than the question they originally set out to answer.

This shift is similar to what happened in engineering with open source. When the cost of building foundational infrastructure collapsed, developers stopped spending time reinventing the basics and started experimenting in entirely new directions. AI is creating a similar inflection point for product teams: as the cost of learning falls, the opportunity shifts from answering known questions to discovering entirely new ones.

Nicole has spent her career at companies operating at very different frontiers—from Airbnb and Slack during their hypergrowth years and, most recently, Kalshi. While the products and technologies have changed dramatically, she argued that the defining characteristic of exceptional teams hasn’t.

The teams that win won’t just build faster. They’ll learn faster.

Jagged edges

Nicole describes the biggest opportunities in a business as its jagged edges—the uncomfortable, unintuitive insights that don’t show up in dashboards or quantitative analysis. They’re the observations you only uncover by getting proximate to customers, competitors, and the product itself.

Most teams spend their time optimizing what they already know, but the best teams spend their time searching for what they don’t.

That distinction becomes even more important in the AI era. As more of the routine work becomes automated, the highest-leverage work isn’t writing documents or synthesizing surveys—it’s discovering the next question worth asking and answering.

So what is a super builder?

Today, people often define a super builder as someone who produces dramatically higher output because they’re exceptional at leveraging AI.

Nicole offered a different definition: a super builder is someone who compresses the learning cycle.

AI has enabled teams to write more code, generate more ad creatives, outbound more customers - all in record time. But the really meaningful unlock in terms of impact is when customer feedback and understanding is as continuous as code deployment. When output and feedback are orchestrated together, we can test assumptions earlier, prototype ideas faster, pressure-test decisions, and absorb feedback before everyone else.

In other words, in Nicole’s definition the advantage isn’t output, it’s learning iteration.

Every additional cycle of shipping, observing, and learning creates another opportunity to uncover a jagged edge that competitors haven’t seen.

Finding the jagged edge

Nicole illustrated this with a story from Airbnb.

Her team spent months investing in professional translations for listings. The hypothesis seemed obvious: if guests could better understand listings in their native language, bookings would increase. After six months, however, the results barely moved.

Nicole traveled to Thailand to meet with real humans, where she realized guests weren’t hesitating because they couldn’t understand the listing. They were worried about accidentally offending their hosts by unknowingly violating local customs—whether that meant hanging clothes to dry instead of using a dryer or navigating other everyday cultural differences that made staying in someone else’s home feel intimidating.

The problem wasn’t translation. It was social comfort.

Rather than investing just in localization, the team built products that reduced social uncertainty. They shipped a badge showing someone from your country had previously stayed at a listing. They surfaced reviews from travelers with similar backgrounds. They made it easier to exchange photos before booking.

While optimization work kept metrics afloat, it was chasing after the jagged edges that became the lightning bolt.

The lesson wasn’t simply to conduct more user research. It was to get closer to reality when the data alone stops explaining what’s happening.

AI changes what great builders spend their time doing

Nicole saw the same pattern again at Kalshi—but with one major difference.

AI had fundamentally reshaped how product teams worked.

Competitive analysis, survey synthesis, experiment write-ups, and research summaries—the work that once consumed entire teams—could increasingly be handled by AI. Rather than replacing product thinking, AI had eliminated much of the repeatable work surrounding it.

That gave the team more time to chase jagged edges.

One example came from observing user behavior, in-person, at major sporting events. Nicole leveraged AI to take over baseline insights and duties, so the team could go to a game and look for the unknown unknowns.

Ideas from that trip have gone on to generate major wins (there were more details, but you had to be there!).

Throughout the conversation with Adverb founders, Nicole returned to the same idea: AI increasingly handles the known work, leaving humans to focus on pursuing the unknown.

Instead of talking about building, she posits that product development is moving towards a process more akin to cooking. Building suggests long planning cycles and reducing uncertainty before work begins, where cooking is iterative. You taste as you go. You adjust. You ask your roommate if they think it needs more salt. You respond to reality instead of trying to predict it perfectly in advance.

The strongest AI teams increasingly operate this way: ship, observe, learn, repeat, where speed isn’t the goal but learning is.

What founders should optimize for

Nicole closed with a broader question: if AI changes how individuals work, how should organizations evolve alongside them?

Her answer wasn’t to hire only super builders.

The strongest teams pair people who excel at finding jagged edges with teammates who bring exceptional craft, execution, and strategic direction. One group discovers the opportunities. Another turns those opportunities into polished products. Leadership ensures the company continues moving toward the right destination rather than simply moving faster.

She left founders with three questions to keep asking:

• Where are the jagged edges in your business that no dashboard will reveal?

  • Interview question bank from tried and true HR and tech experts

  • Questions you can ask to find your business’s jagged edges

• Where do you need something deeply customized, and where is “good enough” actually enough?

• How adaptive do you want your company to become—and does your team structure reflect that?

The AI landscape will continue to evolve, but Nicole’s advice felt surprisingly durable.

The companies that build lasting advantages won’t simply be the ones that adopt the newest tools first. They’ll be the ones that learn from the reality of their users faster than everyone else.

The public conversation around AI is often dominated by discussions of models, benchmarks, and capabilities, and the perception admittedly skews negative outside of tech’s inner circle. But spending a day with founders shaping the future of consumer AI experiences offered an entirely different perspective. The most compelling companies aren’t asking, “What can AI do?” They’re asking, “What aspects of everyday life have become unnecessarily difficult, fragmented, or overwhelming? And how can AI accelerate solutions?”

One theme surfaced repeatedly throughout the day: modern life has become an exercise in coordination. Between work, family, school schedules, homeownership, and an ever-growing number of digital systems, people are increasingly acting as project managers of their own lives.

What we saw at Humans in the Loop was a new generation of companies attempting to solve that problem—not by adding another application or another notification to our phones, but by reducing the cognitive burden that accompanies modern life.

Two presentations from Adverb portfolio companies illustrated this shift particularly well.

Indexed Labs: Reducing the mental load of parenting

Liza Gurtin, co-founder of Indexed Labs, opened with three small stories that probably sound familiar if you’re a parent.

• Will drove to the wrong school for pickup after mixing up his kids’ calendars, then rushed across town feeling like he’d abandoned his kid.

• Maddie missed a note about spirit day buried at the bottom of a school email, and her daughter came home in tears.

• Tiffany missed golf camp registration and spent the next week scrambling for a backup. None of these in isolation are catastrophic, but it’s death by a thousand papercuts.

Parents are drowning in admin, logistics, and coordination. Kids today are in more after-school enrichment activities than ever before. Over 60% of school weeks are non-standard. And the details are scattered across emails, text threads, PDFs, school websites, apps, and portals. The bottleneck isn’t access to information; it’s the work of making sense of it.

Indexed Labs is building Otto to take the logistics of family life off parents’ plates entirely. Otto ingests information from the dozens of places parents already manage their lives—school emails, calendars, portals, websites, flyers, and group chats—and turns it into a single, personalized source of truth. Instead of asking parents to adopt another tool, it proactively surfaces what matters and handles everything else. The opportunity in AI isn’t to generate more information; it’s to make sense of what already exists—and to act on a parent’s behalf.

✨ Join the waitlist for Indexed

Homeownership shouldn’t feel like a second job

Michael Mizrahi‘s presentation on Casa made a similar source of administrative burden: home ownership. Owning a home is one of life’s most meaningful milestones, but managing one often feels like a full-time job. Little things break, bigger things require coordination, and homeowners are suddenly responsible for finding vendors, comparing quotes, and making decisions in areas where they have little expertise.

Michael framed the opportunity by pointing out that affluent households already have a version of a solution: a house manager who knows the property, understands the systems, coordinates repairs, and handles problems before they become emergencies. Casa is building that experience for everyone.

What makes Casa compelling is that it starts with understanding. In a single visit, Casa captures a detailed digital model of the home—everything from appliances and lightbulbs and paint colors to real-time property data—so that when something goes wrong, the system already knows the answers to most questions a vendor would ask in order to service the property. A water heater acting up is no longer just a vague issue; it is a specific tankless make and model due for maintenance, still under warranty, with a gas bill that has been creeping up for months. From there, Casa can recommend the right fix and dispatch a vetted professional who arrives already briefed.

That combination of intelligence and execution is what makes the idea feel durable. The goal is not simply to make home maintenance easier, but to make homes easier to enjoy living in.

🏡 Checkout Casa

What made Humans in the Loop memorable wasn’t the technology itself, but the reminder that AI is ultimately a means, not an end.

Kirsten Green, founder and managing partner of Forerunner, captured this idea well in her reflections after the event, writing that AI’s narrative problem is really a proximity problem. Most people experience AI through headlines, social media, and predictions about what might happen someday. What changes opinions isn’t a better argument; it’s spending time with the people building this future and understanding the values and motivations behind their work.

We left Humans in the Loop with a similar feeling. Not certainty about where consumer AI is headed and when, but confidence in the people shaping it and their attunement to the humans they’re looking to serve.

Casa makes home ownership hassle-free. It’s a concierge-like experience that handles the proactive maintenance your home deserves and eases the burden of the ad hoc mishaps that are inevitable when you own the place you live. The team builds a digital record of every aspect of your home — paint colors, appliance models, the shorthand you use to refer to each room — so that you get the convenience of a personal property manager at a fraction of the cost.

There’s a lot of talk in startup circles right now about “selling the work” — a future in which software completes complex projects rather than just facilitating users doing so themselves. Casa is that thesis in action: a consumer experience that feels like an effortless messaging thread, backed by a rich dataset, AI-powered workflows, and the operational muscle to actually see work through to completion across tree trimming, exterior cleaning, appliance repair, and everything else a home throws at you. In the same way that Uber reduced getting a ride to the airport from a multi-step ordeal to a single button, Casa turns ‘my dishwasher is leaking’ from a three-hour research project into an immediately solved problem.

The hardest thing to build in this category isn’t the app. It’s the team willing to own the messy operational layer underneath it. That’s been the most remarkable thing to watch up close. The Casa founders have assembled a group that treats that operational layer as the product itself, not a compromise around it. Every week, something that was impossible the week before has quietly become routine. That compounding is what takes a company from SF and LA to every city where a homeowner lives.

Being an investor of a company you also love as a customer is a rare joy. Being able to watch, up close, the relentless work that’s gone into bringing Casa to life over the past nearly two years has been a privilege. In this case, we mean this quite literally: we’ve been co-working out of a shared office with this team since their founding. Michael York and Michael Mizrahi, co-founders, previously worked in early roles at Uber and CloudKitchens, and bring an operational grind and customer obsession to the product that we’ve felt as customers. It’s a remarkable experience and we’re thrilled that as of today, homeowners in SF and LA can experience the magic. Sign up today!

Congratulations to the Casa team on sharing with the world what early customers and close observers have known for some time now — that Casa is the next in a series of household names that millions of people will soon be aghast they ever lived without.

Casa: Member Success Manager

Outset: Product Designer

Granola: Product Engineer (Full Stack)

Alpaca Health: Chief of Staff

Atlas Network: Chief of Staff

If you’re exploring, or know someone exceptional who is, you can find all open roles across the portfolio here: https://jobs.adverb.vc/jobs

As always, we’re biased, but we think these are some of the highest-slope environments you can find right now. If something here resonates, reach out. We’re always happy to make intros.

How to Approach App Store Optimization (ASO)

Aja frames ASO around five core pillars:

1. Keyword Prospecting: Identifying the right search terms before competitors find them—especially generic or underserved terms.

2. Metadata Strategy: Dialing in title, subtitle, keywords, and screenshots to match user intent.

3. Feeding the Algorithm: Early traction (from ratings, conversions, and behavioral signals) gives Apple quality indicators that improve ranking.

4. Optimizing for Ratings: Ratings meaningfully influence both conversion rate and keyword rankings.

5. Monitoring Results: ASO isn’t one-and-done; data shifts, competitors launch, and search volume evolves.

Example 1: ADHD Calendar

Aja shared an app she built around ADHD-focused productivity that illuminated how to leverage keyword prospecting and employing a metadata strategy to drive conversion:

What Worked

• Early to generic ADHD keywords: She targeted related search terms before the volume spiked.

• Rode a surge in trending interest: Search volume for these keywords grew significantly—Aja had already established visibility.

• Metadata that converted: Title, subtitle, keywords, and screenshots worked together to match user intent.

What Didn’t

• Competitors entered quickly: Once the category grew, newer apps crowded the space and overtook rankings.

• Ratings were a bottleneck: Without early and frequent ratings, the algorithm deprioritized the app.

Example 2: Citizenship Education App

This example highlighted how ratings timing and onboarding design can dramatically affect organic growth.

What Worked

• Guided path to a “win”: The onboarding flow led users to complete a quiz, which included celebratory confetti in the UI, delivering a moment of delight that empower users to feel confident in beginning their journey.

• Ratings request during onboarding: Following that moment of early success, users were asked for a rating, priming them to leave a more positive review.

What Didn’t

• The market was dominated: Existing apps with decade-long head starts held the top generic keywords.

• Highly commoditized search results: Users tended to download the first result—hard to displace incumbents.

• ASA was too expensive to scale: Apple Search Ads couldn’t profitably acquire users in a competitive vertical.

Apple Search Ads: Helpful, but Complicated

Aja emphasized that ASA can meaningfully shape organic keyword rankings—not just paid traffic. But it introduces strategic tradeoffs:

• Should you bid on your own keywords?: Sometimes it prevents competitors from stealing your brand traffic, but sometimes it cannibalizes your organic installs.

• Category and competitor bidding may still influence organic ranking signals indirectly.

What’s Working for Her Today

From across her apps and experiments, Aja highlighted three areas currently driving growth:

1. Targeting generic keywords: Especially underserved ones where incumbents haven’t optimized metadata.

2. Optimized ratings prompts: Triggered at moments of user success, not randomly.

3. Apple Search Ads on strategic terms: Focused on brand, category, and competitor keywords.

4. Launching on Reddit to drive conversion to app downloads

Takeaways for Founders

A few themes from Aja’s talk stood out for early-stage builders:

• You don’t need a marketing budget to win in the App Store—you need the right keywords.

• Ratings timing for users matters a lot—if poorly timed, users will just dismiss the option.

• “Winning” a category is about early identification of rising search volume, not chasing established terms.

• ASA and organic rankings are tightly coupled, and founders should treat them as a single system.

• Distribution isn’t just ads—Reddit, metadata, and product-led loops can do far more for early traction.

Check out other Fields Notes: our series that takes you behind the scenes of the conversations we’re having with founders and operators in the Adverb community. Each article brings a practical, candid perspective from the people building, scaling, and learning in real time.

What makes TikTok hard to ignore is how much it now shapes culture, taste, and purchasing behavior. But, what makes it useful for founders is something simpler: it offers an unusually level playing field. Small teams, new accounts, and low-budget experiments can perform surprisingly well. Below, you'll find the playbook that Aaron honed to scale Le Walk’s following to almost 200k.

Note: Aaron uses TikTok as an example, but all of this guidance can be replicated across short-form platforms, including Reels and Shorts.

📍#1: Be on TikTok

One of the most helpful mental shifts is to think of TikTok not as “social media” but as a matching engine — a marketplace that routes videos to people to maximize their enjoyment and the platform’s economics. In practice, that means anyone can get reach, and every video has a chance.

A practical starting point is simple: spend time on the platform. Make multiple accounts, and use each one to explore a different niche or a different feature.

Every “Tok” — BookTok, ShopTok, EduTok — has its own norms, references, tone, and humor. Explore widely to get a feel for how content varies across communities, as it’s easier to contribute meaningfully when you’ve absorbed what already exists.

As a launching point, Aaron created an account with hundreds of videos and creators spanning every major industry organized in the Favorites section.

A helpful exercise: Trace how creators evolve

Pick a creator who has reached some degree of prominence (roughly 500k+ followers). For example: https://www.tiktok.com/@tanner.leatherstein.

Start by watching their most recent 20 videos. Pay attention to what they’re doing and try to identify a pattern in their content: a repeatable format, consistent themes, familiar pacing, and a recognizable style of value delivery.

Then, scroll all the way to the start of their feed.

What you’ll see at the bottom almost always contradicts what you saw at the top. Unless the creator has deleted their early videos, those first posts typically lack the structure, clarity, and reach of their later ones. They’re experiments. Prototypes. Attempts at figuring out what resonates.

Tanner Leatherstein is a good example. He began by posting educational content about leather craftsmanship — deep dives into materials, tanning techniques, and the nuances of quality. The content was solid for a specific audience of leather nerds, but it didn’t take off.

Then he pivoted. He began a series of videos evaluating designer handbags from the perspective of a calm, nonjudgmental expert. He took the bags apart — literally and metaphorically — and explained what made them worth the price (or not).

Suddenly the audience was hooked.

These videos tapped into multiple engaging themes at once: questions of value and status, the appeal of expertise, clear narrative structure, and the simple curiosity of watching something get dissected. It became a format he could repeat, allowing viewers to instantly understand the value he brought and making it easy for new audiences to join in at any moment and providing established followers a familiar format.

This pattern — wandering until you find a format that “clicks,” then repeating it — is one of TikTok’s most consistent truths. And you only notice it when you take the time to study creators across their entire arc.

🎯 #2: Know Your Goals

Before making anything, it helps to decide why you’re on TikTok at all. The answer differs by company, and it could be some combination of:

• Establish broad awareness

• Announce launches and drive conversion

• Build in public

• Experiment with brand voice

• Promote job opportunities

• Gather user feedback

These goals don’t need to be fixed or precise. You can change them as you learn. But having an intention makes it much easier to evaluate whether something is “working.”

If you post without a goal, TikTok can feel overwhelming. With a goal, it becomes a series of small, structured experiments.

💭 #3: Content Ideation

We can think of content ideation in the following framework:

• Message: the thing you want to communicate

• Angle: the perspective or creative approach

• Format: the actual structure of the video (e.g., selfie, green screen, voiceover, meme template)

This combination makes ideation much more flexible. The same message can be expressed through dozens of angles and formats. A single idea can generate many small experiments.

A useful exercise is to list 3-5 core messages about your company. Then, pair each with multiple angles and formats. Very quickly, you’ll have a backlog of experiments rather than a blank document.

The goal is not polish. It’s exploration. You’re trying to find “content-market fit,” which often looks like one video slightly outperforming your expectations — enough to signal that you should try a few more like it.

🎥 #4. Content Execution

Early on, the biggest advantage you have is speed. Videos that feel informal or lightly produced are not just acceptable — they often perform better.

Most companies overcomplicate this step. You don’t need a content calendar, a complicated review chain, or agency-level production. You need a workflow that lets you make many small bets.

Who makes the content?

There’s no universal answer. Some teams discover that an unexpected person — an engineer, an ops hire, an intern — who naturally has an intuition for short-form video. It’s worth letting several people try.

Alternatively, hiring a creator (~50k followers) part-time can be surprisingly efficient, versus doing influencer marketing and sponsored content. Many brand accounts that feel “effortless” are actually run by small creator teams who understand pacing, tone, and platform conventions.

The algorithm as feedback

One of TikTok’s advantages is how quickly you get signal. You often know within hours whether a video resonated. Instead of interpreting that as judgment, treat it as information:

• Try a shorter version

• Change the hook

• Adjust pacing

• Recut and repost

• Try a new angle

Most creators iterate constantly. Brands can too.

📝 Some Additional Notes

• Comments are data: They reveal what viewers noticed, misunderstood, or wanted more of.

• Repetition helps: Many ideas only work on the second or third attempt.

• Casual is okay. If a video feels slightly unfinished, it often performs better than one that feels overworked.

• Search matters: TikTok is increasingly used like a search engine. Clear text, keywords, and simple phrasing help your videos get found.

For early-stage teams, TikTok can feel unpredictable. But underneath the chaos is a platform with clear patterns: participation matters more than polish; creativity scales faster than budget; and the best results usually come from people who treat it as a series of ongoing experiments, not as a campaign.

Aaron underscored something simple to anchor on: you don’t need to be an expert to start. You just need to be there, pay attention, and keep an open, patient mind to try things.

Check out other Fields Notes: our series that takes you behind the scenes of the conversations we’re having with founders and operators in the Adverb community. Each article brings a practical, candid perspective from the people building, scaling, and learning in real time.

This is our second installment of Fields Notes, a new series that takes you behind the scenes of the conversations we’re having with founders and operators in the Adverb community. Each article will bring a practical, candid perspective from the people building, scaling, and learning in real time.

When Sara Beykpour started building Particle News, an AI-native news platform that uses LLMs to summarize and personalize news coverage, in early 2023, there was no playbook—no dashboards, no prompt libraries, no evaluation tools to measure LLM output quality. Particle had to build nearly everything from scratch just to get their product out the door.

Two years later, the AI tool options look entirely different. Startups can now spin up evals, caching, and fine-tuning with off-the-shelf tools. But as Sara and Jennifer Jang, founder of StoryGold—which uses generative models to design CAD for jewelry—discussed at our recent Adverb Founder Summit, there’s still one fundamental question every founder building with AI must answer: What do you build yourself, and what do you buy?

🧰 Build vs. buy

The internal tools that Sara and the founding Particle team built early on became deeply intertwined with their product architecture. “It’s so tied into our data model and caching systems now that it’s hard for a third party to beat that,” Sara explained.

That deep integration has become a competitive advantage—the control and autonomy compounded over time. Jennifer, however, learned the build vs. buy tradeoff firsthand at StoryGold. They began with off-the-shelf tools, then moved away from them after realizing that they wanted to have more control around how our eval pipeline integrated with the product.

Both Sara and Jennifer reflected how the landscape has shifted. “OpenAI’s latest suite connects evals, tracing, and fine-tuning in a way none of the early tools did,” Sara said. “If you’re starting now, the baseline is so much higher—you’d have to justify building what OpenAI will do better in three months.”

The new tradeoff isn’t just build vs. buy—it’s control vs. velocity. Mature teams with entrenched infrastructure are slower but have more autonomy over their stack because they’re built from scratch, while early-stage startups can stand on the shoulders of integrated ecosystems and sprint. To help navigate this decision, Jennifer offered the advice that: “If inference or evaluation is critical to your core experience, you want to own it early. That’s where your differentiation lives. You don’t want to outsource your edge.”

⚙️ Prompt and eval design: engineering for reasoning

Particle’s internal evaluation (“eval”) system, which determine the quality of output from an LLM, evolved through trial and error into a few practical rules every AI product builder—should know:

1. Tell the model how it will be graded. Telling the model how its output will be evaluated measurably improves performance. LLMs try to predict what a “good” answer looks like. If you’re clear—concise, factual, friendly, accurate—the model steers toward that target.

2. Order matters—ask for the thinking before the answer. LLMs generate text one token at a time. If you ask for the conclusion first, they’ll often “lock in” and then justify it. Flip that order and ask for the reasoning before the decision. “We got much more consistent answers once we changed our prompts to reason first,” Sara explained. “It sounds small, but it changes how the model thinks.” For example:

• Answer-first (bad): “Is this comment bullying? Answer Yes or No. Then explain why.” Result you might get: the model says “Yes” and then invents shaky reasons to match that yes.

• Reasoning-first (good): “Explain why this comment might or might not be bullying. After your explanation, conclude: Yes or No.” Result you’re more likely to get: a thoughtful explanation that actually supports the final decision.

3. Match structure to logic. Output fields should follow a logical flow—evidence first, summary second, verdict last. “Think of it like a reasoning chain,” Sara said. “Each step feeds the next. That makes it much easier to debug when something goes off the rails.”

🧩 Evaluations vs. A/B Testing: Different tools, different use cases

Sara and Jennifer posited a key distinction between testing: evals prevent regressions; A/B tests drive innovation.

Evals act as guardrails, catching prompt breakage and measuring stability. “If your accuracy drops from 80% to 50%, you see it immediately,” Jennifer said. “Evals are like automated tests for your AI—they tell you if you’ve broken something you already knew worked.”

A/B tests, by contrast, measure real-world impact. “They show you what actually resonates with users,” she explained. “We’d run two prompts in production and look at downstream metrics—engagement, speed, satisfaction. You learn things evals can’t capture.”

As Sara summarized: “Evals keep your baseline solid. A/Bs help you climb higher.”

Live comparisons to surface meaningful behavioral differences and customer impact.

🧠 Context windows aren’t everything

As context windows balloon, retrieval-augmented generation (RAG) become essential for accuracy and cost control. If you’re passing in hundreds of thousands of tokens—the basic building blocks of how language models (like GPT) read and generate text—the model gets diluted and confused. But, a good RAG system improves focus and consistency.

“RAG improves accuracy because if you’re passing in hundreds of thousands of tokens, the model is focused equally on everything that you’re passing in. And the model doesn’t do a good job with finding the right context within the context,” Jennifer said. RAG works by feeding the model only the most relevant information—reducing noise, token bloat, and hallucinations.

👀 The bottom line

Two years in, the LLM toolchain is finally catching up to early adopters. Founders today can choose between tightly integrated commercial stacks and bespoke systems with deep in-house control.

The right path depends less on philosophy and more on context—your data, your scale, and your team’s technical leverage.

As Jennifer put it, “We used to build everything because nothing existed. Now the question is: what’s still worth owning?”

We’re kicking off a new series that takes you behind the scenes of the conversations we’re having with founders and operators in the Adverb community. Each article will bring a practical, candid perspective from the people building, scaling, and learning in real time.

For this first installment, we’re revisiting a session from Adverb’s recent San Francisco Founder Summit where Mike, former head of hardware at Square and current founder of an early-stage AI startup, spoke on one of the most misunderstood growth levers in early-stage startups: pricing.

Mike has spent the better part of his career thinking about how price shapes customer behavior—from reworking Square’s transaction model to designing his current startup’s revenue-sharing structure.

Drawing from hard-won experience and missteps, Mike walked through what founders should know about experimenting, user psychology, and the reality of setting prices when you’re still finding product-market fit.

Here are our biggest takeaways.

💡 Pricing is your biggest untapped growth lever

Pricing doesn’t just shape revenue—it defines your margin, customer mix, and market perception. “Pricing should always be top of mind,” Mike said. “I know it’s exciting to think about a roadmap and hiring and who you’re going to close next. But the single biggest lever that you have for how your business is going to perform is pricing.”

At Square, Mike learned that a one-time price adjustment could outperform years of new customer acquisition. When Square moved its transaction fee from 2.75% to 2.65% + 10¢, the change lifted profit substantially.

Founders tend to obsess over new features and funnel metrics—but price affects every customer, immediately. As Mike put it, “A dollar from price growth can be worth three to ten dollars of new revenue.”

🧪 Always be testing

There’s no perfect formula for pricing. Elasticity shifts by segment, time, and context, but Mike emphasized the importance of small, continuous experimentation.

“There is nothing static about pricing. You should be thinking about every new customer coming to your website and seeing your pricing as a testing opportunity where you can get a sense of maximum willingness to pay. As soon as you stop testing, your data gets stale.”

Even in B2B, where volume is lower, every negotiation can be a test. “A single buyer conversation has three or four test occasions,” he explained. Each turn of a contract or quote provides data about where customers perceive value that can help you build conviction around your pricing strategy.

🎯 Price where the value is created

The best pricing models mirror how your product delivers value. “If your customers are very familiar with paying for your type of product in a certain way, it’s going to be an uphill battle to change,” Mike said. “But if you ground your pricing in the value you deliver, that makes the ROI conversation easier for your customer.”

At Mike’s current startup, that meant tying pricing to measurable business outcomes, not vague productivity promises. Early on, Mike tried selling on labor savings, but customers didn’t believe the math: “The believability of pulling out labor costs was very low.”

Instead, they reframed the value around revenue upside: how much faster drive-thru lines moved and how that lifted average ticket size. Once they could show a 4-15% increase in sales, pricing conversations changed completely.

Today, Mike’s startup offers two plans: a straightforward monthly fee and a model that puts their fees at risk, tied to customer upside. Most customers opt for the fixed plan, but they love that there’s an option built around shared incentives.

🧠 Humans aren’t rational—price with psychology in mind

Founders often approach pricing as an analytical exercise. But the decision to buy is emotional first, logical second. “Even when you’re selling to businesses, it’s a human writing the check,” Mike said.

In consumer terms, grocery stores know shoppers only track the price of milk, eggs, and bread—so they price everything else freely.

Your customers have their own “price of milk.” Learn what that is—maybe it’s percentage of headcount cost, a monthly subscription fee, or a comparable SaaS spend—and position yourself relative to that.

Mike’s favorite example: Square’s transaction fee model subtly reduced friction by taking out Square’s fees before merchants saw the funds. “A penny never seen is a penny never lost,” he said. If they’d billed merchants after the fact, the same fee would’ve felt much more painful.

🧭 Local context matters—don’t export your pricing logic blindly

When Square expanded to Europe, the U.S. playbook needed to be re-thought. The company bundled software and payments into one rate—common in the States, confusing overseas. “They were used to buying POS software and payments separately,” Mike said. European merchants compared only the payments rate and assumed Square was expensive.

Lesson learned: pricing is cultural. Before you scale your model across markets or segments, understand how your buyers are used to purchasing similar tools. Sometimes, success means unbundling or adapting to a new frame of reference.

🧩 Help your buyer make the internal sale

Equip the champion inside your customer’s org to justify the spend on your product. That means framing ROI in their company’s language, fitting into existing budget buckets.

At Square, when a major customer kept threatening to churn, they flipped the model: Square gave away payment processing and started charging for software instead. “It completely changed the economics,” he said. “They were thrilled, and it made them more profitable for us.”

Different bucket, same spend—but a win for both sides.

🔄 Price for the long game

Discounting or pricing low to win early deals can backfire if you can’t later reset expectations. “Once you anchor customers at a lower price point, it’s hard to move them,” Mike cautioned. Unless you have a clear story—like a major feature release or shift in value—raising prices later feels arbitrary and can erode trust.

Growth and pricing aren’t opposing forces, as they need a plan to converge. “You can lean into growth,” Mike said, “but be clear about how that evolves into the margin profile you want.”

Pricing isn’t a one-time decision—it’s a living system of psychology, testing, and iteration. Or as Mike put it: “You’ll get it wrong the first time. The goal is just to get to a better place, faster.”

• 👟 20k tours already taken!
• 🏛️ 5 cities live - Paris, Rome, Florence, Barcelona and London
• 🌍 22 languages - the best guides how you want to hear them

We're excited about their vision to reimagine travel and real world exploration with an AI companion. So much so, we co-led their $4.1M seed round alongside Lerer Hippeau, Origins and P72 Ventures. Hope you can take it for a spin on your next walk.

More from Axios about the company and their launch: https://lnkd.in/eEZcACCg

If you're a software development team, you can use Macroscope to understand your code base. It can answer questions like “what did the team get done this week?” and automate code reviews. Check it out: https://prasso.ai/

And it's an especially fun day for us at 💚 Adverb Ventures because this is one of the first investments we co-led when we started our firm. We've known founders Kayvon Beykpour, Joe Bernstein and Rob Bishop for many years having worked alongside them at Twitter after acquiring their prior start-ups. We wanted to be their partner in whatever they built next. And we are literally working alongside them (again) in this new chapter as they’ve been co-working with us in the Presidio.

We're so delighted to be able to share more about it now that it's public. Congrats Kayvon Beykpour, Joe, Rob and the whole Macroscope team!

A bit more about why we're so excited about this opportunity:

Software development is evolving at record speed, and new challenges in overseeing AI-written code are colliding with the age-old difficulties that come with dozens, hundreds, or thousands of engineers collaborating in a shared codebase.

Ask any engineering leader - or product leader, or CEO - and you’ll find that the lack of visibility into how their codebase is evolving keeps them up at night. And ask any engineer embracing the new era of AI-driven tools, and they’ll bemoan the amount of time they spend in code reviews - for human- and AI-written code alike. The promise of these tools is massive, but if it already took tons of manual “accounting” to understand how an engineering organization is contributing to a code base - well, wait until the AI code engineers show up in your code base!

The age of AI-powered software engineering requires a new toolkit, and presents the opportunity for a new level in the developer stack - codebase understanding. Previously locked up in the heads of the most tenured or senior engineers in your organization, Macroscope (launching today, after a beta in stealth for the past year) will synthesize, summarize, and answer what’s happening in your code base, helping engineers and senior leaders be good stewards of the most precious resource your company has - engineering bandwidth.

Founders Kayvon Beykpour, Joe Bernstein, and Rob Bishop bring experience building at scale at Twitter, and through prior experience as the founders of Periscope and Magic Pony. This is a team we at Adverb would have backed no matter what they were building, but seeing them take the lead on a new category which will be foundational for the AI era has been a joy.

We at Adverb Ventures are grateful to have partnered with them from day one along with our co-leads Thrive and GV and welcome Michael Mignano and Lightspeed as the leads for the Series A!

We are delighted to see them emerge from stealth so more companies can learn first-hand what we already know - that Macroscope is a no-brainer for any software development team.

Join us and a group of fellow founders, engineers, and designers in San Francisco for breakfast and coffee on August 5th. Expect delicious bites, great company, and plenty of opportunities to swap stories, share ideas, and connect.

It will be a small, curated group, so please register to secure your spot:

Register

🔎 Adverb Portfolio Companies Showcased at Humans in the Loop

We were delighted to have four Adverb portfolio companies present at Humans in the Loop, a one-day conference in San Francisco hosted by Forerunner highlighting consumer AI startups.

We’ve held the belief that many consumer categories are ripe for disruption since we launched Adverb in 2023. Call it a “vibe shift”, the inevitable implications of AI on every aspect of life, a post-pandemic reaction, or simply the next generation of companies, but consumers are yearning for new experiences and services and founders are rising to the challenge. Early consumer products often feel like secret clubs — and Humans in the Loop had that buzz. The next generation of breakout consumer companies are very likely already hard at work, even if you haven’t heard about them just yet.

Here is a brief rundown of the Adverb presentations:

Sara Beykpour of Particle demonstrated the platform's latest innovations, including their new focused feeds feature that helps users cut through information overload, and the recently launched daily crosswords that bring a playful element to news consumption.

Aaron Sekhri previewed Le Walk's exciting developments, treating attendees to a custom local experience designed specifically for the Presidio location. The preview highlighted how Le Walk is transforming the way people discover and engage with their surroundings through AI-powered experiences.

Mariam Naficy showcased Arcade AI, which leverages generative AI to enable anyone to dream up a product concept and see it manufactured and shipped directly to them, fundamentally changing how we think about consumer goods and creativity.

In addition to our portfolio companies, April joined the investor panel, sharing insights alongside Alfred Lin from Sequoia in a discussion moderated by Forerunner's Eurie Kim. The panel explored the evolving investment landscape for consumer AI and what it takes to build lasting companies in this rapidly changing space.

The strong Adverb presence at Humans in the Loop underscored our commitment to supporting the most forward-thinking founders in consumer AI. Each of our portfolio companies represents a different facet of how AI is enhancing human experiences—from how we consume information and explore our world to how we create and bring ideas to life.

🤝 Get connected to the Adverb Portfolio

We’d like to remind you of two ways to get plugged into the Adverb portfolio. First, the Adverb Portfolio Job Board is continuously updated with open roles from the incredible companies in our portfolio. Secondly, we would like to invite you to join our talent network to hear about and proactively be considered for roles that align with your background.

Launching today: the Adverb Portfolio Job Board, featuring open roles at the incredible companies across our portfolio! Whether you're looking for an early-stage role or know someone who'd be a great fit, browse current openings at companies we're proud to back.

🤝 Join our talent network

In tandem with our job board, we would also like to invite you to join our talent network to hear about and proactively be considered for roles that align with your background. At Adverb, we're passionate about helping our founders assemble exceptional teams that will define tomorrow's industries, and likewise helping our community connect with these opportunities.

🚀 Share your career interests

By taking 3 minutes to complete this form, you'll help us:

• Connect you with founding team and early employee opportunities

• Proactively recommend you to our founders building their core teams

• Connect you with fellow builders in our community who share your interests

🗓️ Join us for Humans in the Loop 2025: Consumer AI Showcase

On June 26th in San Francisco, we’re co-hosting a gathering for builders celebrating the ambition and momentum in Consumer AI alongside leading investors including Forerunner, Union Square Ventures, Thrive Capital, Sequoia Capital, NEA Ventures, and Y Combinator. On stage, 20+ seed-stage founders, including Sara Bekypour from Adverb portfolio company Particle, will present on what they’re building, representing some of the most creative ideas in the sector.

Whether you're a builder, operator, founder, or investor—please register to come get inspired and preview the companies shaping tomorrow's AI landscape.

More event info here: humansintheloop2025.com

Register

🚀 Outset raises a $17M Series A from 8VC

Outset, the AI-moderated research platform used by customers including Nestle, Away, and Microsoft, announced their Series A fundraise with coverage from VentureBeat. Adverb led their Seed round in 2023. Cheers to Aaron Cannon, cofounder Michael Hess, and team!

Aaron and April aboard Outset's well-attended happy hour for the research community during Research Week

Scott brought his unique perspective as a founder, builder, and investor at the intersection of technology and design, sharing valuable insights for founders, builders, and creatives amidst the seismic technological shift.

Their conversation sparked many insights – here are a few of our favorites:

🪄 Algorithms are optimizing for what keeps us watching, not what we truly want to see — creating a new imperative for authentic storytelling.

As brands increasingly flood the content space through the use of gen AI tools and algorithms, which tailor content that drives engagement and not necessarily our true interests, quality storytelling will be at a premium. The bar to stand out from the noise out will only get higher. Scott’s new role at A24 Films, a studio known for distinctive storytelling, is a unique purview from which to leverage the potential for technology to transform the film industry.

A critical question arises—will personalized content become commoditized? Just as consumers gravitate towards branded, high-quality items (like designer shoes or premium cars) when everyday products become ubiquitous, we may begin to demand more meaningful, authentic content. In this landscape, "brandertainment" is emerging—where brands produce story-driven content to stand out from traditional ads. This shift is particularly exciting for user-generated content, as we’re all empowered to tell our stories in new and creative ways.

🫰As a founder, empathy can beat passion.

Being a successful founder isn’t just about relentlessly driving your vision forward — it’s about listening. The best founders balance conviction with adaptability, pivoting when needed and deeply understanding user needs. “If you're the founder that's struggling shoulder to shoulder with your end user,” Scott explained, “you are likely to get some signal that passion for a solution would never give you.”

This wisdom becomes especially valuable when navigating what Scott calls “the messy middle” — a term he coined in his book that describes the volatile place of repeated attempts and failures that every startup traverses. “You want a positive slope of progress where every milestone is a little better than the previous one,” he advised. However, Scott also challenged the “winners never quit” mentality: “If you're thinking, 'I would never have done this knowing what I know now,' then pivot. There's no shame in quitting and trying something different.” Staying in touch with empathy for your end user can help preserve this balanced perspective of knowing when to persist and when to pivot.

♟️AI doesn’t have taste, people do.

As AI makes building software and generating designs easier, the real value lies in making intentional, differentiated design decisions. Great "taste" becomes a competitive advantage. Taste isn't just a skill; it's shaped by experiences—both positive and negative—that inform our design choices. As designers can get leverage on their rote tasks, they can increase impact by harnessing taste, discernment, and decision-making moreso than pushing pixels.

Scott mentioned an example from when Pinterest was struggling with speed issues. Instead of waiting for complex technical upgrades, the design team created colorful screens that loaded before the actual pins. Overnight, users raved about how much faster the site felt, though nothing had technically changed in performance. “It perfectly illustrates how designers can solve technical problems more quickly than engineers sometimes,” Scott shared. “Having designers at the table throughout product development has consistently helped me find better solutions throughout my career.”

👀 Embracing discomfort is a sign of innovation and change

When encountering something new—whether it's a new brand book or UI—if it feels immediately familiar and easy, it may not be pushing boundaries. True innovation often comes with a bit of discomfort, requiring us to challenge existing norms and think differently. This discomfort is a sign that something fresh is being introduced. Finding those moments that require customers to rethink and retrain their expectations is a key to standing out. Cultivating good taste can help discern these opportunities for change, even if they make us feel uneasy at first.

If you’re interested in listening to the full conversation, respond to this email, and we’ll share the recording.

As a Partner at A24 and former Chief Product Officer at Adobe, Scott has shaped how creative professionals and consumers engage with technology. Together, Scott and April will explore the evolution of creative roles, shifts in consumer interaction with technology, and what his latest role at A24 means for the creative world.

Register now to secure your spot!

🚀 Announcing Workmate: the executive assistant you deserve—powered by AI

Have you ever worked with an EA? Have you wanted to but it’s been out of reach? Will Ruben set out to make this level of support accessible to everyone, not just executives. Workmate helps professionals manage scheduling, clear conflicts, and boost productivity—without the high cost of a traditional EA. Now in beta, Workmate is already tackling email support, reminders, and more. You can sign up for the early beta here.

🏦 Accend expands into underwriting with AI-powered financial analysis

Accend has officially launched its AI-driven financial statement spreading and analysis solution. This innovation decreases manual processing times by up to 80%, reduces costs, and accelerates credit approvals with AI-generated credit memos. Early adopters like Slope have already seen a 4x boost in efficiency.

​🖌️ Adverb.vc has a fresh coat of paint!

When we set out to plant our flag in the crowded venture capital landscape, we wrestled with the words we would use to convey who we are. We always knew that the best way to answer “who is Adverb?” would be by showing our work—sharing the list of founders we’d backed. We looked forward to the day we could unveil a website featuring the humans behind our portfolio, and the time has finally come!

These 42 humans, the founders in our portfolio, are breathing life into our vision for Adverb. And we’re happy to show them and their companies to you in our re-designed adverb.vc.

👯‍♂️ Mozi is a private social network for seeing your people more, IRL — no more missed connections!

We are delighted to support Molly & Ev’s vision to help you stay close with your people, whether they're down the street or across the world. Mozi is social without any media - it's the best way to privately share with friends where you are and where you're going.

🎥 Hypernatural is an AI-video storytelling platform that makes it easy to create beautiful, ready-to share videos.

Rebecca Kossnick and Taylor Hughes are using AI to help you turn content into leads. Hypernatural is specifically designed to enhance video content for marketing, enabling brands to create high-quality, AI-powered videos with unprecedented ease and efficiency.

🗣️ Untold is a voice journal that writes back.

Ashwath Rajan is democratizing mental health and emotional well-being access. Untold provides a safe and reflective space for people to process their thoughts and emotions aloud via audio journaling. The AI-powered journal responds with feedback and insights empowering people to navigate their personal journeys with greater clarity and resilience.

Opportunities at Adverb Portfolio Companies:

• Founding Software Engineer at Alpaca Health – more details and application here

• Operations Lead at an early-stage consumer company improving the homeowner experience – more details and application here

• Web & LLM Tools Engineer at an early stage consumer company – email liza@adverb.vc with a blurb on your background if you're interested

It’s all about the people. At Adverb, we live by this mantra. Our early stage investing is centered on founders and now, 18 months in, we have a growing community of founders in our portfolio and friends of the firm who we had the privilege to convene together for our first ever retreat in October. There is no better reminder about how lucky we are to do what we do than to spend a few days together with this community.

Our programming kicked off with Liza interviewing Billy Beane, former GM of the Oakland A’s and subject of Moneyball, on how he tackled baseball in a radically different and data-driven way. On Day 2, we dove into practical topics for our founders: research leader Behzod Sirjani presented on deepening customer understanding, OpenAI GTM Lead Maggie Hott shared her tips for selling, and serial founders Kayvon Beykpour and Jeff Siebert spoke about their founder journeys and navigating a fast-evolving technology and market backdrop. We wrapped up the day by zooming out with Microsoft’s Corporate Vice President of Design and Research Mike Davidson and even got a rapid fireside chat with Vijaya Gadde, board member at Planet, Guardant Health and Mercy Corps and previously long-time Chief Legal Officer at Twitter.

The time away from regular routines made for deeper conversation and strengthening relationships between the members of our early Adverb Founder Community. We are all looking forward to the next one and the stories the founders we’ve backed will have to share then.

Portfolio Spotlights

• We welcomed Michael Gao and Bao Van of Alpaca Health to the Adverb portfolio! They are building an operating system for pediatric autism clinics. Their toolset leverages modern technology (yes, AI and more) to create more efficiency and improve the quality of care and ultimate outcomes for clinics in the US. Michael and Bao were members of the South Park Commons and based in NYC.

• We welcomed Aja Beckett of Shotsy to the Adverb portfolio! Aja has built the #1 companion app for people on GLP-1 medications (e.g. Ozempic, Zepbound). Shotsy helps users better understand and manage the medication, its results, and its side effects, and Aja’s first-hand experience as to the transformational potential for these medications has set her sights on being the digital partner through this journey and beyond. Shotsy is available for iOS and as of this week, for Android.

• Cove AI launched their AI workspace with an amazing reception (TechCrunch, ProductHunt). Cove is the first consumer product we’ve seen which pushes the boundaries of how to work with AI for everyday planning and collaboration. The team behind it is prolific and previously helped build Google Street View and Uber Eats.

Adverb Community and Programs

In October, April spoke at TechCrunch Disrupt alongside Glean CPO Tamar Yehoshua and Amplitude Analytics CEO and Founder Spenser Skates on product-market fit to a packed house at Moscone.

April is Speaking at Primary Venture Partner’s “On the Business” Mastery Day for senior women execs in SaaS on Tuesday, January 21, 2025

The one-day program will include topics on “first team thinking” and P&L Fluency, along with speakers currently or recently in C-level roles at companies including Toast, BILL, Demandbase, and Ticketmaster.

April will share her experience on board room management in conversation with Erica Brescia, MD at Redpoint. Apply here to join.

Adverb will be hosting a founder happy hour in LA during the week of Upfront. Stay tuned for more information in January.

Reflecting on 2024

We are often asked what we’ve learned since we started Adverb. We can say without a doubt that the pace of innovation and market activity is faster than anything we’ve seen in our careers. Every week brings new technical breakthroughs, product launches, and funding rounds which reconfigure the competitive landscape in which early-stage founders are building.

We come away from this year grateful to be in business with a growing community of founders who seek the balance between setting a bold north star and staying the course in spite of the noise with responsiveness to new information and the agility to change their plans along with their assumptions.

Throughout the year, we’ve hosted over 20 events for founders and builders to convene with and learn from each other, and find connection in what can otherwise be a lonely journey in those early days of building new companies. Finally, we are grateful to you, our early subscribers, for tuning in to our updates.

More to come in 2025 – much like the companies we back, Adverb is just getting started. Happy Holidays and Happy New Year – we can’t wait to see what the next year brings!

Best wishes, Team Adverb

Adverb Team News

The Adverb investing team has grown! Please join us in congratulating Liza Gurtin on her transition to a full time investing role and welcoming Gary Pesola to the team.

Liza Gurtin has transitioned into an investing role!

Liza Gurtin has been a key part of the Adverb team since before we even closed Fund 1, and joined us last summer as our first full-time hire as Director of Products, Programs, and Community. Liza has worn many hats – from helping us build out our toolset for managing our investing pipeline to organizing more than 30 community events with founders and builders to build awareness of Adverb and attract deal flow. Even better, Liza has herself attracted exceptional founders to the Adverb portfolio, including an investment we co-led in Q2 yet to be announced.

Liza’s not a first-time investor: she was on the investing team at Matrix for a year prior to joining us and has angel invested in over 20 companies. But her depth of experience on the building side really stands out - she was employee #18 at Robinhood, and early on the Product team at Slack, where she tackled a number of the product and scaling challenges our portfolio companies will face in the years to come. 

We are delighted to have Liza transition into investing full-time, and hugely grateful for the many contributions she’s made to date. Liza is based in NYC, and holds a degree in Symbolic Systems from Stanford. 

You can reach Liza at liza@adverb.vc.

Welcome Gary Pesola to Adverb!

Gary Pesola joined us in September as a full-time investor and has hit the ground running (he literally did a half ironman the weekend before his first day with us). Gary was most recently on the investing team at Conviction looking at all things AI, and was previously at General Atlantic where he studied and covered more mature companies. 

In every one of our conversations and references Gary’s insatiable curiosity, infectious enthusiasm, and ability to build rapport with founders stood out. He also brings an AI-native network and perspective to our team, from everything to products, how we work, and who we know. We’re delighted to have him on our team!

Gary is based in SF and you can reach him at gary@adverb.vc.

New Digs

In addition - we’ve moved to a new location in the Presidio, where we are subletting space. We’ll be hosting events here later in the fall so we hope to see you here soon!

Connect with the Adverb Community this Fall!

🏕️ Special opportunity: Adverb’s Inaugural Founder Retreat

We are excited to offer 2 spots to founders recommended by you, our community. If you have a founder of a pre-seed to Series B company who you think would be a great addition to our first-ever founder retreat, please email April by Tuesday, October 8.

Who: Adverb Portfolio Founders and a few select Friends of Adverb
When: Wednesday, October 16 - October 18, 2024
Where: Sonoma County, California

Invitation Only: email nominations to april@adverb.vc

🫡 Chief of Staff Dinner

Who: Chiefs of Staff or Operations Leaders at Seed to Series B companies
When: 10/10
Where: NYC
Email liza@adverb.vc to sign up

🍸AI Happy Hour

Who: AI Founders, Builders, and Investors 
When: 10/22
Where: Adverb’s office in The Presidio, San Francisco
Sign Up Here

📣 April speaking on “Product-Market Fit: how to find, and keep it” at Techcrunch Disrupt 

Who: Founders, Builders & Investors
When: Monday, October 28, 2024
Where: San Francisco
Sign Up Here ($ registration fee for conference)

Founders – we want a vibe check!

Founders! We are seeking study participants for Adverb’s Inaugural Founder Study, conducted in partnership with Behzod Sirjani of Yet Another Studio and powered by Outset (an Adverb portfolio company). We like to flex some of those old muscles used while building companies like Slack and Twitter in our work as investors, so we are kicking off this qualitative research study to better understand our customers - founders like you! We’d love to hear your perspective as a venture-backed founder in 2024 - in your own words.

If you’re a founder between inception stage and Series C, we want to hear from you. Click here to complete the study.

Portfolio Spotlights

Particle is standing out as the best new source for news. The team at Particle set out in the early days of the Gen AI wave to see how it could be applied to news. Could AI help consumers get beyond bias and go deeper on the news stories they care about most? We’re particularly excited about the audio experience that generates an audio version of your personalized feed of breaking news. Particle announced an $10.9M Series A, led by Mike Mignano (founder of Anchor) at Lightspeed in June (Axios, Techcrunch) . 

You can get on the waitlist for access here.

Piramidal announced their seed round (VentureBeat, Techcrunch). Piramidal is leveraging AI to build a foundational model for the brain to modernize diagnosis of neurological conditions with EEG data. They are partnering with hospitals and research institutions who have an interest in being at the forefront in this area. If you’re connected to neurologists or institutions who may want to learn more, please let us know.

Outset.ai launched Outset Interactive. Outset.ai uses generative AI to give their customers the depth of qualitative research with the speed of a survey. AI asks the questions to unearth insights from study participants, and now with the launch of Outset Interactive, it can be married with real-time product usage data to replicate usability studies critical to software development teams’ understanding of their customers.

GrantScout is now Arwin! Despite being on track to $500K this year, the GrantScout team took stock of what they’d learned from their first year in-market leveraging AI to author government grant applications and were decisive that the opportunity wasn’t big enough to match their ambitions. In just 2 months, they pivoted to an adjacent space - leveraging AI in the taxing RFP process for software vendors which slows down sales cycles and stands in the way of revenue. They’re working with their first design partner already and relaunched as getarwin.com. Founder and CTO Jennifer Jang wrote a great post about the pivot here.

Adverb Community and Programs: Summer 2024

Thanks to friends in our network, we were able to offer portfolio founders support in mission-critical areas through our programming this summer: founder-led sales, and working with the fast-evolving foundational models from OpenAI and Anthropic. 

Adverb partnered with Anthropic to host a virtual brown bag session for founders building on top of Claude to hear more about their roadmap and get real-time guidance on how to get the most out of their platform. Several of our founders have been able to access higher-touch support from the Anthropic team after the session.

Adverb hosted a fireside chat on Founder-driven Sales with Saji Wickramakasara, CEO of Benchling. Founders learned from one of the best about how to close early deals and get early market feedback.

Adverb attended a closed-door session for the VC community with OpenAI. We got insight into their roadmap, an open Q&A with Sam Altman, and got to extend early access to their Real-time  API (highly performant 2-way audio) to a couple of teams in the Adverb portfolio. 

Opportunities to Work for Adverb Portfolio Companies:

• Full Stack Engineer at Accend (AI-powered Compliance for Fintech)

• Full Stack Developer at Arwin (AI for RFPs, formerly known as GrantScout)

• Software Engineer, Backend at Particle (AI-powered news app)

Last month, we hosted a fireside chat with Aaron Zamost (former Head of Comms, Square) and Lauren Goode (Senior Writer, WIRED) in San Francisco to pull back the curtain on how tech press gets made, how some founders manage to break through a noisy and competitive landscape, and common pitfalls for early-stage founders and their companies.

Our biggest takeaway: it all starts with compelling storytelling, and like all things, empathy for the person on the other side of the table (in this case, a reporter). It turns out that reporters are humans, too – understanding their interests and how to develop the kind of rapport that results in the press founders dream of is a long game.

We’re organizing a similar event in NYC later this summer - let us know if you'd be interested in joining.

Read on for a summary of this valuable conversation.

A conversation with Aaron Zamost (ex-Head of Comms, Square) and Lauren Goode (Senior Writer, WIRED) on Navigating the “Tech Press”

The conversation has been edited for clarity.

Build real relationships

Aaron asked Lauren how many pitches she gets every month. The answer wasn’t pretty. “Hundreds, if not thousands. I have 34,000 unread emails in my inbox.”

So as one of those hundreds (or thousands) of people pitching a top tier reporter, how do you get Lauren’s attention?

Lauren: “If you're a startup founder and you're thinking of hiring [a communications person] internally, or you're thinking of outsourcing PR to an agency, you want to make sure that they are people who can get into reporters’ inboxes, and not in a sales driven way. It's got to be a direct relationship, like ‘I'm in this person's iMessage or Signal. We talk regularly. We go for coffee, we go for lunch. We go for lunch for no reason sometimes. And I know how to get in touch with them, and even if I’m not going to place a story, I’m at least going to get helpful feedback.’”

You don’t always have to pitch something

You might have better luck with a reporter if you’re up front with them that you don’t expect them to write.

Aaron: “I’m a big fan of the no-pressure conversation. I worked with an AI startup, and I told them, ‘There are a thousand AI companies, and reporters can only write about so many of them.’ So I called a reporter I know well, and I said, ‘I know you're trying to get the lay of the AI land these days, learning more about AI generally, and funding, and valuations, and hiring.’ And I said ‘I've been working with a company and I think they have something to say that will be different, and will you give them 25 minutes?’ That was my whole angle. Like, ‘There's no pitch here. I just think when you talk to them you're not going to hear the same things you hear from everybody else and it might be useful to you.’ And she agreed and took the meeting, and now the company has a relationship with a reporter at a major publication.”

Set realistic expectations and take wins where you can

Just because you think you’re interesting (and you very well might be), it doesn’t mean a reporter is necessarily going to write a story about you, especially when you’re small. That’s OK! Communications and storytelling are a long game. Be happy and willing to participate even if you’re not going to be the headline. 

Lauren: “When you’re a smaller startup, even a midsize startup, you're probably not going to get a big feature story yet. Of course you would love that, but if you temper your expectations and realize that even if you are just mentioned in a bigger story as a company example, or your quote is included in a bigger trend story, it's still really good visibility for you. And, I'm not sure that all PR people realize that.”

Know what reporters need, not just what they’re interested in

Too often founders and comms people associate a reporter with a specific beat, and think about how best to position themselves within that beat. But reporters are curious people, and they are constantly looking for new and different angles to pursue. Be willing to go “off-script,” and see how you can help the reporter, not just how the reporter can help you.

Aaron: “Often people will pitch a reporter by saying, ‘You recently wrote a story about Topic X. My company is also involved in Topic X, and we have a lot to say about it. Obviously you would be interested in talking to us, right?’ And most reporters think, “Huh? I just wrote about that thing. Why would I write about it again so soon?’”

Lauren: “It’s maddening. The likelihood that I'm going to take your company's [pitch] when I literally just published a story about it, it's slim to none.”

Aaron: “But isn't it intuitive that I would say, ‘Well, you cover wearables, and I am a wearable fitness company, so obviously I would want to reach out to you about that right?”

Lauren: “Yes but here's an example. I had a call with someone who is a startup founder, who's building something in generative AI. And I'm probably unlikely to cover his company directly, but he was so willing to just talk shop about the industry. Like, he was willing to talk about bootstrapping and that’s something I'm very interested in right now as a potential story. And he was willing to talk about the funding environment and some of the challenges of that right now.”

Aaron: “What you're saying is that he was malleable to your interests.”

Lauren: “Yes, at the end I said I thought to myself, I'm going to reach out to him in the future because he was just willing to kind of jaw, and I really like that.”

There’s always an angle

Be creative. Sometimes the best way to pitch your company is to tell a story about something else instead.

Aaron: “A woman on my team at Square led our Hardware comms and we were releasing the new Square register, which was fully integrated, [no longer required an iPad] and ran Android. And she said, ‘I want to pitch Lauren. She would be perfect because she writes device reviews.’ And I said, ‘There is no way Lauren is going to write a review of a hardware point of sale. She's writing about consumer products, like goggles, she's not writing about the register.’ But then we thought about it, and I said if we were going to try, here's what I think we should do. And we pitched ‘The best Android tablet on the market is a cash register.’”

Lauren: “[The pitch] was totally wise. And it's a story that I wrote and it blew up.”

Aaron: “Because that story was not ‘Square announces a new thing.’ Instead it was, well if we take a shit on Android tablets, I’ll bet a reporter would be super into that.”

Lauren: “It was a beautiful tablet. It looked like a Macbook, with an aluminum unibody, but yes it was Android, and that’s what made it more interesting. And then I would go to coffee shops around San Francisco afterwards [and see the register] and be like, I wrote about that.”

And finally: offer exclusives, not embargoes

Lauren: “The fact that the White House, or medical journals, or the IMF use embargoes all makes sense in the context of national security issues or potentially consequential health information or major market-moving information. But when you're a startup no one's ever heard of and you do an embargo blast to a dozen different outlets and think your embargo is sacrosanct, it might even result in less interest.”

Aaron: “Exactly. This is why I’m a fan of exclusives. Companies often want to pitch a bunch of reporters with the hope that they can get one or two to write, but I always tell them, if I could guarantee you that one top-tier publication would write, wouldn’t you take that? Instead of the risk that no one writes because they think someone else will?”

Opportunities to Work with Adverb Portfolio Companies:

• Multiple roles including non-technical roles at Particle! Particle, the AI-powered news app, is growing! Now hiring for an SRE, a Backend Engineer, a Community / Editorial Lead, and a News / Media Partnerships Lead.

• Founding AI Engineer: Build the future of video generation. Details here.

• Founding ML/AI Engineer at Workmate Labs: Build the AI executive assistant for work. Details here.

• (Unpaid, but fun!) Minecraft players: a stealth company is looking to connect with adult Minecraft players who would be interested in sharing user perspective and trying an early version of their product. Reply directly if you’re interested.

• Neurologists: Piramidal is looking to connect with neurologists within research hospitals. If you know a neurologist interested in connecting a team building a foundational model for the brain, we’d love to introduce them.

• General Opportunities: If you are a builder interested in roles at an early-stage company, share your information with us by replying to this email. We will keep you in mind when our portfolio companies are growing.

Admiringly,

Team Adverb 💚

Helping founders access capital is a mission near and dear to us as investors — but beyond venture capital, there is a massive source of non-dilutive capital many business owners overlook: the US Government. Felicia Chen and Jennifer Jang, co-founders of GrantScout, have set out to make it easier than ever to find, apply for, and win government grant funding. We’re delighted to announce we’ve co-led their seed round and are partnering with them on this mission.

At Adverb, we back extraordinary founders in the earliest stages of building breakout companies. We met Felicia and Jennifer in the Fall 2023 HF0 incubator batch, a program where repeat founders live together in SF for an intensive period of early company development. Their combined experience at companies like Waymo, Anduril, and Apple and the company’s early traction with grant applicants stood out to us.

Typically, applying for a government grant is tedious and time consuming. It can take hundreds of hours and even require hiring a specialized grant writer to complete the cumbersome government driven process. Grantscout leverages AI to simplify this. With a basic description of a company, their AI powered search engine can identify specific grants which may be applicable for a given company. And with additional company information, their fine tuned LLM can draft the application while still leveraging the expert final review of an in-house grant writer.

GrantScout is starting with a focus on Small Business Innovation Grants (SBIRs) — a source of $3.5B awarded annually and ultimately aims to tackle the much larger government grant and contracting opportunity.

GrantScout is already being used by customers building innovative products in healthcare, defense, climate, robotics, etc from early stage to growth stage. And 95% of early customers have already come back for a second grant application. Check it out yourself to see if your team might scout a match!

Sara Beykpour and Marcel Molina met at Twitter, where they were early engineers on the service before it was really clear what it was for — but it was clear that users couldn’t get enough of whatever it was (as evidenced all too often by Twitter’s infamous outages and Fail Whale). How early was it? One of the initial features Marcel helped ship was the first version of Retweet, and Sara built the first version of Twitter for Android (and later, for Vine, Periscope and Secret).

We had the opportunity to work with them there and observe their technical and product leadership up-close. And together, we and other early employees along with Twitter’s early users figured it out — and Twitter became a critical part of how we got information about what was happening around the world, in real-time. Since that time, Sara and Marcel have held senior roles in product and engineering at Twitter and Tesla respectively — they’ve been a part of and seen what it looks like to build something brand new that becomes mainstream.

Now, more than a decade later, we’re in the midst of a tectonic shift. AI is cracking open endless new possibilities for how information is consumed and shared and deeply entrenched consumer habits are shifting — products like ChatGPT and Midjourney have grown to massive scale with unprecedented speed. And this is just the beginning.

So when Sara and Marcel set out to start Particle — an app focused on giving users more powerful tools to get their news, powered by AI — we jumped at the chance to partner with them. Their work at the forefront of social innovation, and the technical challenges they’d faced and overcome while running the world’s largest real-time service through rapid growth, make them uniquely suited to see the gaps that even still exist in getting reliable news about the topics important to you.

Sara and Marcel have been quietly building — the product as well as their team — and are hiring for a Senior iOS engineer. If you’re eager to try the product, join their waitlist.

This one was especially fun for us as it’s the first investment we made out of our first fund — Adverb I. We are joined by Kindred Ventures, Ev Williams, Scott Belsky, and others.

Hear more about their vision, straight from the source.

April and Jessica