Jay Choi, CEO of Typeform, joined The Product Podcast to walk through how he’s answering that question:

The tension Typeform faces isn’t unique to them, so here are 5 lessons for AI Builders from Jay:

Defense before offense

“For our traditional forms business, we think about it as, how can we be defensive? And so we really put it on ourselves to create that same conversational interface into our core forms business.”

The first move Typeform made in response to the AI threat was to harden the core. The defensive thesis was simple: if someone could replicate your product’s basic functionality in a chat window, you needed to bring that same conversational interface inside your own product first. Typeform built an AI form builder that pulls in a URL, draws on millions of data points, and generates a contextualised form from a plain-language prompt.

The happy churner problem

“We call them happy churners. People are not mad at the platform. They just use it, and then they are done with it, and then three months later, they use it again, and then they turn it off.”

Traditionally, churn users are seen as dissatisfied, but Jay introduced the concept of “happy churners” – users who weren’t unhappy with the product, they just didn’t need it anymore. They’d complete a project, cancel, come back three months later, and repeat the cycle. The insight was that churn here wasn’t a product quality problem but a symptom of the use case depth problem.

Typeform wasn’t desperate for better onboarding, a cheaper plan, or better customer service. They needed to find more applications beyond their traditional product.

Triangulating on use case bets

“Research and growth tended to overlap, and so we did a bit of triangulation there, where we could get a lot of product depth with similar investments across our platform.”

Typeform narrowed from three possible vertical bets – growth, research, and talent – to two, with feature overlap being the deciding factor.

Growth and Research flows shared infrastructure requirements. Growth and Talent did not – talent required anonymity architecture, while growth required enrichment and automation. Building both simultaneously would have meant divergent platform investment with compounding complexity.

Breadth as an accidental moat

“The farther we go along the product breadth, it just seems to touch more surface area that feels more challenging and not worth it to replicate. Breadth matters now more than it did before.”

Jay’s reframing of product breadth is counterintuitive.

Conventional product strategy pushes toward depth – go narrow, go deep, own one thing completely.

Jay disagrees. In a world where simple, discrete use cases can be vibe-coded out of existence, the breadth of workflow coverage becomes the moat. The more surface area a product touches – form, enrichment, lead scoring, automated follow-up, CRM integration – the less worth it the alternative becomes.

Model agnosticism as product strategy

“We want to be model agnostic so we can, one, always choose the best model, and then two, some models are better for some use cases within our own product, and we don’t have to all-or-nothing our entire platform on any particular use case.”

Typeform made an early commitment to being model agnostic – and it wasn’t primarily a technical decision, but an important strategic move.

The reason is simple: the pace of model releases makes any fixed bet a liability. Different models already serve different use cases within the Typeform product. And without an AI observability platform, you can’t even answer the question of whether switching models actually improved anything.

Key Takeaway

Jay described a kind of strategic honesty that's harder to practice than it sounds. Typeform started by acknowledging they wouldn't win every use case – and built from there.

That shaped every decision downstream: which verticals to bet on, how to position AI without confusing customers, how to price closer to value without overcomplicating the model.

For product leaders navigating the same pressures, that starting point might be the most transferable thing from this conversation.

🔗 Watch the full episode on our website

In a recent conversation on The Product Podcast, Ajit laid out what it actually looks like to build a mission-driven product at scale when every major competitor is a trillion-dollar company. The conversation covered open source strategy, the economics of a nonprofit-owned for-profit, AI agents, and why the choices browsers make right now will shape whether an open internet survives the next platform shift.

Here’s what senior PMs should take from it:

1. Competition as a product principle

“Competition is probably the most important part of making sure that there aren’t gatekeepers that prevent access from an open internet.”

Firefox didn’t survive because it out-resourced Chrome or Internet Explorer.

It survived because it understood that competition itself is the mission. When Microsoft pulled engineers off Internet Explorer – because a rich web ecosystem didn’t serve its core business – it was Firefox that pushed Web 2.0 forward, making Gmail, Google Maps, and Facebook technically possible.

2. Open source as a trust architecture

“In order to build trust, you have to be able to prove what you believe to be happening to actually be happening.”

Firefox being open source isn’t a philosophical stance but the mechanism that makes trust verifiable.

Ajit draws a direct line between inspectability and credibility: if users can’t verify what a browser is doing, they’re taking its privacy claims on faith.

The Anthropic partnership, which involved Anthropic using the Firefox code base to harden the browser against security vulnerabilities, started not with a corporate agreement but with one engineer at Anthropic who believed in an open internet and started contributing.

3. Defaults are a values statement

“Our default experience is something that we are very unique in, is that we optimize for privacy. And so it is like by default, if you do nothing and you just download it and you set it up, it is optimized towards the more private side of the spectrum.”

Firefox ships with privacy-optimized defaults.

Power users can access thousands of configuration options, and a settings redesign is coming to reduce friction for users who want to tune their experience.

But the baseline experience is set deliberately – and that choice communicates Mozilla’s priorities more clearly than any positioning document.

4. The agent disruption most product teams aren’t ready for

“You just don’t end up in a world where there’s a single agent that’s owned by a single multi-trillion dollar company that’s then deciding what restaurants you’re able to go to.”

AI agents don’t consume the web the way humans do:

They don’t watch ads. They don’t generate pageviews. They don’t need a visual interface.

The potential upside is real: agents could remove middlemen (his example is food delivery platforms taking 30% margins) and give individuals direct access to the open internet in a way that hasn’t been possible since AOL’s closed garden failed.

The downside is equally real: if agents consolidate around one platform owned by one company, the result is a different kind of closed garden.

5. The double bottom line problem

“Revenue is important to fulfill our mission, but then the mission is also important to fulfill for the sake of the mission itself.”

Mozilla’s structure – a for-profit company wholly owned by a nonprofit foundation – is unusual by design.

The for-profit enables competitive behavior that a nonprofit structure can’t.

But revenue isn’t the goal – it’s the enabler.

Ajit describes it as a “double bottom line”: revenue matters because it funds the mission, and the mission matters for its own sake.

That means Firefox makes deliberate choices – like making ad tiles fully optional and transparent – that a purely profit-driven browser wouldn’t make.

Key Takeaway

What Ajit describes at Mozilla is a bet that users still care about who controls the infrastructure their digital lives run on – and that a product built with genuine transparency and choice will compound value in ways that pure revenue optimization can't replicate.

Whether that bet pays off at scale is still an open question. But it's one every senior PM should be thinking about, because the same structural tensions exist inside almost every product organization with a genuine mission.

🔗 Watch the full episode on our website

Linear is running a different experiment.

With 25,000 paying customers – including Fortune 100 companies – and a team of 140, COO Cristina Cordova has had to think carefully about what actually drives scale and what is just the appearance of it.

In the new The Product Podcast episode, she gets into hiring, brand, enterprise go-to-market, and what it means to build a product development system for an era when agents are doing a meaningful share of the work:

1. Small teams are a strategic choice

“You can keep a really small team and have a very large, very impactful business. A lot of people ask, “how big you are in terms of company size and of employees?”, but they’re not really talking about how many customers you have, what your revenue is, and how fast you are growing.”

Linear has stayed lean on purpose.

There’s an assumption that headcount and business impact move together. Linear’s operating model challenges that directly. The question Cristina’s team asks is not “what does a company at this stage normally look like” but “is this a role that is absolutely needed right now?”

Keeping a high hiring bar naturally slows growth, which is the point.

2. Work trials are important because they reveal how candidates think

“It’s not just about the end output. I would say that’s maybe half the value of understanding “can this person produce great work?”, but it’s also in how does this person collaborate, communicate, what happens when you’re in a room and we’re intentionally disagreeing with you.”

Linear runs paid two-to-five-day work trials for every hire.

The output of the project is only part of the signal. What Cristina’s team is actually evaluating is how the person collaborates, whether they push back appropriately, and whether they update their thinking when challenged.

The process is intentionally strenuous – partly to find the right people and partly to filter out candidates who are just collecting offers.

3. Going upmarket requires a new messaging layer

“You have to really shift over time your value-based messaging. So, “what is the value that Linear itself is providing?”, and then “how do we talk about that?”, not just to say “Linear offers this value”, but ideally have other customers who are also sharing their own stories.”

Linear started with feature-forward messaging – keyboard shortcuts, dark mode, a product you will actually love using.

That worked for early-stage startups; it does not land with a CTO at a 10,000-person company weighing the cost of migrating away from an established tool.

The shift to value-based messaging – efficiency gains at the organizational level, agentic coding support, build velocity – requires customer proof that takes time to accumulate.

4. Enterprise sales require more than a good product

“We’re not typically going and embedding with a customer and saying, “we’re gonna touch your code”. But we can demo really deep integrations and deep use cases with a lot of the tools that you might be using in conjunction with Linear. And that usually gets people enough of the way.”

Getting through procurement is not the finish line.

Cristina’s team has built out Solutions Engineers who can demo complex integrations and prove that migration is achievable before a customer commits.

The guiding principle is clear: show the customer how easy the switch can be, but do not become a services company in the process. Knowing that line – what you will do for a customer and what you will not – is an operating decision as much as a commercial one.

5. Non-engineers are becoming the fastest-growing AI power users

“I feel like if you fully delegate all of your thinking around AI, you’ll have no native thinking of how these products actually work on their own.”

The data from within Linear’s own platform is telling: the sharpest growth in agentic feature adoption over the past year is coming from people in non-engineering roles.

PMs in particular are using Linear Agent to do things they would previously have needed engineering support for – converting meeting transcripts into issues, generating PRDs, triaging backlogs.

Cristina herself built a competitive analysis tool in Codex to independently verify a market research claim.

Her point is clear: if you fully delegate your AI thinking, you won’t have any idea how these tools actually work and what their capabilities are.

6. Agentic workflows are the default

“We’d rather teams focus on building rather than managing Linear.”

Linear’s positioning shift – from “the issue tracker you’ll actually love to use” to “the product development system for teams and agents” – reflects a genuine product philosophy.

The triage intelligence feature, the agent integrations with Slack, Gong, Intercom, and Zendesk, and the forthcoming ability to run any popular coding model from within Linear all point to the same bet: the product development workflow is being rebuilt around agents, and the tools that treat that as an add-on will lose to the tools that treat it as the default.

Key Takeaway

Linear is very intentional. They have made deliberate choices to stay lean, to hold a high hiring bar, to evolve the messaging without abandoning the original customer base, and to build agentic workflows into the product's foundation rather than layering them on top.

None of those choices is easy to hold under competitive pressure. That’s why Cristina’s insights are worth a listen.

🔗 Watch the full episode on our website

Quality gates have moved. They no longer live in a PRD or a Figma file – they live in working code. This is how the AI-native-from-day-1 teams build. It’s uncomfortable. It’s fast. And it’s increasingly hard to argue with.

Six insights from Meaghan that are most valuable to AI Builders:

Titles signal a specialty vs the scope of work

“I think of titles a little bit as a specialty you might bring to the team, but it doesn’t delineate what you can contribute to the project.”

Anthropic organizes work in small pods of three to five people. The makeup varies – engineers, designers, PMs, in different ratios depending on what’s being built. Everyone contributes across functions. Meaghan pushes code. Engineers make design calls. Nobody waits for permission to contribute outside their lane.

The quality gate has moved

“We’ve now pushed that decision-making into live working code. You need to be using it yourself. You need to experience the product in its actual workflow and that’s when you make a decision on quality because it’s so much more reflective of what the actual experience is going to be.”

This is the shift that will be most uncomfortable for experienced product and design leaders. At Anthropic, quality is no longer evaluated at the PRD stage or in Figma review. It’s evaluated in live, working code – experienced in its actual workflow context.

Internal adoption is not a soft signal but a hard launch gate

“We look for a specific number of internal daily active users who are truly adopting these products into the workflow. And once we see that growth, then we’ll release it externally.”

Claude Code was not released externally until the team saw a specific number of internal daily active users genuinely incorporating it into their workflow. Customer zero goes beyond being a philosophy at Anthropic – it’s a measurable threshold with real consequences for when the external door opens.

PLG and Enterprise are not sequential phases

“We’ll have individual developers who are using the product in their personal projects and their side projects and then they’ll become advocates internally for expanding Claude Code usage within their organizations.”

Claude Code reached large enterprise customers primarily through individual developers who adopted it in personal projects and became internal advocates. The product’s architecture – hackable and secure out of the box – meant it cleared enterprise procurement without heavy customization.

Bottoms-up momentum did the work that a top-down sales motion would typically require.

Token usage is a good indicator but not an ROI metric

“I still go back to the original impact metrics like what is the adoption, what is the retention, what is the revenue – those all still hold true no matter if you’re using AI or not and those are the things we should really be looking at.”

The pressure to invent new AI-specific success metrics is real, and mostly misguided.

Meaghan’s take: zero token usage is a warning sign, but chasing high token counts is equally misleading. The original metrics – adoption, retention, revenue – still apply.

Everyone is managing agents now

“I see all of my team and a lot of individual contributors at the company like mini managers right now because everyone’s managing a fleet of Claudes that are also working.”

Meaghan’s framing of individual contributors as “mini managers” who are each running a fleet of Claudes is one of the most impactful metaphors.

The management skill set – setting direction, reviewing output, course-correcting – is no longer reserved for people managers. It’s the default mode of getting work done.

Key Takeaway

The willingness to make decisions later than feels comfortable, in working code rather than in documents, and to let the product itself be the arbiter of quality is a moat. It’s how AI-native companies operate, and that’s a significant shift in both culture and processes.

🔗 Watch the full episode on our website

Today, his much-anticipated second book Incorruptible is hitting the shelves in the US. Below is an exclusive recap of the episode’s best moments.

Fifteen years ago, Eric Ries gave builders a language — MVP, build-measure-learn, validated learning — and the product world reorganized itself around those ideas. The Lean Startup sold 2 million copies. The concepts it introduced became so pervasive that people started complaining about them, which, as Eric notes, is exactly how you know something has won.

But winning created a new problem. The companies that the generation of builders created — trustworthy, high-integrity, genuinely valuable — became targets precisely because of that value. And nobody had taught the founders how to protect them.

His new book, Incorruptible, is the blueprint for fixing it.

The best practices are pulling you in the wrong direction

“I feel like I taught them how to build things worth protecting. But I didn’t teach them how to protect them. I didn’t even know they needed protection.”

Eric’s central provocation is that most of what founders are taught as best practice — shareholder primacy, quarterly optimization, conventional governance — systematically undermines the integrity of what they are building.

It is a structural feature of modern capital markets that has been normalized to the point of invisibility.

For PMs: The product roadmap is not the only thing that needs a long-term orientation. Governance, incentive structures, and decision rights need the same scrutiny you would apply to product strategy.

Why Eric never says “stakeholder”

“Who would you rather die than betray? That’s really the question.”

Eric made a deliberate choice to strip modern business jargon from the book: no “stakeholders,” no “culture,” no “ESG.” Instead: fiduciary, ethos, character.

The reason is precise. The word “stakeholder” has become so diffuse it gives everyone permission to feel good while committing to nothing. A fiduciary relationship has legal and moral teeth.

For PMs: When your team talks about serving users or employees, ask whether those commitments are encoded anywhere that would survive a change in leadership — or whether they exist only as slide copy.

Harder is easier

“Every time there’s a difficulty, they feel like that overlap zone is going to be a place where they’re going to find a breakthrough. And so they use these three critical words. They say, figure it out.”

The leaders Eric studied most closely — including Jim Sinegal at Costco, Yvon Chouinard, Matthew Prince at CloudFlare — shared a common habit: they treated constraint and conflict not as problems to minimize but as signals pointing toward a breakthrough.

When someone said “you can’t have quality and speed,” they said “figure it out” and built organizations that eventually proved it was possible.

For PMs: Splitting the difference between two competing priorities often produces mediocrity in both directions. The better question is what would have to be true for both to be achievable.

Values on walls are not a governance system

“The people who put asbestos in baby powder walked by the credo every day on their way to work. That said, we put patients, doctors, and nurses first. How is this possible?”

Eric brings the example of Johnson & Johnson — a company with one of the most famous credos in corporate history, inscribed on 10-foot limestone blocks, still in the lobby today. For decades, they knowingly contaminated baby powder with asbestos and internally documented efforts to hide it.

It shows how product principles that have never been tested by a real decision — a feature killed, a revenue source declined, a deadline missed to protect quality — are not principles. At best, they are aspirations.

Individual contributors have governance leverage

“You don’t understand when you’re climbing the corporate ladder, how much the company is obsessed with what you will or won’t do.”

One of the most practical moments in the conversation was Eric’s advice to people who are not founders, not on the board, not in the C-suite: the builders in the middle who want to push their organization toward integrity but assume they have no standing.

His argument: leadership is constantly watching what employees will and won’t do. Asking one clear question in a job interview like “is this mission in the corporate charter?” can make the entire C-suite wonder and change policies.

For PMs: Even if you’re below CPO title, you can already start thinking in governance terms. Practicing mission-level framing now, before you have authority, is both how you drive change and how you earn the positions where more change is possible.

Watch the full episode on your favorite platforms:

Product School website

YouTube

Spotify

Apple

The board wants an AI strategy. The CEO is asking for ROI.

And somewhere in the middle, a working group is explaining why they cannot move forward yet because the data is not clean enough, the governance framework is not finished, or the right talent has not been hired.

Baris Gultekin has heard every version of this conversation. As VP of AI at Snowflake, he works with some of the largest data-driven organizations in the world, and he has a clear-eyed view of what separates the companies that are actually transforming from the ones that are perpetually preparing to transform:

“There is no AI strategy without a data strategy”

AI feeds on data, and fragmented data produces fragmented results. Before an agent or model can be trusted to surface accurate insights, the underlying data has to be consolidated, governed, and semantically enriched. Platform decisions and AI decisions are not separate roadmaps.

Piece of advice: If your AI workstream does not include ownership of a data infrastructure milestone, it is incomplete. This is not someone else's problem to hand off.

Context goes beyond just data – it is a shared reality

“Context in my mind is a shared reality for agents to operate on. You first need to have broken down all the data silos so that the agent has governed access to all of your data. The second, you need to have the business semantics next to that data.”

Baris defines context as three distinct layers working together:

1. Raw data itself

2. Semantic layer that explains what the data means and how key metrics are calculated

3. Documented workflows that describe how the organization actually operates.

An agent with access to data but not to business semantics will produce answers that sound plausible but are wrong in ways that are hard to catch.

Piece of advice: When scoping an AI product, ask explicitly which of these three layers exists and which needs to be built. Missing layers are real accuracy risks.

Scope governance before you scale it

“Start small, to start with what are the core set of use cases that will move the needle in terms of ROI. For those kind of small set of use cases, then we help them build a semantic model and tell us for this use case, what are the business semantics, and focus on that.”

Waiting for enterprise-wide governance before starting is a strategy for never shipping. His recommendation to Snowflake's customers: identify the highest-ROI use cases, build the semantic model and governance layer for that specific scope, prove value, then expand. Perfect is the enemy of started.

Piece of advice: The AI initiative you are waiting to launch until governance is "done" should be reframed as the pilot that helps you define what governance needs to look like.

Top-down mandate meets bottom-up access

“What moved the needle for the whole company was one, top down from our CEO, having very, very clear mandate that this is not an optional thing. We need to have the whole company operating differently.”

Snowflake's internal AI transformation is a great example. The shift that moved the needle was not one thing – it was two things happening simultaneously. A clear CEO-level mandate that AI adoption was not optional, combined with direct tool access for every employee through their internally built Cortex Code agent. The mandate created urgency; the access created agency.

Piece of advice: If you are leading internal AI adoption, check which of these two ingredients is missing. You likely have one but not the other.

AI as a revenue driver vs a cost lever

“It wasn’t a productivity gain only, it was essentially a series of products that we built to increase the value that we’re offering to our customers and thereby drive more revenue.”

Because Snowflake operates on a consumption-based model, AI created a direct revenue mechanism:

More data processed → more insights generated → more value delivered → more consumption.

The AI capabilities made the platform more useful to more people, which drove more usage.

Piece of advice: When building the internal business case for AI investment, the efficiency argument alone is too narrow. The stronger case is the expansion of value surface area – new users, new use cases, new revenue.

Key Takeaway

AI transformation is less of a technology problem and more of a sequencing and scoping problem. The companies that are moving are the ones that picked a use case, built the right context layer for that specific scope, and proved something real before expanding.

🔗 Watch the full episode on our website

Rahul Vohra, founder of Superhuman, spent years turning it into something else entirely: a repeatable, measurable, roadmap-generating engine:

Ask the one question that actually predicts growth

“After benchmarking hundreds of startups, Sean found that the companies that struggle to grow almost always got less than 40% “very disappointed”, whereas the companies that grew most easily almost always got more than 40%.”

The foundation of the engine is a single survey question borrowed from Sean Ellis, who ran early growth at Dropbox and Eventbrite:

How would you feel if you could no longer use the product?

Users choose between "not disappointed," "somewhat disappointed," and "very disappointed."

After benchmarking hundreds of startups, Ellis found that companies with more than 40% "very disappointed" responses almost always grew easily. Companies below that threshold almost always struggled.

Piece of advice: Experiment with bringing this benchmark into your CS strategy and add the question to the next survey.

Change the market before you change the product

“I think most people forget this about product market fit. There’s actually two pieces: product and market. We rotate around changing the products because that’s what we do all day as product people, but it’s way easier to change the market.”

When Superhuman ran the survey in 2017, their score was 22%.

Rather than immediately rebuilding features, Rahul's team segmented the results by persona – founders, executives, engineers, salespeople – and asked: what if we only counted the segments that already loved the product?

By simply removing the low-fit segments from the calculation and narrowing the target market, the score jumped from 22% to 32% without a single product change.

Piece of advice: Most teams default to building their way to product-market fit, but adjusting the market definition is faster and often more accurate.

Ignore the users who would not miss you

“As painful as it is, we have to ignore the not disappointed crowd because they are so far from loving the product that they are essentially a lost cause.”

This is the step that often gets skipped:

The "not disappointed" cohort will submit feedback, request features, and generally behave like engaged users. Rahul is ruthless: do not act on any of it.

They are too far from loving the product to be converted, and their requests will pull the roadmap in directions that help no one.

The same rule applies, with nuance, to the "somewhat disappointed" group – but only to those whose stated main benefit does not match what your best users love.

Piece of advice: The discipline of ignoring certain feedback is the protective mechanism for your roadmap. Don’t overload it with requests that won’t bring you user love and loyalty.

Split the roadmap exactly in half

“If we only double down on what users love as vision-driven founders or vision-driven product leaders tend to do, we would not increase our product market fit score because we’re just building more of the same. But if we only address objections... then a competitor will eventually overtake you and out-magic you.”

Once you know why your best users love the product and what is holding the convertible "somewhat disappointed" group back, the roadmap writes itself:

50% of capacity goes toward doubling down on what users already love.

50% goes toward systematically removing the friction that holds the convertible segment back.

Vision-driven teams skip the second half. Data-driven teams skip the first. Both approaches eventually stall.

Piece of advice: The 50/50 split is a forcing function – it prevents both the overconfidence of pure vision and the conservatism of pure data.

Track it constantly – because it will move

“Product-market fit never lasts forever. And in fact, with that kind of attitude, it is easily lost.”

Superhuman's PMF score went from 33% in summer 2017 to 58% within three quarters.

But Rahul admits that the score wobbles with every growth spike, every new segment, every adjacent persona that enters the funnel.

The PMF engine is an ever-changing mechanism that requires constant nurturing.

Piece of advice: If your team last measured product-market fit at launch or during a funding round, it’s time to do it again – and repeat regularly.

What Rahul built at Superhuman works across multiple product lines. The uncomfortable truth it uncovers – that some users will never love your product no matter what you build for them – is just as relevant for a Director of Product at a Fortune 500 company as it is for a seed-stage founder.

The question is whether your team has the honesty to look at the number and the discipline to act on what it tells you.

🔗 Watch the full episode on our website

Arnie Katz, Chief Product and Technology Officer at GoFundMe, made a different call two and a half years ago. When LLMs arrived, his team did not ask how they could make engineers more productive; they asked what was now possible for customers that had never been possible before. That sequencing decision has a direct read-across to how any PM team should be building their AI roadmap right now.

AI-first means customer impact first

“We said LLMs is a game changer. It can allow us to create new experiences and capabilities and features to our customers that were not possible before. And that’s where we focused our effort first.”

GoFundMe's initial AI investments went entirely into customer-facing features. Title generation, story enhancement, and eventually a full agent-led creation flow – all before any internal developer tooling. The reasoning was clear: LLMs were immediately capable of the exact tasks their users struggled with most.

Piece of advice: The question "where does AI give us the most leverage?" should be answered from the customer's perspective before the engineering team's. Productivity gains compound over time. Customer unlocks can be immediate and measurable.

The moment of peak customer friction is the moment of peak opportunity

“Asking for help is hard. I’ve been at GoFundMe now for two and a half years and I wanted to raise money for WatchDuty. The act of starting a fundraiser for this great cause was still hard, even for me, after two and a half years here.”

Arnie returns repeatedly to a single mental model: someone sitting in a hospital parking lot in the dark, trying to write a fundraiser title for their child's medical expenses. That person is not a marketer. They are in crisis. The friction of asking for help – even on a platform designed for it – was still real enough that Arnie himself felt it when launching a fundraiser for a cause he deeply cared about.

Piece of advice: Map your AI investments to the emotional and cognitive low points in your user journey on top of the highest-frequency interactions. That is where friction costs the most and where relief creates the most loyalty.

Three marketplace failure modes

“False positive growth – it’s not that the marketplace is growing, it’s just that one specific supplier is growing. And of course, over time, the marketplace would lose value and that supplier will be able to transact directly with the consumer.”

Arnie names three structural failure modes for marketplaces:

1. Cold start failure: Supply or demand does not materialise,

2. Imbalance failure: One side overwhelms the other,

3. False positive growth: One supplier drives all top-line metrics while the platform stagnates.

Each requires a different intervention, and each looks different at scale.

Piece of advice: Growth metrics can conceal structural fragility. Healthy platform growth is distributed across many suppliers and many demand sources. Concentration in either direction is a leading indicator of future brittleness.

The CPTO model accelerates the moment of new learning

“That decision can now come to my leadership table and we can make it together and pivot those resources, move them around from the different squads, maybe different tribes, and we can react within days to it.”

Arnie has held the CPTO title since 2019 across multiple companies. The core advantage he describes is the speed with which a team can respond to a new experimental finding. When product, engineering, design, and data all report into one leader, reallocating resources in response to a new insight can happen in days rather than weeks.

Piece of advice: Org structure is a product decision. If your structure slows the moment between "we learned something new" and "we changed what we are building," that latency is costing you compounding returns.

Metric alignment removes an entire category of conflict

“It’s a great place to be where your primary metric and basically your financial statements are very much aligned with the mission, which is also very much aligned with the good of the world.”

GoFundMe's primary metric is gross donation volume. It is simultaneously the mission metric, the business metric, and the customer success metric. Arnie is explicit that this alignment is a structural advantage – it removes the internal tension that exists when a team's performance metric and the company's stated purpose point in different directions.

Piece of advice: Prioritisation debates often originate from misaligned metrics. If what your team is measured on and what your company says it stands for are different numbers, every roadmap conversation will carry that tension invisibly.

Watch the full episode on our website

When a product leader has been at the same company for almost a decade and still speaks about the product with genuine excitement, he’s 100% worth listening to.

"Do you feel proud of what you're shipping?"

“I think this is where the team being proud of it becomes really important. If the leaders are applying that judgment, then it’s naturally going to be slow because they’re going to bring something to you, then you’re not going to love it, then it’s going to go back for another cycle.”

This is the core metric Abhishek uses to calibrate product quality across the organisation. It sounds simple; its effect is not. When individuals doing the work hold themselves to this standard, the quality bar stops being something enforced from above and starts being something the team defends from within. The result is fewer late-stage rework cycles and a much shorter feedback loop between "not good enough" and "let's fix it."

Piece of advice: If quality enforcement lives only with leadership, it creates bottlenecks and resentment. Building a culture where the team asks this question before the review happens is the scaling problem worth solving.

The barbell strategy for UX

“Any product we design should be really approachable for new customers. And we have to nail the product for the new customer. And then within that same kind of framework and experience, we also have to nail the advanced user use case.”

Robinhood designs every product by holding two extremes simultaneously: the beginner user who needs the most guidance, and the advanced user who demands the most capability. The argument is that if both extremes are genuinely well served, the middle ground resolves itself. Designing for an average user, by contrast, tends to produce a product that is mediocre for everyone.

Piece of advice: Using two extreme user personas as the frame for product reviews forces sharper design decisions earlier in the process and surfaces the real tensions that a single average-user model would hide.

AI is collapsing ideation timelines – but not judgment

“The early stages of product development are so much faster now in the sense that, like, what used to be two to three weeks of ideation and prototyping and a design sprint can now be compressed into one week.”

What once took four to five weeks – prototyping, design sprints, stakeholder alignment – now takes two to three days. Abhishek is direct about what this changes and what it does not. The speed is real. But the judgment required to decide which ideas are worth pursuing, which moments in a product deserve deep investment, and what will actually stand out in the market – that has not been automated.

Piece of advice: The window for slow ideation processes has closed. But the premium on taste, judgment, and the ability to evaluate options quickly has never been higher.

Compliance and legal should be co-owners

“You want to make those partners feel as owners in the product. As product owner, the engineering manager feels, or, the designer feels, right? The legal partner or the compliance person should feel as motivated to deliver an incredible product to the market as everyone else.”

Robinhood's move to a GM org structure in 2022 was partly a product velocity decision. When compliance, legal, operations, product, and engineering all roll into the same GM, they are structurally on the same team. The mindset shift this enables – from regulatory partners as blockers to regulatory partners as co-owners – is what Abhishek credits for unlocking faster, better product delivery in a highly regulated environment.

Piece of advice: If your compliance or legal partner does not feel motivated to ship a great product alongside you, the problem is probably less the market regulation and more the relationship structure or the org design.

AI as an integrated feature

“Our thought has been, how do we integrate AI into the rest of the Robinhood experience? Not like create. This like bespoke thing that sits off to the side that doesn’t really understand the rest of Robinhood.”

Robinhood Cortex was not built as a standalone AI feature sitting beside the main product. The deliberate choice was to integrate AI into existing customer workflows – from customer support, to stock movement explanations, to active trader scanning tools. The guiding question was how to fit AI into what customers were already doing instead of creating a new surface for AI to live.

Piece of advice: The most durable AI features in financial services – and arguably most products – are the ones embedded in the workflow customers already have, not the ones that require customers to learn a new interaction model from scratch.

Robinhood Social: verified trust as a product moat

“All the identities of people using Robinhood Social are KYC’d because they go through Robinhood brokerage onboarding all the trades that people post are real trades because we know that they did those trades so there’s like a level of verification and safety built into the network.”

Before it was a brokerage, Robinhood started as a social app. That origin story is becoming relevant again. Robinhood Social is being built on a foundation that generic social platforms cannot easily replicate: every identity is KYC'd, and every trade shared is verified as real. As the industry gets more and more polluted with fake screenshots, inflated crypto claims, and performative investing content, verified authenticity is the product differentiator.

Piece of advice: In markets where trust is the core problem, the product moat is the structural guarantee that what users see is real.

The full conversation covers org design, product velocity, the future of financial advisors in an AI-native world, and why Robinhood's earnings calls look like an NBA post-game presser.

Watch the full episode on our website

There is a version of the AI productivity story that stops at “engineers write code faster.” It is true, measurable, and easy to report to a board. It is also incomplete.

The more interesting story – the one that has real implications for how product organisations need to operate – is what happens after you solve for velocity. When the engine is fast, the constraint shifts. Suddenly the quality of your thinking, your strategy, and your stakeholder alignment becomes the binding variable.

The COVID era was a stress test

“We thought it was basically over a decade worth of advancement in both acceptance of remote work as well as in the maturity of the tools to do it in a matter of months.”

Zoom tripled its headcount in roughly 12 months during the pandemic. The product went from enterprise communication tool to infrastructure for a planet in lockdown. The architecture held because of decisions made years earlier – luck had little to do with it.

As Jeff put it, "it was an overnight success. No, it was 10 years in the making."

Piece of advice: Reliability and quality floors are strategic investments that compound silently. The teams that build above market expectations while demand is low are the ones that do not break when demand spikes.

Multi-product expansion must follow the user's job, not the revenue map

"I don't go to a meeting just to have a meeting. I go to a meeting to accomplish something."

Post-COVID, video quality alone stopped being a defensible moat.

Zoom’s response was to map the full lifecycle of why someone has a meeting – preparation, execution, follow-through – and build surfaces that address every stage.

Calendar, mail, whiteboard, and async coordination are not diversification bets. They are answers to a specific question: what does someone need before and after a meeting in order to accomplish something?

Piece of advice: Adjacent product moves earn legitimacy when they are traceable to a gap in the user’s workflow. If you cannot name the job it is completing, it is probably a feature in search of a strategy.

Coexistence is a real strategy

“I’ve got to offer something that has an order of magnitude higher value in order to overcome those switching costs.”

Competing against entrenched tools – Google, Microsoft, Salesforce – means accepting that switching costs are structurally high.

The answer is not to demand displacement. It is to integrate, to expose Zoom's capabilities via MCP into third-party surfaces, and to stay close to what users actually need regardless of which ecosystem they live in.

Piece of advice: Map your strengths against the competition's weaknesses, then design for coexistence first. Earn the right to full adoption over time.

When engineering becomes abundant, strategy becomes the bottleneck

“The challenge becomes that pushes upstream the need to, am I doing the right stuff? Is the strategic direction, the outcome driven decision making, that becomes so much more important.”

Zoom was already shipping at a rate users struggled to consume before AI. With AI-assisted coding, that engine got even faster.

The team's response has been to restructure the spec process entirely – bringing stakeholders into the conversation earlier and using AI Companion to draft the spec from that meeting in real time.

Piece of advice: If your engineers are moving faster than your strategy is being updated, you do not have a velocity problem: you have prioritization and alignment problems.

Engagement depth is the right metric for AI value

“It’s aligned incentives, right? So I want deeper engagement because it makes people more retained in our platform, but they’re only retained if they’re getting more value. So that’s the – I like that where it’s not something that I’m extracting. It’s a partnership between us and our users.”

Auto-summarising every meeting is high frequency. Generating a board-ready presentation from months of meeting transcripts, Salesforce signals, and deep research outputs is high value. Jeff tracks the distance between those two states as an engagement funnel – and the goal is to move users down it.

Piece of advice: Design engagement ladders that reward progressively deeper value extraction instead of more frequent touch points. If your AI metric is "used the feature," you are measuring motion without understanding the impact.

MCP and the agentic layer are not optional experiments

“I can’t be closed off in a silo and expect that I’m gonna be able to serve these people to the best of my ability.”

Zoom has made a deliberate choice to play on both sides of the agentic layer – building AI capabilities within its own surface and exposing them via MCP for use in third-party environments.

Jeff's reasoning is pragmatic: "We don't know how this is going to evolve. And so what we need to do is we have to play in each one of these spaces."

The companies that close themselves off to this layer will find themselves unable to serve users who live across multiple ecosystems.

Piece of advice: You do not need a perfect agentic strategy. You need presence and learning signals. Start experimenting now so you are not calibrating from scratch when the layer matures.

The throughline across all three eras Jeff described is the same: stay close to the user, build above what the market expects, and resist the temptation to let competitive noise drive your roadmap.

What makes this moment different is the speed of the feedback loop.

The teams that are honest about where their actual bottleneck sits – velocity, strategy, alignment, or all three – will be the ones that use this era well.

Full episode of The Product Podcast is live on our website.

The most insightful thing David Kaufman said on The Product Podcast this week was also the most awakening:

"The perception of the business was that launching the website was the hard part. That's not the hard part. The hard part is getting people to care and come to your site."

Kaufman is VP of Product and Head of Monetization at TikTok. Before that, he led monetization ecosystems at Facebook and Google. He has spent his career at the intersection of advertising, e-commerce, and platform product, and the conversation covered a lot of ground.

These three ideas are most worth your time:

1. The funnel is a trust problem

Kaufman mapped out the full purchase journey for a new merchant customer: ad impression, click-through, website landing, product detail page, add to cart, account creation, email verification, payment entry.

At each step, he said, you lose half the traffic.

The implication is that the gap between desire and purchase has to close entirely. On-platform commerce – where a user discovers a product in their feed and buys it without leaving the app – is the structural fix:

"You discover the product in feed, you click, you're already a user... and you can really get that serendipitous discovery and purchase right at the same time." That is not a UX improvement. It is a different product architecture.

2. AI at catalog scale is what makes this commercially viable

Kaufman was direct about the creative density problem: if a merchant has 100,000 items in a catalog, generating compelling creative for each one using traditional production is "formidable."

The application of generative AI – adding motion to product images, generating scripts from prompts, automating creator affiliate commissions through the GMV Max solution – is what allows the platform to offer "sell you revenue at a discount and take care of everything else" as an actual value proposition.

He also made a strong point for anyone who works in AI product:

"One of the open secrets about AI is that ads ranking has been using it forever, right? We called it machine learning instead of AI and the terminology has evolved."

The technology is not new – only the interface to it is.

3. Live shopping is a scheduled business operation

The merchants generating the highest commercial volume from live streams are not improvising. Kaufman described a specific operational model:

Announce the live in advance → run ads against the scheduled broadcast → partner with creators who have embedded audiences → build an internal approval workflow nimble enough to shift ad spend mid-stream toward whichever product is performing.

"Stores were doing a million dollars an hour off of their live stream sale."

The US is catching up to what Asia has been doing for years. The businesses that get there first are the ones treating live as a production vs a casual broadcast.

The throughline across all three ideas is the same one that Kaufman traces back to his time at Facebook:

If you build a monetization solution that honors the user experience, the revenue takes care of itself.

The question is whether you have the product infrastructure to make that true at scale.

Full episode of The Product Podcast is live on our website.

The design-to-code gap is finally closing – and product teams need to catch up

This week on The Product Podcast, Ed Bayes (Design Lead at OpenAI, working on Codex) and Matt Colyer (Director of Product at Figma) walked Carlos González de Villaumbrosia through a new integration that closes that loop in both directions. Here are the core ideas they worked through — and what each one means for how product teams operate now.

The silos were already collapsing. The tools just caught up.

The handoff model was already a workaround and not a system built to last. If your team still operates with hard separation between design and engineering, that friction is no longer a constraint but a choice. The round-trip workflow between Codex and Figma formalizes something high-performing teams were already doing informally.

Ed Bayes:

“More and more designers are in code, and because the velocity of engineering has sped up so much, more and more engineers are also designing – designing in prod, building new components. Engineers are jumping into Figma, iterating on design systems. And designers are hopping into code more – jumping into last-mile polish, upleveling the craft, and closing the fidelity gap between the initial design and the final product.”

Agents absorb the translation layer. Judgment stays human.

This is the clearest articulation of the AI leverage model for product teams. The repetitive, high-stakes, low-creativity work – syncing component states, maintaining parity, translating specs – is exactly what agents handle well. What they don’t replace is taste, direction, and the judgment call about what something should feel like. That ceiling is still set by the people in the room.

Matt Colyer:

“Design taste and direction still matters just as much as it did before – but now we don’t have to spend our time moving layers around and implementing slots and props inside Figma. We can ask agents to help us do that.”

Ed Bayes:

“It’s about automating low-level work so you can focus on the high-level primitives.”

The interface barrier is lower than most product leaders think.

The entry point to these tools has been redesigned around how humans already communicate. The Codex app, the Figma plugin, the Write to Figma capability – none of them require a terminal. If you’ve been watching from the sidelines because the tools felt like they were built for engineers, that argument expired.

Ed Bayes:

“We’ve moved from a side chat on the right to putting chat front and center on the left. The primary focus becomes the conversation. And it turns out anyone can have a conversation. At its core it’s just a conversation, and that’s accessible to anyone.”

Skills are institutionalized domain expertise – and anyone can write them.

The API layer requires engineering to build. But the skills themselves – the patterns that tell an agent how to build a component correctly – can be written by any practitioner who knows what good looks like. The people creating compounding value in an AI-first product org are not necessarily the most technical. They’re the ones with the sharpest domain knowledge and the discipline to codify it.

Matt Colyer:

“Think of skills like recipes: we provided all the ingredients, which is the complicated engineering part. But anyone in the Figma community can totally write a skill, because it’s about taking the domain knowledge and expertise the community has. Knowing, for example, that when you build a design system component, you should use tokens so you can support light mode and dark mode in the future – that’s expertise from the design profession that a layperson might not have.”

The cost of experimentation has dropped far enough to change the meeting.

This is a direct challenge to current operating norms. The prototype is now often faster than the calendar invite. The question for product leaders is how many of your current decision rituals are optimized for a world where building something was expensive. That world is changing faster than most team processes are.

Matt Colyer:

“Don’t have a meeting, just make a prototype. The number of people in a meeting and the time spent debating ideas – some of those ideas could just be tried. Was it A or B? Let’s do both, try it, and then decide.”

Two final pieces of advice from Ed and Matt

Ed Bayes:

“If you’ve always thought coding wasn’t for you or it’s scary, just give it a go. Once you start building, it’s not just prompt-and-ignore – if you’re curious, you start asking what’s going on. It’s like a tutoring approach.”

Matt Colyer:

“It doesn’t matter what role you have in a business. These tools can affect almost anything you do. To anyone on the product side, whether early or late in your career, I’d encourage curiosity and play. We haven’t even tapped the full potential.”

Bottom line, a little bit less reading and watching about AI — a little bit more just prompting it and going from there.

Watch the full episode on our website.

The most interesting AI product conversations are about consequences.

What happens to your UI when the AI has already done the work? What happens to your pricing strategy when AI compresses switching costs? What happens to your team’s decision-making culture when the pace of iteration increases by an order of magnitude?

Diya Jolly, Chief Product and Technology Officer at Xero, has been sitting with these questions for a while. She came to the role having led monetization at YouTube and served as CPO at Okta – a background that spans consumer, large enterprise, and now the SMB space.

This week on The Product Podcast, we covered four areas worth slowing down on:

The SaaS interaction model is being inverted

Diya’s argument is direct: most companies are adding AI to existing UIs without rethinking what the UI is for. Her view is that the interaction model itself is changing in two ways. First, users will move from taking action in apps to reviewing what has already been done. Second, they will move from clicking buttons to asking questions in plain language across whatever surface they are already using. The product that does not account for this is only accumulating UX debt.

Accuracy is a design problem, not just a model problem

In a regulated environment like accounting, the instinct is to wait for AI to be perfect before deploying it. Xero’s approach is more operational: build algorithmic guardrails that constrain what the LLM is allowed to return, and build human review workflows directly into the automation layer so users can verify and correct quickly. According to Diya, users in certain workflows are saving roughly 22 hours a month. That payoff is what sustains trust when the AI occasionally gets something wrong.

One-way doors require optionality. Two-way doors require speed

Diya draws a clean line between decisions that are hard to reverse and those that are not. Pricing is her canonical example of a one-way door: start high, because there is always a path to lower prices but almost never a credible path back up. For everything else – the 80% to 90% of daily decisions that can be structured iteratively – the answer is to experiment and correct, not deliberate and delay. Most product teams invert this. They move fast on irreversible decisions and build consensus machinery around decisions that should just be tested.

Focus is the strategy

Xero serves businesses with zero to 200 employees, with a core emphasis on one to 50. When asked about the pressure to move up-market, Diya’s answer was grounded: the needs of an SMB customer and an enterprise customer are fundamentally different, and a product built to serve both will serve neither well. The SMB customer needs consumer-grade simplicity layered over enterprise-grade functionality. That is a specific, difficult design problem – and it is enough of a market to build a substantial business around.

Watch the full conversation on The Product Podcast website page or YouTube.

As the most senior product leader at Shopify reporting directly to the CEO, Vanessa oversees the product strategy for a platform with a $165B+ market cap and recent 30% year-over-year revenue growth.

A two-time YC founder, Vanessa brings a builder’s mindset to leadership, emphasizing the need for technical fluency and courage as a service when making high-stakes decisions.

What you’ll learn:

• AI-Native Playbook: How Shopify transitioned from traditional SaaS to agent-led commerce.

• Internal AI Evals: How to use LLM-based judges to grade and ensure product quality.

• Technical Product Leadership: Why staying involved in API details and technical cohesion is critical for senior roles.

• The Versioning Framework: The strategy Vanessa used to successfully push back and convince her CEO to version Shopify’s API.

Key Takeaways:

• Beyond the PRD: Shifting focus from rigid specs to training AI models for non-deterministic outcomes.

• Courage as a Service: How to leverage deep domain expertise to find your screw it moment and drive organizational change.

• The Founder Mindset: Maintaining autonomy and an experimental spirit within a 10,000-person global organization.

The full episode is also available on our website

Robby is steering the most significant shift in Search history: the transition to AI. He oversees a massive portfolio, including the new AI Mode, which has already scaled to 75 million daily active users. Drawing from his time building Stories and Reels at Instagram, Robby breaks down how to build zero-to-one products inside a tech giant and why you need high conviction to push through early bad data.

What you’ll learn:

• Building Like a Startup: How to maintain crazy speed and focus on zero-to-one initiatives while protecting a massive core business.

• True Product-Market Fit: Why flat or J-curve retention in early cohorts is the only reliable indicator that a product is actually working.

• The Value of Colossal Disasters: The untold story of how early failures with Instagram Reels and Close Friends were necessary steps to global success.

• Agentic Search: How Google is moving beyond providing links to executing complex, multi-step tasks by deeply understanding personal context.

Key takeaways👇:

• Start Small to Win Big: Even at Google’s scale, massive AI products begin with just 500 trusted testers and a focus on solving specific user complaints.

• Look for the Golf Shot: When building AI, look for that rare moment when the whole system works perfectly to build the conviction needed to keep iterating.

• Leaders Must Co-Create: To move fast in a large org, leaders shouldn’t just approve from the top; they need to form working groups and operate in the details.

  ---

  The full episode is also available on our website

🗺️ The AI Executive Roadmap: Moving Beyond Features to Ecosystems

Stop building temporary AI features. Start defining your long-term AI strategy. For VPs, CPOs, and Founders, the challenge in 2026 isn’t just adding AI—it’s restructuring the product roadmap to build a defensible, AI-native ecosystem.

This February, we have curated a high-level series featuring Product leaders from Google, Walmart, YouTube, and Vercel to help you validate your strategic direction.

These sessions are designed to help you cut through the noise and lead your organization toward scalability, retention, and a sustainable competitive advantage.

What You Will Learn

• Strategic Roadmap Prioritization: How leaders at Notion and Stripe use AI to filter signal from noise, aligning product decisions with long-term revenue impact.

• Infrastructure as Strategy: Why successful transformation requires a deep focus on the backend foundations—from GPU unit economics to latency—that support global scale.

• The Orchestrator Model: Shifting your organization’s mindset from managing outputs (static PRDs) to orchestrating outcomes (autonomous agents and high-fidelity prototypes).

• Defensible Localization: Scaling platforms that balance global efficiency with the hyper-local nuances required to build customer trust.

🎙️ Podcast Episode: How Real AI-Native Products Operate and Ship Faster

Speaker: Aparna Sinha, SVP of Product at Vercel

Wednesday, Feb 4 · 11:30 AM PST

In this episode, Carlos Gonzalez de Villaumbrosia, CEO & Founder at Product School, interviews Aparna Sinha, SVP of Product at Vercel, the cloud platform recently valued at $9.3 billion following a $300 million Series F. Aparna joins us to discuss how Vercel is powering the next generation of AI-native applications.

Drawing from her experience at Google Kubernetes and Pear VC, Aparna reveals how Vercel empowers Teams of One to ship faster than ever. She explores the cultural shift required to build in the AI era—moving from rigid planning to rapid experimentation and iterating to greatness.

• Key Insight: Speed is Survival. In the AI era, waiting for perfection means falling behind; agency over hierarchy allows small, autonomous teams to outperform rigid structures.

RSVP FOR FREE

Transformational AI Product Strategy for eCommerce

Speaker: Udit Agarwal, AI Product Leader at Google

Tuesday, Feb 10 · 10:30 AM PST

Online shopping has become a daily habit for millions, yet the journey—from selection to payment, fraud checks, and returns—remains far from seamless. With the rise of AI, retail is on the cusp of adopting autonomous actions that simplify these complex transaction lifecycles.

Join Udit Agarwal, Lead PM at Google, as he deconstructs the strategy behind AI-native retail. He will move beyond the frontend user experience to reveal the backend infrastructure required to scale, including the critical trade-offs between model latency, GPU investment, and unit economics.

• Key Insight: Infrastructure is Strategy. Successful transformation requires connecting front-end personalization with back-end delivery logistics through deep focus on GPUs and TPUs.

RSVP FOR FREE

Product Strategy Lessons from Notion, Stripe & Google

Speaker: Birkan Icacan, VP of Product at Enterpret

Thursday, Feb 12 · 10:00 AM PST

Most product teams drown in customer feedback but struggle to prioritize what actually matters. Join Birkan Icacan, VP of Product at Enterpret and former Head of Enterprise Product at Notion, as he shares how he scaled product decisions at Google, Stripe, and Notion using AI and customer intelligence, and how these principles continue to inform how teams plan and prioritize for 2026. Learn how leading teams move beyond volume-based prioritization to make strategic decisions grounded in revenue impact, retention, and business metrics.

You will walk away with two frameworks: a maturity model to identify where your team is today, and an operational playbook for leveraging AI and customer feedback across the product lifecycle, from quarterly planning to launch impact measurement. See real examples of Google, Stripe and Notion on what breaks at scale and how to transform overwhelming feedback into your competitive advantage.

• Key Insight: Quantify Beyond Volume. Leverage AI Knowledge Graphs across the product lifecycle to accelerate time-to-insight and move from gut-feeling to metric-grounded decisions.

RSVP FOR FREE

Leading AI Products: Speed & Orchestration

Speaker: Kalyan Ganapathisubramanian, Group Product Manager at YouTube

Tuesday, Feb 17 · 10:00 AM PST

The traditional Product Management playbook is broken. In an era where AI can synthesize months of research in minutes and generate code on demand, impact is no longer bottlenecked by execution, it is driven by product instinct.

Join Kalyan Ganapathisubramanian, Group PM at YouTube, as he reveals the new frontier of AI Product Leadership. Drawing on 11+ years of experience across YouTube and Google, he will demonstrate how to evolve your role from managing outputs (static documentation) to orchestrating outcomes (scalable prototypes and evaluation frameworks). This is not just about using tools; it is about leading at the speed of thought.

• Key Insight: The New Deliverable. The era of the static PRD is ending. Replace documentation with interactive, scalable prototypes and Governance as Code to ensure quality at the speed of thought.

RSVP FOR FREE

🎙️ Podcast Episode: Scaling AI-Powered Localization Across Hundreds of Stores Worldwide

Speaker: Tim Simmons, CPO at Walmart

Wednesday, Feb 18 · 11:30 AM PST

In this episode, Carlos Gonzalez de Villaumbrosia, Founder & CEO at Product School, interviews Tim Simmons, Chief Product Officer at Walmart International, the retail giant serving 255 million customers weekly across 18 countries. Tim is leading a massive transformation to move from decentralized tech stacks to global platforms that empower local innovation.

In this conversation, Tim explains why complexity is actually a competitive advantage when training AI. He dives deep into Agentic AI and the concept of Orchestrators—systems that manage workflows between agents to automate tasks like user story generation with 88% accuracy. Tim also shares the strategy behind the Walmart Translation Platform (WTP), which has cut translation costs by 99% while increasing speed and trust.

• Key Insight: Complexity is Data. The more you expose AI to your organization's unique complexity, the more resilient it becomes. Focus on Orchestrators to automate workflows with up to 88% accuracy.

COMING SOON

Welcome to the AI-Native Era of Product Management

As the global leader in Product Management and AI training for product teams and individuals, Product School is defining this new standard with the world’s most advanced AI Product Management Certifications. We are trusted by Fortune 500 companies and a community of over 2 million professionals. 

Explore our Certifications →

The conversations were bold, dense, and deeply practical: how to reframe growth in a world of AI-powered discovery; how to graduate from AI experimentation to native execution; what it really means to build products, platforms, and agentic experiences at scale.

Thank you to the over 1,000 product leaders who joined us in person and to the more than 10,000 of you who tuned in online. If you missed it, don’t worry! You can watch all the talks on our website and share them with your team.

The Collapse of Traditional Distribution and the Rise of Product-Led Growth Loops

Elena Verna, Head of Growth at Lovable

Self-described “growth titan” and three-time ProductCon speaker Elena Verna kicked off the day with a fiery session on what she calls the collapse of traditional distribution. From her seat leading growth at Lovable—the fastest-growing software company in history—Elena made one thing clear: great products don’t win; great distribution does.

She took the audience on a crash course through the new laws of growth, arguing that funnels are dead and loops are everything. Drawing on examples from her time at Dropbox, Miro, and SurveyMonkey, she showed how compounding product loops—not marketing campaigns—create sustainable, defensible growth.

“Your users should be doing your marketing for you,” she said. “That’s how you build predictable momentum.”

With SEO-driven acquisition collapsing and social algorithms clamping down, she urged product leaders to rethink how users discover and adopt their products. From embedding viral loops into onboarding to leveraging founders’ personal brands and creator partnerships, Elena’s message was simple: distribution is the new product moat.

“There are millions of great products we’ve never heard of—but only because they didn’t have distribution. A great product is not enough. Great product plus distribution wins every time.”

Transforming Your Product Team from AI-Curious to AI-Native

Carlos González de Villaumbrosia, Founder & CEO at Product School

Coming home to San Francisco reminded me how far our product community has come. Twelve years ago, we were just starting to define product management. Now, we’ve found a way to lead in the AI era.

It wasn’t easy! 93% of AI transformations fail. This is because most companies chase tools instead of transformation. Real change happens when you align people, tools, and processes under a new operating model.

We learned those lessons ourselves at Product School, and they’ve directly shaped our offerings. In our AI training for product teams, we align executives first (commitment is the first bottleneck), uplift organization-wide fluency, and rebuild workflows around data governance and human–agent collaboration.

Then we launched three new courses that teams actually need now: AI Prototyping (speed + de-risking), AI Evals (reliability for nondeterministic systems), and Advanced AI Agents (orchestration at scale). We then infused every legacy program to make it AI-native, end-to-end.

Becoming AI-first ourselves is why we can teach it credibly. We’re not a marketplace or consultants—we’re practitioners who lived the transformation and turned it into a repeatable playbook for product teams. As I said onstage:

“Transformation can’t be outsourced. It has to be modeled.”

The future belongs to teams who don’t just use AI, but integrate it into every aspect of the product development process.

The Velocity Advantage: How AI-Native Teams Ship Products Fast and Safe

Aparna Sinha, SVP of Product at Vercel

This hands-on talk, which included live demos onstage, framed AI development as a moving target and argued that your release infrastructure must move with it. The core playbook:

• Flexible infrastructure to swap models quickly

• Fast CI/CD pipeline

• Feature flags for safe rollouts

• Instant rollback when experiments fail.

And security can’t be bolted on; it has to be built in from the start, especially as AI workloads attract more malicious actors.

Sinha demonstrated how Vercel automatically detects frameworks, configures infrastructure, integrates security by default, and uses a scalable, pay-as-you-compute architecture. Beyond developer velocity, tools like Vercel’s v0 let product managers and designers prototype multi-model experiences, share previews with select users, and hand engineers working code. The company’s “work in public” culture encourages early shipping, tight measurement, and iteration—yielding products like their PR-review agent with a 70% acceptance rate.

“Building in AI is like building in an earthquake. I guess that’s appropriate for San Francisco. But the ground is always moving. The best model changes at any time, sometimes multiple times a day.”

From Core to Platform: Leading Product Expansion Without Losing Focus

Andrey Khusid, CEO at Miro — live Product Podcast recording

During our live Product Podcast recording, Andrey and I explored how Miro evolved from a single whiteboard tool into a full-fledged innovation platform—without losing its core. Over 15 years, he’s steered four major reinventions, from browser-based whiteboarding to the upcoming AI Innovation Workspace. His mindset remains rooted in what he calls “day one thinking”: questioning how he’d rebuild Miro if starting fresh today.

Andrey shared that staying ahead means evolving alongside shifting market dynamics, balancing consolidation with AI experimentation, and knowing when to push for reinvention—even when things are going well. Culture, conviction, and customer obsession remain his anchors through every phase of change.

“At the end of the day, it’s all about how fast you are learning as an organization and how truly agile you are. Not just what agile practices you follow, but how fast you recognize signal, separate that from noise, and act on that signal. The faster you learn from failures and successes, the more probability you have to win long term.”

Miro’s next act is about speed of learning, not just speed of shipping—a principle every product leader in the room could relate to.

Building a Product Culture of Ownership around User Impact, Not Shipping Features

Lisa Kamm, Head of Product at Dow Jones

Lisa reframed the classic “influence without authority” paradox and argued for ownership anchored in outcomes, transparency, and shared responsibility instead of a tally of shipped tickets. Lisa pushed teams to define clear goals, make tradeoffs visible, and co-own results across product, engineering, and design. She also challenged incentive structures that reward launches over lasting impact, advocating recognition for killing features when necessary and hard calls that improve user value. Her practical framework, GROW (Goals, Responsibilities, Outcomes, Wins), gives leaders a simple way to align up, down, and across without falling into RACI theater.

“Stopped talking about launches and started talking about landings. Are you landing features? Did it have outcomes? Did it have impacts? Did it do the right thing?”

My takeaway: if we want durable product momentum, we need to measure what matters to users and the business, celebrate the work that lands, and make accountability everyone’s job, especially when it means saying no.

Building Security-by-Design in AI Products

Jeetu Patel, President & Chief Product Officer at Cisco — live Product Podcast recording

I sat down with Jeetu to unpack what “security-by-design” really means in the agent era. His thesis: the shift from chatbots to autonomous agents forces a ground-up rethink of architecture and trust. Three constraints now shape product strategy:

• Infrastructure: limited power, compute, and network capacity to support rapidly growing AI workloads.

• Trust: users won’t adopt unless systems are reliably safe and secure; requires model validation and runtime guardrails.

• Data: fragmented governance and readiness make it hard to harness enterprise data for AI.

Security can’t be bolted on later: validate models (toxicity, jailbreaks, prompt injection), add runtime guardrails, and design platforms that are loosely coupled, tightly integrated so customers can mix vendors without losing protection. Jeetu also pushed leaders to partner widely—even with competitors—to meet customers where they are and earn trust.

“This is the first time that you’re seeing that security and safety is not looked at at odds with productivity. It’s actually looked at as a prerequisite of productivity. In the past, you would have security and safety or you could have productivity. But now, if people don’t trust an AI system, they’re not going to use it.”

Product teams need to treat security as a core product feature, wire it into CI/CD and rollout policy, and keep the model ↔ product feedback loop tight. Partner for speed; measure for trust.

Enterprise AI Adoption to Drive ROI

Moderated by Product School’s AI Product Leader Divya Sabade, the panel brought together SK Krishnamurti (Dovetail), Sean Tindale (Optimal), Sean Collins (Boomi), and Tanya Littlefield (Amplitude).

The discussion centered on how enterprise product teams can move from pilot projects to measurable ROI while keeping trust, governance, and agility at the core.

Panelists shared lessons from their own companies on bridging the gap between experimentation and value. Tanya Littlefield emphasized the importance of aligning finance and product teams around efficiency metrics instead of chasing short-term revenue impact, while SK Krishnamurti illustrated how Dovetail helps teams turn massive volumes of unstructured feedback into real-time insights.

Sean Collins and Sean Tindale both spoke about trust as the critical barrier to scaling AI—trust in data quality, integrations, and decision automation. Across the board, the message was clear: enterprises that pair strong governance with fast, iterative experimentation will be the first to realize AI’s business potential.

As Sean Tindale (Optimal) put it:

“The enterprises that are experimenting right now are the ones we’ll still be talking about in ten years.”

Design, Product, Engineering: Blurring the Lines in the AI Era

Dylan Field, CEO & Co-founder at Figma — live Product Podcast recording

I sat down with Dylan Field to explore how AI is collapsing the walls between design, product, and engineering—and why that raises the bar for all of us.

What struck me first was how deeply he still builds: he carves out hands-on time, stress-tests new models (even jailbreaking them), and uses that learning to sharpen product judgment. We dug into Figma’s evolution from a single product to a platform (Design, FigJam, Dev Mode, Slides, Make, Draw, Sites) and the choice to meet users wherever they work—right down to invoking FigJam inside ChatGPT so ideas can move from prompt to polished artifact fast.

On AI’s UX, Dylan calls today a “prompt-centric MS-DOS moment” and argues the next leap is new, more intuitive ways to steer models—often agentic—beyond text boxes. His challenge to product leaders was clear: brand, craft, and end-to-end experience now decide winners.

As Dylan put it:

“We’re in this world where software is easier to create than ever before. The curve has gone vertical instead of exponential. So how do you win? It’s design, it’s craft, it’s point of view—it’s also your brand and your marketing. And it all has to connect and work together. If you’re not able to internalize that and articulate it in a way that feels joyful and cohesive for the user, you won’t win.”

Orchestrating Agents at Scale to Turn Complexity into a Competitive Advantage

Tim Simmons, Chief Product Officer at Walmart International

Tim Simmons, Chief Product Officer at Walmart International, explored how one of the world’s largest organizations is using AI to turn scale and complexity into a competitive edge. He described how Walmart operates across 18 countries, 22 languages, and 30 brands—complexity that once slowed innovation but now fuels it through agentic systems.

His first example, the Walmart Translation Platform, combines neural machine translation, LLMs, and human linguists to translate intent, not just words. It now processes millions of real-time translations across 15 languages at 1% of the original cost—learning and improving from cultural nuance and user feedback.

Another example is a model used to orchestrate up to 15 specialized AI agents across the product lifecycle to automate work, enhance accuracy, and free PMs to think strategically. In one pilot, a project estimated at 32 weeks reached testing in just eight.

Simmons’ message was clear: as AI accelerates, the challenge isn’t to slow it down but to manage it intelligently.

“AI is moving faster than any of us can track, but that’s not a problem to solve—it’s an opportunity to harness. The more complexity we throw at these systems, the smarter they become. Our job isn’t to contain the chaos, it’s to orchestrate it into something that makes us faster, sharper, and ultimately more human.”

Implementing Customer-Facing Agentic Experiences into Your Product

Rachel Obstler, SVP of Product at Contentsquare

Obstler shared hard-won lessons on layering chat and agentic capabilities onto a non-AI-native analytics product. Start with problems, not hype: map customer pain and outcomes, then intersect them with what LLMs do well (summarization, interpretation, retrieval).

She urged us to invest deliberately, starting with cross-functional “AI Accelerator” squads to prototype quickly, using bootstrap datasets, and shipping iteratively. But there’s no “minimum viable” version for trust. It’s non-negotiable: show the agent’s plan, explain the recommendations, and capture consent with in-product click-throughs.

“Users do not like black boxes. If you’re making recommendations, you need to tell them why you made that recommendation and how you got there. And frankly, there’s no reason not to because after all, generating text is something LLMs are particularly good at.”

Planning > Plans: The Importance of Embracing Speed Now More than Ever

Rob Seaman, CPO at Slack

Rob closed us out on a high note! He argued roadmaps are ill-fit for an AI boom + economic uncertainty, encouraging us to replace feature lists with outcome-based plans and rapid hypothesis testing. Slack scales fast decisions through product principles:

• Don’t make me think

• Be a great host

• Prototype the path

• Seek the steepest utility curve

• Take bigger, bolder bets.

In practice, he suggests forming uncomfortably small proto-teams (often designer+engineer), get to code, learn, and iterate.

“We should not be planning individual features. We should be thinking about the outcomes we want to achieve for our company and for our customers. And we should treat our plans as a series of hypotheses that we’re going to go test. And in doing so, we should test those hypotheses through rapid prototyping.”

This really felt like the best ProductCon yet. Of course, our goal is always to make every event better than the last. But this time we pulled it off because we’re deeply invested in the conversations surrounding AI, the questions, challenges, and opportunities it presents for every product team worldwide.

Product School is doubling down on what product leaders need now. After 12+ years training Fortune 500 teams, we’re leading the shift to AI-native product organizations with programs built for real impact. We focus exclusively on AI training for product teams and for PMs, focused on the skills that product leaders need to become AI-native. Learn more about our enterprise training and individual certifications.

Make sure to join us for the next ProductCon, whether online or in person! I look forward to seeing you at ProductCon 2026 in London on February 24th and in New York City on May 20th.

And we still have far to go. With AI, the old playbook no longer applies. Many of you are already running AI pilots, but you’re finding it difficult to turn these experiments into scalable, revenue-driving initiatives. According to a recent report from MIT, about 95% of generative AI initiatives fail to generate meaningful financial returns or scalable impact. Only 5% of these programs achieve rapid revenue growth.

Just as we’ve helped define and elevate the role of the Product Manager, Product School is ready to lead product teams through this new era of transformation. After over 12 years of experience training Fortune 500 companies, we know what enterprises need for sustainable change that guarantees return on investment in training.

Starting today, Product School is exclusively focused on AI training for product teams and AI certifications for product managers. Why? Because the future of product leadership isn’t just about using AI—it’s about becoming AI-native.

Our courses prepare product organizations and teams to become AI-native leaders who deliver ROI, efficiency, and sustainable growth. We’re not a marketplace of random courses. We’re not consultants who leave behind slides no one looks at again. We are the global leader in AI-first product training, trusted by some of the world’s top companies and a community of over 2 million product professionals worldwide.

The AI Gap Is Holding Product Teams Back

In an AI-first world, our mission is simple: turn product organizations and managers into AI-first leaders so you can move faster, build smarter, and deliver ROI.

The same report from MIT found that AI transformations stall because they aren’t tied to workflows, don’t adapt to company processes, and suffer from what the report calls a “learning gap” between the tools and the people who use them. To succeed, leadership must move beyond proving concepts; you need adoption, execution, and systems that embed AI into everyday work.

This is particularly urgent because coding is accelerating exponentially, meaning the bottleneck is no longer engineering; it’s product. Experts like Andrew Ng have suggested the developer-to-PM ratio could soon flip to 2:1, as opposed to the previous norm of 1:4-6.

The truth is, many of you feel behind. The sprawl of tools is overwhelming. It’s unclear where to start, which skills matter most, or how much technical depth is really required. AI basics and early experiments fizzle when they don’t lead to adoption or impact revenue growth. The product teams that will win in this AI-first world are the ones that connect AI to measurable ROI by building the right skills through training. It’s the only way to embed AI into the DNA of the product function from ideation to launch, across every role, and throughout every workflow.

This is where Product School comes in.

A Bold, New Approach to AI Training

As I said above, product is going through a major transformation. This means we at Product School have had to adapt, too. AI is now the foundation of everything we do. The result is a training system that works for both the product team that needs to adopt AI at scale and the individual learner who wants to future-proof their career.

Every course in our new catalogue has been built to be AI-first and product-focused. That even includes our fundamentals certification because every product manager today must be an AI product manager.

Here’s what makes us different (and better):

• AI-first, product-only: Every program is designed for product managers, either in teams or as individuals, with practical AI built into every training, project, and resource.

• No generic, one-size-fits-all engagements: We start by assessing your team through our AI Maturity Index and benchmarking relevant metrics. We then work with your team and design customized training programs that align with your company’s specific objectives.

• Measured impact: We work with clients to identify their organizations' most important leading indicators; our North Star metric is return on investment in training.

• Live, small, hands-on: No passive videos. Your team learns directly from AI-first product leaders who are doing the work themselves. Unlike consultants who drop in and disappear, we stay with you end to end. You’ll have a dedicated Customer Success Manager and a handpicked training team aligned to your goals of embedding new skills directly into your team’s daily work.

• Knowledge that sticks: Our Learn → Illustrate → Apply → Reinforce model ensures AI skills leave the classroom and embed themselves in daily workflows.

How Our AI Training Delivers ROI

AI for product teams is about increasing revenue and efficiency. That’s why return on investment in our training is our most fundamental measurement of success. With that in mind, before we do a single training, we benchmark the most relevant KPIs that act as leading indicators of higher revenue and lower costs. These vary by org, but often include:

• Experimentation Velocity: Teams run more experiments, test faster, and reduce decision cycles, driving quicker product iteration and discovery.

• Product Adoption: New features are adopted faster and with greater success, indicating stronger product-market fit.

• Strategic Product Influence: Your product teams move from shipping outputs to driving outcomes that are directly tied to revenue and business goals.

• Time-to-Value: Users realize value faster through better onboarding, clearer value propositions, and more intentional product design.

By tracking the metrics that integrating AI in product workflows will impact, we can tie our training directly to your product team's performance and business growth.

Critical Capabilities for AI-Native Product Leaders

Our new, AI-first curriculum is a comprehensive training system designed specifically for product teams. Developed in collaboration with top AI product experts, our programs equip your team with critical skills that underpin successful AI transformations. During the AI maturity assessment, we work with clients to create the most effective training path to meet their growth needs.

Here is a sample of some of the core competencies we cover:

Theory doesn’t transform companies. Action does. That’s why our training is hands-on, cohort-based, and leverages the most powerful AI product tools available.

AI training for every level of product leader

While much of our work focuses on transforming entire product teams, the AI-first courses above are also available for individuals. This includes everyone from entry-level PMs building their foundation to senior leaders mastering advanced strategy. We’ve completely upgraded our five existing certifications to become AI-native, ensuring every program teaches the latest, AI-first product practices and launched three brand-new certifications for the next wave of product leaders: AI Prototyping, AI Evals, and Advanced Agents.

Together, this means our training now covers the full journey:

• Entry-Level PMs who want to learn the fundamentals of AI-first product development.

• Product Managers learning to design and ship AI-powered products.

• Senior PMs & Leads guiding teams through AI-native product development.

• Directors, VPs, and Executives leading AI-driven transformation at scale.

No matter where you are in your career, Product School gives you the live training, expert coaching, and global community to become an AI-Native PM and stay ahead in the AI era.

The Stakes Have Never Been Higher

The reality is simple: AI has turned product development on its head. Teams that fail to adapt will fall behind. Teams that embrace AI-first product management will lead the next decade of innovation.

I’ve been here before. When I founded Product School, people doubted whether Product Managers needed formal training. We proved them wrong. Today, as AI reshapes the industry, I see the same doubt, confusion, and fear of change.

Once again, Product School is stepping up to lead. If you’re ready to take the next step, now is the time. You have a choice: wait and risk being left behind, or equip your teams with the skills, strategies, and systems to lead in an AI-first world.

At Product School, we’re ready to partner with you on that journey. We know what it takes to build lasting capability. We’ve done it before. And we’re doing it again—with AI.

—

Carlos

CEO, Product School

🌟 Discover How AI Agents Are Transforming Product Development: Join Our Webinar Series!

Episode 1: Discovery and Requirements: Identifying AI Agent Opportunities

Wednesday, Sep 10 · 10:30 AM PST

In this first part of a four-part series webinar, “AI Agent Product Development: Build, Scale and Drive Growth”, Elio Damaggio, AWS AI Product Leader, dives deep into advanced strategies for identifying and validating AI agent use cases that can drive significant impact across your organization. The session focuses on real-world application, using tried-and-tested techniques to evaluate inefficiencies and uncover automation opportunities that deliver both operational efficiency and measurable ROI.

RSVP FOR FREE

Episode 2: Design and Architecture: Human-Agent Collaboration Frameworks

Wednesday, Sep 24 · 10:30 AM PST

In this second part of a four-part series webinar, “AI Agent Product Development: Build, Scale and Drive Growth”, Elio Damaggio, AWS AI Product Leader, will explore cutting-edge strategies for designing seamless human-agent collaboration frameworks that maximize both product performance and user trust. Through practical case studies and frameworks, you’ll learn how to design and refine agent workflows to ensure that both AI agents and humans can effectively collaborate, enhancing both the user experience and overall business outcomes.

RSVP FOR FREE

🤖 Building the Future with AI-Driven Product Innovations

The Collapse of Pre-AI Distribution Moats | Lovable Head of Growth

Monday, Sep 8 · 10:00 AM PST

Elena Verna, Head of Growth at Lovable, reveals why traditional product distribution channels are collapsing in the AI era—and what it takes to build moats that actually hold. From the death of SEO to the rise of DIY software builders, this talk is a wake-up call for every product leader navigating the next wave of growth.

RSVP FOR FREE

How to Build High-Quality AI Features | Dow Jones SVP of Product, Consumer

Monday, Sep 1 · 10:00 AM PST

Ryan Daly Gallardo, SVP of Consumer Product at Dow Jones, shares the behind-the-scenes journey of integrating AI-powered article summaries into a newsroom built on trust and editorial excellence. Learn how her team turned a risky experiment into a high-value user feature—without sacrificing depth, substance, or journalistic integrity.

RSVP FOR FREE

Storytelling for Product Leaders in the Age of AI | TED Head of Product

Monday, Sep 29 · 10:00 AM PST

Tricia Maia, Head of Product at TED, shares why storytelling is the hidden edge product leaders need to navigate AI pressure, align teams, and build human-centered features that resonate. Drawing from TED’s 40+ years of idea-sharing mastery, she explains how to communicate product strategy with clarity, conviction, and purpose.

RSVP FOR FREE

Create Your Own AI Operating System as a Product Leader | Pendo Field CPO

Monday, Sep 15 · 10:00 AM PST

Dave Killeen, Field CPO at Pendo (EMEA), returns to ProductCon to reveal how product leaders can architect their own AI operating system. Designed not just to enhance productivity, but to unlock deeper thinking, strategic clarity, and creative breakthroughs, this talk offers the blueprint for building a truly augmented mind.

RSVP FOR FREE

Rubrik CPO on Business Model, Org & Product Transformation | Anneka Gupta | E270

Wednesday, Sep 17 · 11:30 AM PST

In this episode, Carlos Gonzalez de Villaumbrosia interviews Anneka Gupta, Chief Product Officer at Rubrik. We explore Rubrik's journey through three major transformations: shifting from perpetual licenses to a SaaS-first model, expanding from a single product to a multi-product platform (with two new product lines approaching $100M in ARR), and repositioning as a cybersecurity-first company. We'll dive deep into the challenges and strategies behind managing such significant organizational change at scale.

COMING SOON

How PM can get the most out of Cursor | Arize AI Head of Product

Monday, Sep 22 · 10:00 AM PST

In this episode, Aman Khan, Head of Product at Arize AI, reveals how PMs can partner with AI-native development environments powered by autonomous agents to go beyond prototyping and into execution. Learn how these agents help PMs read, debug, and even contribute to code—without leaving their spec-writing workflow.

RSVP FOR FREE

At Product School, we’re leading the way in AI training and transformation, helping teams and individuals become AI-native product managers. Our mission is to shape the future by equipping product teams with the skills and knowledge needed to thrive in the AI-powered world. Learn how we’re enabling product leaders to master AI and level up their teams through our AI training for teams and AI Product Certification programs.

Check out key takeaways below, and watch the full recordings to get all the insights!

By the way, we’re already putting together the incredible lineups for the next ProductCons. Don’t miss out—get your tickets now!

How to Build High-Quality AI Features

Ryan Daly Gallardo, SVP of Product, Consumer at Dow Jones

Ryan Daly Gallardo kicked off ProductCon AI 2025 by showing us exactly how Dow Jones built high-quality AI features that deliver real value to their users. She walked through her process of balancing editorial trust with AI automation, demonstrating how careful, data-driven decisions led to impactful results. Attendees left with a practical playbook on how to approach AI integration in a way that enhances user experience without compromising quality.

Key takeaways:

• Balancing editorial trust and AI automation: How to build trust with your audience while using AI to automate and scale your content efficiently.

• Lessons from A/B testing AI features: Real-world examples of A/B tests on AI features, with actionable insights into what the data revealed about user behavior.

• Quality, oversight, and thoughtful UX: Why these three elements are critical to successfully integrating AI into your product, ensuring that the technology works in harmony with your users' needs.

"Progress doesn't come from playing it safe. It comes from asking bold questions, testing them thoughtfully, and listening."

Watch Ryan’s full talk.

Building a High-Performance Product Team of Humans and Agents

The Product Podcast featuring Noam Lovinsky, Chief Product Officer at Grammarly

I had the pleasure of interviewing Noam Lovinsky on a special edition of The Product Podcast. Noam shared Grammarly’s exciting vision to become an AI-native productivity suite, including how they’re evolving into an agent-driven platform and integrating with partners like Coda. Noam's insights revealed how product roles are rapidly changing, with PMs managing not only people but also intelligent AI agents.

Key takeaways from our conversation:

• Every PM at Grammarly is accountable for revenue: Growth is a team-wide effort, and every product manager plays a direct role in driving company revenue. This insight helps shape a growth-focused mindset that extends beyond the typical PM duties.

• AI agents as core teammates: AI is no longer just a tool but an integral part of product teams. Learn how Grammarly empowers PMs to collaborate with AI agents to create more dynamic, effective products.

• Great leaders know how to build: Noam emphasized the importance of product leaders being builders first, and knowing when to step aside and allow your team to shine. This approach encourages autonomy and fosters a culture of innovation.

"With all these new AI tools, it’s going to be a lot easier for anyone who has even a modest desire or instinct to become a builder, because there are many things you no longer have to learn the hard way."

Watch the conversation.

The Collapse of Pre-AI Distribution Moats, and How to Build New Ones

Elena Verna, Head of Growth at Lovable

Elena Verna delivered an insightful talk explaining how traditional growth strategies like SEO are being disrupted by AI, urging product leaders to adopt new strategies like growth loops. Elena highlighted the critical importance of building strong, AI-era moats to remain competitive, especially as customers themselves now have the power to build competing solutions.

Elena Verna shared how Lovable, a company on the cutting edge of AI-driven software creation, is transforming the product development landscape. Recently, Lovable raised $200M in one of Europe’s largest Series A rounds ever, bringing their valuation to $1.8B.

Key takeaways from Elena’s session:

• Traditional growth levers are outdated: SEO and paid social are no longer as reliable due to AI’s ability to disrupt traditional channels and allow competitors to emerge quickly.

• AI is democratizing software creation: With tools like Lovable, AI is empowering the 99% of people who can't code to create and launch their own software—making users powerful competitors.

• The new moats are all about speed and integration: Success in the AI era is about moving quickly (velocity), creating a strong brand, leveraging data, and building integrated ecosystems that create lasting value.

“Having a great product is not enough. There are many incredible products out there that have never seen the light of day, which means that having a great product does not equal a great company. Product plus distribution is what actually makes a great company.”

Watch Elena’s full talk.

Create Your Own AI Operating System as a Product Leader

Dave Killeen, Field CPO, EMEA at Pendo

Dave Killeen took us on a very relatable journey out of "the yellow sticky room," the chaotic mess of scattered ideas every PM knows too well. He showed us how AI-powered workflows can organize our ideas, tasks, and insights, giving us space to breathe, think strategically, and be truly creative. Dave emphasized the value of letting AI handle the tedious stuff so we can focus more on connecting the dots and achieving a healthier work-life balance.

Key takeaways from Dave’s session:

• Escape scattered notes with a smart AI system: Use AI to organize your thoughts, tasks, and ideas in one connected platform, eliminating the chaos of scattered notes and fragmented tools.

• Automate insight capture with AI-driven workflows: Automate the process of capturing valuable insights using AI tags, workflows, and agents to ensure nothing is lost and everything is actionable.

• Reclaim time for deep thinking, rest, and creativity: With AI handling routine tasks, product leaders can focus more on high-level thinking, innovation, and even taking necessary breaks to recharge.

“The goal of your AI operating system isn't just to be more productive; it is to be more strategic, more creative, and ultimately much more human.”

Watch Dave’s full talk.

Integrating AI into Product Strategy to Grow Revenue

The Product Podcast featuring Inbal Shani, Chief Product Officer at Twilio

Inbal Shani joined us to talk about how AI is transforming customer engagement at Twilio. With an extensive background in AI and machine learning, including leading the launch of GitHub Copilot, Inbal now oversees Twilio's product strategy and R&D, working to redefine customer journeys using AI. She shared how AI is integrated across Twilio’s products to improve customer experience, and how product managers can leverage these tools effectively.

Inbal discussed the key shift in thinking about AI as a tool to meet customer needs. This customer-first approach, combined with measurable success criteria, is how AI is truly adding value at Twilio.

Key takeaways from my discussion with Inbal:

• AI is a tool, not a strategy: AI should never be treated as the strategy itself. Customer need is always the strategy. AI is a powerful tool that helps you solve specific customer problems, whether that’s improving automation, enhancing productivity, or delivering personalized experiences.

• Define success through customer impact: For AI to succeed, you need clear success metrics. Whether it’s reducing support tickets, improving customer satisfaction, or accelerating response times, measure AI’s effectiveness through tangible customer outcomes.

• AI in customer engagement: Twilio is pioneering the use of AI to enhance customer interactions, whether it’s through conversational AI, fraud detection, or personalized marketing campaigns. The focus is on making each interaction more relevant and impactful for the customer.

Watch Inbal’s full talk.

How PMs Can Get the Most Out of Cursor

Aman Khan, Head of Product at Arize AI

Aman Khan demonstrated how product managers can dramatically streamline their workflows using Cursor, an interactive development environment. He emphasized how PMs can not only simplify their technical workflows but also enhance their productivity by using AI to collaborate more effectively with engineering teams.

In his demo, Aman showed how Cursor allows PMs to immediately dive into the codebase, gain insights, and perform technical tasks without needing to be a coding expert. For example, he used Cursor to clone a repo and instantly ask the system, “What does this code do?”—a feature that enabled him to understand the product's architecture quickly, without writing a single line of code.

Key takeaways from Aman’s demo:

• Easily interact with your codebase: Use Cursor to quickly understand the structure and function of any codebase by asking direct questions like “What does this code do?” or “How do I run this code?”—no need to wait for engineering to provide all the details.

• Streamline your documentation and specs: Instead of manually writing PRDs or documentation, let Cursor generate them for you based on your codebase, helping you save time and ensure consistency in your documentation.

• Collaborate more effectively with engineering: Cursor acts as an “AI tech lead,” enabling you to navigate code, debug, and even help set up environments—all within a unified platform, making you a stronger partner to your engineering team.

• Use AI to iterate quickly: With Cursor, you can brainstorm and prototype faster. For example, you can prompt the AI to generate new product ideas or features by simply inputting natural language requests, turning concepts into concrete action items without needing to write every line of code yourself.

“The PM role is changing. The expectations on us and our teams have increased in this new world where AI is reshaping how we use technology and build products. It feels like roles that used to be distinct are now overlapping across product, engineering, design, and even extending into sales and marketing.”

Watch Aman’s full talk.

Storytelling for Product Leaders in the Age of AI

Tricia Maia, Head of Product at TED

Tricia Maia captivated the audience by diving into the power of storytelling for product leaders, especially in the age of AI. At TED, they’re experts in how to communicate complex ideas simply and effectively, and Tricia encourages product leaders to apply these same principles to AI product development. She shared how product leaders can harness the art of storytelling not just to explain AI features but to influence decisions, align teams, and lead with clarity.

In her talk, Tricia discussed the paradox of human storytelling in an AI-saturated world and how PMs can use AI to sharpen storytelling skills. She emphasized the importance of grounding AI features in a bigger picture need, ensuring that each product feature has a meaningful story that connects with users.

Key takeaways from Tricia’s session:

• Storytelling is essential for PMs in the age of AI: As product managers, your role isn’t just about launching functionality—it’s about guiding your users through a transformation. Great product leaders don’t just ship products; they inspire and persuade through compelling stories.

• Start with the user, not the tech: Don’t start with the tech and reverse-engineer the user story. Begin with understanding the real needs and challenges of your users, and only then build solutions that address those needs.

• Use AI to enhance your storytelling: AI can be used as a tool to refine your storytelling process. It can help gather insights, automate repetitive tasks, and generate data-driven narratives, but the core message and emotional resonance should always come from the product leader.

"Every product is a story, whether you tell it intentionally or not. You're not just launching functionality; you're guiding people through some sort of transformation."

Watch Tricia’s full talk.

Architecting the AI-Native Organization: Infrastructure, Agents, and Teams

Panel Discussion featuring Marcus East, VP & GM of Digital & eCommerce at Autodesk & Ed Macosky, Chief Product & Technology Officer at Boomi. Moderated by Divya Sabade, AI Product Leader at Product School.

We wrapped up ProductCon AI 2025 with a powerful panel discussion moderated by our own Divya Sabade. Marcus East and Ed Macosky shared their insights on how to build AI-native organizations and the infrastructure necessary for scaling AI solutions. They explored the transition from assistive AI, which aids human workflows, to agentic AI, which drives processes autonomously, and highlighted the essential governance and organizational changes required to support AI-driven work at scale.

Key takeaways from the panel:

• Moving from assistive AI to agentic workflows: Shifting from AI tools that support human decision-making to fully autonomous AI systems that handle workflows independently requires new approaches to integration, autonomy, and control.

• Governance is non-negotiable in scalable AI systems: As organizations scale AI, robust governance is crucial to ensure ethical, efficient, and secure deployment. AI systems must be built with clear oversight mechanisms to avoid risks and ensure accountability.

• The organizational structure shifts needed for AI-native work: To truly embrace AI, organizations must evolve their structures, creating cross-functional teams that blend product, engineering, and AI expertise, and developing new roles that support AI-driven workflows across the company.

"The best AI implementations take advantage of underlying data structures that make sense, discipline around CI/CD, and good governance." — Marcus East

Watch the conversation.

A huge thank you to the visionary speakers, moderators, and everyone who joined us at ProductCon AI 2025. Your passion and dedication continue to shape the future of AI and product leadership. This year’s event was truly one of a kind, packed with actionable insights, real-world use cases, and hands-on demos.

ProductCon AI is the premier event for product leaders who are eager to learn how to build better products with AI. I loved seeing attendees leave with AI playbooks they can directly apply to their business, having gained real AI strategies from leaders who are already shaping the future.

As AI continues to evolve, we remain focused on providing you with the tools to build smarter products and lead more effectively. The product leaders you heard from at ProductCon AI 2025 are showing how AI can be leveraged to drive real change, and we are proud to have shared this journey with you.

Ready to become a leader in AI product management? The Artificial Intelligence Product Certification (AIPC) will equip you with the tools and strategies to build and integrate AI throughout the product lifecycle. Learn directly from AI product leaders like the ProductCon AI speakers above to master AI product growth, create scalable AI solutions, and craft effective strategies for executive buy-in. Learn more.

We’ll keep talking about AI and the topics that matter most to product leaders at upcoming ProductCons. Join us online or in person to be part of the conversation:

• ProductCon San Francisco on October 8th, 2025

• ProductCon London in February 2026

See you there!