Typeform CEO on Why Breadth Beats Depth as an AI Moat and How to Build a Defensive and Offensive AI Strategy | Jay Choi | E301

In this episode of The Product Podcast by Product School, Carlos González de Villaumbrosia sits down with Jay Choi, Chief Executive Officer at Typeform. Typeform is the AI engagement platform trusted by more than 150,000 customers, including 95% of the Fortune 500. Before Typeform, Jay spent seven years as Chief Product Officer and General Manager at Qualtrics, where the company scaled from $100M to over $1B in ARR.

What you'll learn:

• Breadth of surface area as a stronger AI moat than depth of use case, and why going broad is the right strategic bet right now
• The dual posture Typeform built: a defensive strategy to make their core product impossible to replicate, and an offensive strategy to expand into full customer workflows
• Research Flow, their new product that compresses 50 customer interviews from weeks into hours using AI-moderated research
• Being model-agnostic from day one, and what they learned when switching models without an observability platform in place
• The pricing experiment framework Jay uses: 30 simulations before a single market goes live

Key takeaways:

• When AI threatens to commoditize your core product, expanding surface area is a stronger defense than adding AI features to what you already have
• Positioning AI capabilities in plain language, not technical terminology, is the difference between adoption and abandonment
• Happy churners are a product problem, not a marketing problem: the fix is finding structurally always-on use cases.

Credits:
Host: Carlos Gonzalez de Villaumbrosia
Guest: Jay Choi

Social Links:

• Find out more about Product School here
• Follow our Podcast on TikTok here
• Follow Product School on LinkedIn here

Transcript

Typeform's AI Defensibility and the Vibe Coding Threat

Jay Choi | Typeform 00:00:00

We don't leverage AI, agentic AI, or that terminology so much, because we have found with our customers that what they want to know is: what does this do for us in particular? Imagine trying to talk to 50 customers, especially for product people. Talking to 50 people will take weeks, maybe even months. With AI-moderated research, you can literally get answers in hours.

For our traditional forms business, we think about it as: how do we be defensive? So someone can just type in, "Hey, build me a lead generation form for a product-led podcast," and it'll suck it in and give you best practices derived from millions of our data points. Going broad actually starts to matter a little bit too. The broader surface area we create, the harder it is to replicate.

Carlos González de Villaumbrosia | Product School 00:01:00

Hey, this is Carlos, CEO at Product School and your host on The Product Podcast. Today's guest is Jay Choi, Chief Executive Officer at Typeform. Typeform is the AI engagement platform trusted by more than one hundred and fifty thousand customers, including ninety-five percent of the Fortune five hundred. Originally built around conversational forms, the platform has expanded into AI-powered workflows that connect lead capture, enrichment, nurturing, and conversion in a single platform.

Before Typeform, Jay spent seven years as Chief Product Officer and General Manager at Qualtrics, where the company scaled from one hundred million to over one billion dollars in ARR and went public in a thirteen billion dollar IPO in twenty twenty-one.

In our conversation, we cover breadth of surface area as a stronger AI moat than depth of use case and why going broad is the right strategic bet right now. The dual posture Typeform built: a defensive strategy to make their core product impossible to replicate, and an offensive strategy to expand into full customer workflows. Research Flow, their new product that compresses fifty customer interviews from weeks into hours using AI-moderated research. Being model-agnostic from day one and what they learned when switching models without an observability platform in place. The pricing experiment framework Jay uses: thirty simulations before a single market goes live.

Let's get into it. Jay, welcome to the Product Podcast.

Jay Choi | Typeform 00:02:00

Thank you, Carlos. Thank you for having me, and thank you to all your listeners.

Carlos González de Villaumbrosia | Product School 00:02:05

I'm excited to have you because I started using Typeform many years ago, and I know you've gone through a lot of evolution, especially as the world becomes more AI-driven. I'm very curious to dive deeper into that. Maybe we can start with you — what is your story on how you got into Typeform and became CEO?

Jay Choi's Background: From Qualtrics to Typeform CEO

Jay Choi | Typeform 00:02:30

My background has bounced between product and GM roles for most of my career. Starting off early on in consulting, then after business school I took a role at a company called Danaher, which is more industrial tech, doing a product and GM role. Then I joined a company called Qualtrics, and that's probably the longest tenure in my career and the most applicable to the product space.

When I joined Qualtrics, it was at about $100 million in revenue, on a growth tear but still relatively young in its journey. A big deal for us at the time was $10,000, selling to a research director. By the time I left, we were negotiating $10 million deals to the CEO of companies. It had been quite a journey, and it was really fun to be part of that growth.

After a couple of stints elsewhere, I joined Typeform's board of directors as an independent member. What they were looking for was someone who knew the space really well and was focused on driving growth — that's always been what I get excited about. Very shortly after I joined the board, there was an opportunity to join as CEO. Typeform is at a very similar phase to when I joined Qualtrics: about the same size, just cresting over $100 million, looking for the next stage of growth. I've been here just a little over a year now.

Carlos González de Villaumbrosia | Product School 00:04:00

That's one of the unique things that got me excited about this conversation. You come from a very strong product background, joined the board as an independent director, and now you're full-time CEO. Let's talk about the potential disruption — or opportunity — because Typeform has the word "form" in the company name, and you're at $100 million ARR. We see a lot of headlines every time Anthropic or OpenAI make a new announcement, and it feels like a few other companies just die. How are you thinking about that from a defensibility standpoint, to make sure Typeform not only doesn't die, but actually can take advantage of the opportunity?

Defensive and Offensive AI Strategy

Jay Choi | Typeform 00:05:30

It's a fantastic question, and it reminds me of when I first joined as CEO. If you look 16 to 18 months ago, we weren't in the midst of the SaaSpocalypse — AI was still on the earlier end of the curve. When I first joined, I talked to a lot of bankers and investors who were introducing themselves, and the question always emerged: how does AI impact your business? It was always a question from an investor standpoint, but also: what does this mean for your growth going forward?

The way we digested that is: businesses need both a defensive posture and an offensive strategy. The defensive question is: how do you know someone can't do the same thing in ChatGPT or Anthropic? Why can't they just vibe code your product? The offensive question is: if AI continues to be more prolific, where does that benefit your business?

We've really internalized both perspectives. For our traditional forms business, we think about it as: how do we be defensive? We put it on ourselves to create the same conversational interface in our core forms business. Someone can just type in, "Hey, build me a lead generation form for a product-led podcast," put their URL in, and it'll suck it in and give best practices derived from millions of our data points. That's what we think of as our defensive posture.

On the offensive side, AI is now enabling us to do things that were probably never possible before. One of our focus areas is going beyond forms to entire workflows. An example is our lead generation flow: we take a lead that comes in, enrich it with AI, and create AI-generated automations on top of that. None of that was possible before.

A new product we just launched is called Research Flow, which allows customers to do AI-moderated research. Imagine trying to talk to 50 customers — for product people, if you're making a decision and want feedback fast, talking to 50 people will take weeks, maybe even months. With AI-moderated research, you can literally get answers in hours. That's how I think about it: what are we going to do to defend ourselves in AI, and what are we going to do to be an AI benefactor and go on offense?

Carlos González de Villaumbrosia | Product School 00:08:30

I like that duality — thinking defensively, like "okay, we're going to die, so why don't we kill ourselves first" as an opportunity to then resurrect and thrive, and then more offensively, expansion. One of the things I hear about your expansion is enterprise, and more use cases not just for individuals. When I first started using Typeform, I wasn't even using it for my own business — it was just a cool, easy, free form. I'm really curious about that connection. You still have a bunch of consumers, but most of the business is probably on the B2B side now.

From Horizontal Platform to Vertical Use Cases

Jay Choi | Typeform 00:09:00

When we thought about how we were going to grow the business, we really looked at our customer base and started segmenting it. We love personal use cases — weddings, birthdays, invites — because it gets people used to the product, exactly like you described. In our free product, we get broad usage and penetration. But where we found people really got a lot of value was on business use cases.

Over the past couple of years, there was a great product strategy already in place, and I was just layering on top of it. There are these great use cases people are already building on the platform that move them from a one-off project to a longitudinal, more strategic, higher-value program — such as lead generation with a product called Growth Flow, or research with a product called Research Flow.

Our goal is to build vertical products on top of that horizontal base so we can expand more and more. We still love people using our free and basic plans to get used to Typeform. But we also want to create paths for them to use more verticalized solutions: higher-value, longitudinal, always-on programs where we can offer so much more to the entire workflow. That's what we've been building over time.

Carlos González de Villaumbrosia | Product School 00:10:30

One product that came to mind is Calendly. We hosted their CPO a few months ago, and their original feature of calendar booking became commoditized. They had a whole story on moving enterprise and expanding to end-to-end workflows. One of the challenges I've seen with that is everyone trying to use grandiose terms like "agentic platform," which gets confusing. If everyone's trying to be the agentic platform and integrate with everyone else, how do you play here? How do you position your product in a way that's understood by a customer?

AI Positioning: Plain Language Over Jargon

Jay Choi | Typeform 00:12:00

We've been trying to be more focused on end-user value. If you look at our website, we don't leverage AI, agentic AI, or that terminology so much, because we've found with our customers — we serve SMBs and very small businesses — what they want to know is: what does this do for us in particular?

A good example is translation. It's one of our most popular features. It's an AI feature, and we just call it Translate AI. We don't call it an agentic AI translator. It was so painful before — people would have to take a form, pull it out, translate it, put it in new columns, and then upload it back in. Now, with a click of a button, it translates to French, German, Hebrew, Chinese, whatever language you want.

We've always tried to position all of our AI capability in the most straightforward manner possible. As agentic terminology and usage patterns become more prevalent, we'll adopt more of that language. But for now, we've kept it as simple as possible on the benefits, and that's worked out well.

Carlos González de Villaumbrosia | Product School 00:13:15

I'm smiling because I've seen the opposite — companies that say AI in every single sentence, and you don't know if it's real, and even if it is, it's just so confusing.

Jay Choi | Typeform 00:13:25

Exactly. Our thinking is: if they can't understand it, it's going to be really hard for them to get the benefit from it.

The Happy Churner Problem: Finding Always-On Use Cases

Carlos González de Villaumbrosia | Product School 00:13:35

I want to keep peeling off layers of your strategy. You mentioned you identified three specific business use cases and doubled down on them. From a product perspective, how do you make those trade-offs to ensure you're prioritizing the right way?

Jay Choi | Typeform 00:14:00

It's very similar to my time at Qualtrics. One of the big insights we had there was asking: which use cases are high-value enough that we can turn them into end-to-end solutions? Not just people using it once and shutting it down, but something longitudinal.

We did the exact same thing at Typeform. We polled our users and asked: where do we believe people are using our platform in a more programmatic manner? One of our challenges is churn, and we call them happy churners. People are not mad at the platform — they just use it and then they're done with it. Three months later they use it again, and then they turn it off. We asked: what could we do to make that always-on?

Our research team did a lot of analysis and found that the top two or three use cases we could turn into high-value strategic opportunities were: growth use cases like lead generation and lead qualification, research use cases, and talent use cases.

To your point — three bets is really hard to take on at the same time. So we focused on two things. First, which use cases could deliver the highest value to our existing customer base? We did a willingness-to-pay study and a prioritization study, and growth use cases surfaced to the top, followed by research. Second, which of these had the most overlapping features? Growth and talent, for example, have very different requirements — growth is about enrichment, lead generation, and workflow automation, while talent is about things like building anonymity structures for honest feedback. That would have been very divergent. But research and growth tended to overlap, so we could get a lot of product depth with similar investments across the platform.

That's how we landed on those two, and that's what we're focused on now. Talent-specific features will come online as we mature.

Carlos González de Villaumbrosia | Product School 00:17:30

As part of this prioritization, the buyer within the same company can be different — the decision-maker on a talent product versus a growth product. You had a strong PLG motion. I imagine that still remains strong, but how do you go about expanding within existing accounts for the new use cases?

PLG, Pricing Experiments, and the Path to Consumption

Jay Choi | Typeform 00:18:00

The majority of our revenue still comes from product-led. It's such an advantage because PLG takes many years to build that engine — all the traffic, the millions of visitors a month, hundreds of thousands of signups per month. What we've really spent a lot of time on is pricing and packaging, in two ways.

First, a big change we're making is positioning packages for specific use cases: here are forms for everyone, here are forms for marketers, here are forms for HR. Different functions, different functionality, pointing people toward the right use cases. Second, understanding the key features that drive a lot of value and putting them in the right packages so people can jump straight to those areas.

The one thing we've learned over the past year or so is that it's all about experimenting. We can talk a lot about what we want to do, but we get the best feedback from just putting it into the market and seeing how people react. We've made more changes over the past year than we have in a long time because we want to get signal and get better at it.

Carlos González de Villaumbrosia | Product School 00:19:30

I want to learn more about that, because pricing is such a sensitive area. Experimenting with pricing is probably different than experimenting with copy on a landing page. How do you go about experimenting with features and packaging, especially when you're selling to an enterprise client that wants clarity?

Jay Choi | Typeform 00:20:00

We do simulations first. The first thing is simulations on maybe 30 different variations of our pricing page — what I'd call roughing it in. We pick the two or three winners and measure overall mix, what it does to the ASP, what it does to annual plan adoption, and optimize those pieces.

From there, we take it to geography testing. Maybe we'll test it in the UK, or in smaller markets where it's easier to get enough signal behind it, and then implement it more broadly. We always try to have a guiding north star: what are the pricing principles? Are we trying to make pricing easier to understand? Are we trying to make it easier to adopt new features? We lay those out and judge results based on our goals. And we always have a saying inside: there's no such thing as perfect pricing and packaging. It's just the next evolution.

Carlos González de Villaumbrosia | Product School 00:21:00

Speaking of that evolution — especially in SaaS — a lot of companies are moving away from seat-based pricing, and the opposite extreme would be outcome-based pricing. I've seen examples like HubSpot or Intercom, which just renamed to Fin, on that extreme. Where do you see yourself in that spectrum?

Jay Choi | Typeform 00:21:30

We've thought a lot about this. My personal opinion is that seat-based pricing has been under pressure for a while — it's super easy to understand, but it's also relatively challenging to connect it directly to value. I wish it were as clean as Intercom's resolved-call outcome-based model, but that's not available to all SaaS.

What we've been thinking about is how to ease into it. In areas where we can provide something closer to value-based pricing, we'll do that. For example, on our Growth Flow product, one of the new features is enrichments. We'll price on the number of enrichments, or on the number of automations completed. It's a little closer to outcomes and value, and as the whole world gets more used to consumption models, we're thinking about how to move to some sort of consumption-based metric so people feel they're only paying for what they use.

Model-Agnostic by Design

Carlos González de Villaumbrosia | Product School 00:23:30

How are you thinking about your integrations with different LLMs and evaluating which ones perform better for your use cases?

Jay Choi | Typeform 00:24:00

One of the big strategic imperatives we made from the beginning was to be model-agnostic. Every month a new model comes out with something new, so we're constantly adjusting. We felt that pattern would persist over time. Being model-agnostic means we can always choose the best model, and we've also found that some models are better for some use cases within our own product — we don't have to go all-or-nothing on any particular use case.

A great example: translation might be better served by one model, and another model might be better at the core Typeform AI experience — "Hey, build me a form that does X, Y, or Z."

Internally, before we invested in an AI observability platform, we would switch to a new model and ask: was that better? It was definitely more expensive — but was it better? We had to invest in an observability platform to test whether it was actually better and worth the additional change management. We've definitely leaned into being agnostic, and I think we'll continue to live in a world where new models come out from different companies every other quarter.

Carlos González de Villaumbrosia | Product School 00:25:30

Two thoughts come to mind. One is a comment I remember about strategy for companies trying to build their own models — most people are still trying to pick a very specific use case and assume they have some edge over the model, while others focus on the assumption that models are only going to get dramatically better, and it's not about doing something better than the model because you have more data, it's about something completely different. Most people are still using the first strategy, which has been proven wrong by the pace of change. The other thought is what you mentioned about optionality — not fixating on one model, but considering different use cases based on price, performance, latency, and other variables.

Jay Choi | Typeform 00:26:30

Exactly. It's difficult to imagine exponential growth — how much something will exponentially improve. We want to be flexible so that if something great comes out that we weren't even imagining yesterday, we can make it easy to adopt within our own platform.

Carlos González de Villaumbrosia | Product School 00:27:00

The switching cost is just so high right now that you have to be so much better than what's already out there. As companies today try to vibe code their own solutions and realize there's a big part of the solution that isn't just the front end — where do you see your moat in this new AI era, and how do you position yourself against competition as you expand your solution?

Breadth as the New AI Moat

Jay Choi | Typeform 00:27:30

Part of this was just being really honest: I don't think we're going to win every use case. Some things may get vibe coded, especially if they're very simple. We needed to start from a clear-eyed conversation about the fact that we won't win everything. But where we want to win and add more value, let's spend a lot of time and energy on that.

Contrary to what product strategy I've used in the past — which is usually to go super deep in a use case and make that one really great — I'm finding personally that going broad actually starts to matter a little bit too. The broader surface area we create, the harder it is to replicate. Maybe it's really easy to vibe code a lead generation form for an event or a podcast. But the integration into putting it into your CRM, then enriching the lead, then scoring that lead, then generating an automated email that goes out right after — 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 has before.

The second moat is integrations. We still live in an ecosystem world, and organizations want to live in an ecosystem — they want us to be part of it. Third is security posture and enterprise-grade admin infrastructure. It's still hard to replicate that when you're willing to take on your customers' information but you're not a security expert. If you think about Typeform's penetration — we're in ninety-five percent of the Fortune five hundred, with eight hundred people able to access our workspaces and tens of millions of people on the platform. That scale, that latency, that ability to handle security — it's just challenging to replicate.

Carlos González de Villaumbrosia | Product School 00:30:00

Ninety-five percent of the Fortune five hundred?

Jay Choi | Typeform 00:30:05

Yeah, ninety-five percent.

Enterprise Adoption: Security, Admin, and Shadow IT

Carlos González de Villaumbrosia | Product School 00:30:10

That's remarkable. Before you mentioned that stat, I was actually going to ask about it, because one of the things we experience as a business when we do trainings or implementations for large enterprises is that they've made their bets on certain tools. They say, "We're a Microsoft shop — maybe there are other tools out there that are better, but we need to make it work with our current tool stack." I'm curious how you've been able to speak the enterprise language and adapt your product to cover all the bases.

Jay Choi | Typeform 00:30:45

When enterprise customers ask for enterprise features, a lot of it comes down to admin, security, privacy, and SSO. SSO is a big driver — people want single sign-on. When they want to move from a free license to something more, it's usually: "We want enterprise-grade security and enterprise-grade admin functionality." They want to remove the dozens or hundreds — sometimes thousands — of Typeform free accounts and move away from shadow IT into a centralized platform. That's been a fairly pragmatic driver for us on the enterprise front. Our ability to talk about our security posture is usually a big driver for us in those conversations.

Carlos González de Villaumbrosia | Product School 00:31:50

I love this. I think it's been a masterclass on how to survive and thrive after the SaaSpocalypse. Thank you so much for your time.

Jay Choi | Typeform 00:32:00

Thank you, Carlos. It has been such a pleasure. I've really enjoyed talking to you — I love your questions and engaging with the whole broader community.