Linear COO on Rebuilding the Product Development Lifecycle for Teams and Agents — From Issue Tracker to Shared Operating System | Cristina Cordova | E299

In this episode of The Product Podcast by Product School, Carlos González de Villaumbrosia sits down with Cristina Cordova, Chief Operating Officer at Linear, the product development system built for teams and agents. Linear raised $82 million in a Series C round in June 2025 at a $1.25 billion valuation. The company has been profitable since 2021, and serves over 20,000 paid business customers, from seed-stage startups to Fortune 100 enterprises, with a team of just 140 people. Before Linear, Cristina joined Stripe as one of its first employees, and led Platform and Partnerships at Notion.

What you'll learn:

• Why keeping headcount intentionally lean is a strategic advantage
• Replacing traditional interviews with paid two to five-day projects
• Why PMs are the fastest-growing power users of agentic tools

Key takeaways:

• A small team is not a small business. Revenue, customers, and growth rate matter more than headcount.
• If you fully delegate your AI thinking, you lose your native understanding of how these products actually work
• Agentic workflows are now the default, not a feature. The companies that treat them that way will pull ahead.

Credits:
Host: Carlos Gonzalez de Villaumbrosia
Guest: Cristina Cordova

Social Links:

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

Transcript

Introduction

Cristina Cordova | Linear 00:00:00 I feel like I've joked that to keep up with everything that's happening in AI, you basically have to be unemployed. If you fully delegate all of your thinking around AI, you'll have kind of no native thinking of how these products actually work on their own. Sometimes people mistake that for saying you wanna keep the team small, that means you want the business to be small.

Cristina Cordova | Linear 00:00:00 No, you can keep a really small team and have a very large, very impactful business. Those two things aren't one and the same. Typically at a company you might have an onsite interview as your last stage. Maybe that's like four to five hours worth of interviews. At Linear, we have a project that you work on with us anywhere from two to five days, and we pay you for your time.

Cristina Cordova | Linear 00:01:00 There are some people who are in the Tesla age of AI, where it's like, "Ooh, my car can park itself for me, but I'm still driving it." And then you have the Waymos of the world, which are like, "I'm in the backseat. I'm not even looking at the road. I'm not paying attention at all, and I'm just letting the AI do the work for me." And so we have to build products for every single one of those people.

Carlos González de Villaumbrosia | Product School 00:02:00 Hey, this is Carlos, CEO at Product School and your host on The Product Podcast. Today's guest is Cristina Cordova, Chief Operating Officer at Linear, the product development system built for teams and agents. Linear raised $82 million in a Series C round in June 2025 at a $1.25 billion valuation. The company has been profitable since 2021 and serves over 20,000 paid business customers, from seed-stage startups to Fortune 100 enterprises, with a team of just 140 people. Before Linear, Cristina joined Stripe as one of its first employees and led Platform and Partnerships at Notion. In our conversation, we cover why keeping headcount intentionally lean is a strategic advantage, replacing traditional interviews with paid two to five-day projects, and why PMs are the fastest-growing power users of agentic tools. Let's get into it.

Carlos González de Villaumbrosia | Product School 00:02:00 Cristina, thank you so much for joining me on The Product Podcast.

Cristina Cordova | Linear 00:02:00 Of course. Happy to be here.

Carlos González de Villaumbrosia | Product School 00:02:00 I'm excited. I had your CEO, Karri, on the podcast around a year and a half ago, right as Linear was starting to become the go-to platform for product teams. A lot has happened in how teams build software, especially now with AI and agents. So I'm really excited to explore what that really means and how you've been scaling your team, your product, and your go-to-market.

Cristina Cordova | Linear 00:02:00 Excited to talk more about it. A lot has changed and evolved in our business.

Background: From Stripe and Notion to Linear COO

Carlos González de Villaumbrosia | Product School 00:03:00 So you joined Linear as COO around two years ago?

Cristina Cordova | Linear 00:03:00 Almost three years ago. Next month it'll be three years.

Carlos González de Villaumbrosia | Product School 00:03:00 And before that, you worked in ops and go-to-market for companies like Stripe and Notion very early on. So I think you experienced some of that hypergrowth.

Cristina Cordova | Linear 00:03:00 Yes, and product too. I had pretty much done a lot of different roles. Unfortunately for the other people I managed, I was never a designer or an engineer, but I happened to manage designers and engineers. Luckily they put up with me. At Linear, I'm very much on the go-to-market side — everything that's not engineering, product, and design. It's nice to be able to switch that up depending on the company and the context. Most of my roles have been somewhat on the business side, somewhat on the product side, a little bit in between. Linear is much more on the business side, so a little more focused.

Carlos González de Villaumbrosia | Product School 00:04:00 At the same time, we are building for builders, right? So you are somehow involved in the kitchen regardless of the title.

Cristina Cordova | Linear 00:04:00 Yes. That was one of the things that attracted me to Linear. I loved my experience at Stripe working with these very discerning developers who were kind of the first people to really fall in love with Stripe as a product. In coming to Linear, I realized that our customer base is very much the same audience, but actually broader. At Stripe you'd often be selling to the payments team within a company. At Linear, we're selling to anyone in product, design, and engineering. Everyone on those teams has familiarity with Linear and what we're doing. I'll get text messages from friends at certain companies saying, "Oh, I saw you shipped this update or that update." It's a really fun part of getting to do this job — to build for so many people who also build really wonderful products.

Carlos González de Villaumbrosia | Product School 00:05:00 You are the first COO I've hosted on the podcast, because you have such a unique background. You cover ops and go-to-market, but you also have a really strong understanding of product, and that's what is going to make this episode special.

Scaling Lean: 25,000 Customers, 140 People

Carlos González de Villaumbrosia | Product School 00:05:00 I want to start with the different ways you've been scaling teams. At Stripe and Notion, there were a lot of people, but maybe balanced to the revenue you were generating. As I look at Linear today, you're at over 20,000 paid customers and a team of barely 120 people. That sounds very lean compared to your revenue scale. I'm curious to know how you're going about scaling headcount differently now.

Cristina Cordova | Linear 00:06:00 Well, we're 140 now, so maybe we've been growing fast. But I talk to a lot of people and they're shocked when I say that. They're like, "Oh, you're not like 400 people? You're 140 people?" That's surprising to them. I think that's the power of a great brand and a product that touches so many different people inside of companies — a product you use day in and day out. And 25,000 paying customers, far more that are trying Linear for the first time. That's really exciting to see.

Linear at its core when I joined was a very product-led growth company. The focus was on how you build for these people and these teams. A lot of our early-stage customers looked a lot like us — same size, same stage. That works for a company that's early on in how it builds. Stripe was very similar. We were building for other early-stage companies like ourselves. And then you have to continue to build the company to scale beyond a customer that looks like you anymore. That's what we're starting to see. We have Fortune 100 customers now, many public customers, which is exciting. They have strong standards when it comes to security, privacy, how their data is used, and what they feel about AI products. We're trying to build for those kinds of companies, which require a lot in terms of thinking around the product and how we evolve it for different use cases.

For us right now, it's scaling the actual product surface area, ensuring that Linear can be used for teams and organizations and then multi-thousand-person companies. That's really our focus.

So one of the things you see at a lot of other companies is they try to accomplish that goal by just bringing on more bodies, hiring more people. We've taken the tactic of asking, "Is this really a role that is absolutely needed right now?" versus hiring people because, "Oh, that looks like a job." And just being very conscious of how we're building teams and how we're growing the company — and keeping a really high bar. Turns out when you keep a really high bar in terms of who you're hiring, that naturally slows down your growth, because you can't just hire lots of people. There aren't an infinite number of incredibly talented people who are the right fit for Linear at this stage. Which means we have to wait around for exactly the right people to be ready for us and for us to be ready for them.

We've kept the company very intentionally lean, and I think sometimes people mistake that for saying you wanna keep the team small, that means you want the business to be small. To that I say no — you can keep a really small team and have a very large, very impactful business. Those two things aren't one and the same. A lot of people ask how big you are in terms of company size and employees, but they're not really asking how many customers you have, what's your revenue, and how fast are you growing. Those are the metrics that matter more than how many bodies you have on your team.

Carlos González de Villaumbrosia | Product School 00:09:00 We're both in the San Francisco Bay Area, and I remember many years ago it was all about how much money you raised, how big is your team. Nobody would ask about revenue or even if you're profitable. But speaking about that bar, when the margin for error now is even smaller — if you're hiring fewer people, they better be great. How do you approach that process now, maybe slightly differently than when you had a little more leeway?

The Work Trial Hiring Model

Cristina Cordova | Linear 00:10:00 We've always had a probably longer, more strenuous hiring process. In particular, we do something called work trials that everyone participates in. Typically at a company you might have an onsite interview as your last stage — maybe four to five hours of interviews. At Linear, we have a project that you work on with us anywhere from two to five days, and we pay you for your time. We're not asking people to do free work for us. What we're trying to do is give them a real project. For an engineer that might be a product feature we actually want to ship, relatively small in scope, something you could get done in a few days but something we'd actually push out the door. For marketing it might be, "Here's an audio recording of a customer interview, turn that into a customer story," and we might actually ship it at the end.

So we try to mimic what feels like real work, and then get a really good sense of what it's like to actually collaborate with someone. Because it's not just about the end output. That's maybe half the value — understanding whether this person can produce great work. It's also in the how. How does this person collaborate, communicate? What happens when you're in a room and we're intentionally disagreeing with you? Sometimes the right answer is to push back and say, "No, here's why I think I have the right course of action." And other times you're like, "Actually, that's a good point, and maybe I wanna think differently about it." It's more about how that interaction works — do we feel this is someone who'll challenge us but we can also challenge them?

Our process has always been strenuous in a way. It's not an easy process to get through. It's not a quick process. So if you're a candidate just trying to collect as many offers as you possibly can, maybe you don't come to Linear, because it's just gonna be more work. And we think that's okay. We're not looking for every talented person. It's finding someone who's talented and who is ready for the kind of challenge this work presents. That's what matters most to us.

We've kept our process very similar over the years, but we've tried to add efficiencies. We'll often repeat projects. You're not usually getting a completely unique project that no one has ever done before. We're doing work trials pretty consistently, so it helps to have a consistent rubric for judging performance on those projects. Having them be repeated means we have a good sense of what the upper echelon of candidate looks like versus someone who's on the edge. Those pieces are important to understand.

It also optimizes for people who really want the role and the opportunity to work with us. It also takes a lot of time from us to collaborate and work with that person. So we hope it shows to candidates that we really care about the process and we care about finding the right people — and that the thing that makes working at Linear great is being able to work with people who you would really want to spend years and years of your life with.

Carlos González de Villaumbrosia | Product School 00:14:00 I think it also forces you and your recruiting team to be super picky, because if you're going to spend a lot of time with someone, you better be right. So for you as an executive, there is so much pressure out there. It seems like we are demanding all of these AI skills from individual contributors. As an executive, how do you actually catch up with all the new AI stuff? How do you make time to build yourself and make sure you're staying super relevant?

Keeping Up with AI as an Executive

Cristina Cordova | Linear 00:14:00 I feel like I've joked that to keep up with everything that's happening in AI, you basically have to be unemployed. There's just no way to keep up with everything. We have a channel in Slack where people will just post different things that are happening, and it's many different updates throughout the day every day. It's a lot.

I feel like you have to have surface-level knowledge across a wide variety of things, and then as it comes to your competitors, or very deep partners, or other folks in the space that you really care about in your ecosystem, you tend to go deeper and try to really understand what are they launching, and then try it. Those are the pieces I tend to pay attention to. There's a lot getting released. Not all of it's getting used or making an impact in a lot of situations. So often there are people I tend to follow who try a lot of these products out really early, and I care about their opinions on what's making an impact — specifically in the software development process and life cycle. That's probably what I pay attention to most. I probably pay more attention to agentic coding specifically than other types of work.

For me, I try out as many things as makes sense. I care a lot about privacy and security, so there are certain things I have a pretty high bar for. Some things I'm like, "I'll leave that for my personal computer and not my work computer," if there's something I really want to try and don't exactly know how it's gonna go. But I think you just have to have a very experimental approach. You have to have people within your company who are really excited about new tools and new products and what's coming out, and they share them with you. And then at the same time, a willingness to try things and see how those products are shifting and evolving over time.

Carlos González de Villaumbrosia | Product School 00:17:00 Cristina, last time we spoke, I remember you mentioned that you've been using Claude Code, and I was like, "Oh my God, that sounds incredible," especially for someone who doesn't come from a technical background. Is there anything specific that you do to carve out time to make sure that you not only read and see what others are doing, but you actually do it yourself?

Doing AI Work Yourself: The Codex Experiment

Cristina Cordova | Linear 00:17:00 Generally I think about this as: someone in my role often by default delegates. That's kind of what you're supposed to do. That's what your job is. But if you fully delegate all of your thinking around AI, you'll have kind of no native thinking of how these products actually work on their own. I don't have a lot of personal use cases for AI. People always talk about trip planning or shopping or things like that, and I try to use AI for those things, but it's just totally different from using AI for work purposes. So I think by default I actually try not to delegate if there's something I see as an opportunity and I think I can do it myself.

A good example of this recently: there was a study or research report that came out saying one of our competitors was used by all of these companies based on it being mentioned in their job descriptions. And I was like, "That's weird, because a bunch of these companies on this list we know are our customers. And we're not showing up on this list. Why not?" So I wrote a prompt in Codex and said, "Look at all of these companies' careers pages, go through every single job description, and create a table. When any of these tools are mentioned, mark those down, including us, and any competitors as well. Come back in 10 minutes or so." Not that many companies, not that many job descriptions to go through — and it was complete. Now you have this question of where did this data come from? By default you'd be like, "Who knows how they ran their survey or did their analysis?" And now you actually have the truth. It looks like they're not actually using that competitive tool — it was mentioned in a list of tools like X, Y, and Z. So I'm okay with that. I don't think we're actually losing customers to this competitor.

It's a good example of saying, "Well, why can't I just do that myself?" And it helps being a small company. We have a data person and I could say, "Hey Tim, would you like to work on this project?" Or I could just do it myself. That's what a lot of these tools give people access to do.

When we look at the data we see within Linear itself, the biggest gaps between people who were using AI agent features six months ago and are now heavily using them — that spread of people who've come really far in a short period of time — are typically in non-engineering roles. A lot of people in engineering came up on this in a particular way because they had tools like GitHub Copilot for a really long time before some of these other AI tools started taking off. But PMs and product folks in particular have come a really long way in a short period of time — in the last year. I'm seeing much more growth in the agentic product features they're using heavily. And I think that's because historically they wouldn't default to doing things on their own. Now you see them default to doing things on their own rather than saying, "Hey, can an engineer do this? Can a data person do this? Can someone slightly more technical than I am take this on?"

I'm actually seeing similar things on my side. AI can have this supercharging effect for executive leaders — all of this information gets made available to you that historically you would have to wait for somebody to go find. Like, I can upload all of our Gong call transcripts into a custom ChatGPT project and say, "How are the conversations that our salespeople are having with customers changing in the past six months compared to the six months before? What products are we talking about that we weren't talking about before? What are the customer concerns that are happening now, and how are those different from six months ago?" Maybe you get some of it via anecdote from your sales team, but it's actually really powerful to understand what's really happening without the bias. The deals I pay attention to are typically all of our large deals. I'm not really paying attention to the long tail. To be able to do that kind of analysis that quickly is really powerful.

Go-to-Market: Brand Evolution and Enterprise Messaging

Carlos González de Villaumbrosia | Product School 00:23:00 I want to shift gears to go-to-market, because that's another important area of your scope. I know you care about brand. Brand is this thing that comes and goes — sometimes companies seem to care a lot about brand, then they seem to care more about demand gen. And before, you mentioned that your early customers looked a lot like your company. But now as you expand into SMBs and large enterprises, I'm curious to know how you are evolving your own brand.

Cristina Cordova | Linear 00:23:00 We invested a lot in brand early. And I think that is partially an investment in design — what is the look and feel of the website, how does that match the look and feel of the actual product, so that you have a taste for what this brand feels like before you actually get into the product itself. We had great spokespeople in our founders, really trying to talk about the product building process, their beliefs in building companies, and just trying to be out there more than a lot of other founders, which also helps. And that helps you attract other startups and early-stage companies.

But is the VP of engineering or the CTO of a 10,000-person company hanging out on Twitter following our tweets? Maybe not. So how are we going to acquire this customer in a different way? Well, we've learned actually that a lot of these people do hang out on Twitter. You'd be surprised at how many say, "Oh, I saw that blog post." And we didn't email it to all of our customers — the only place we shared it was on social media, so that's probably where you found it.

And then it's about thinking about channels. How do you ensure that the people getting certain kinds of messages are the right people? When we talk about the type of Linear user who loves keyboard shortcuts and really likes Linear in dark mode — those kinds of things — that's probably the individual contributor who cares about those things. That's probably not your VP of engineering or CTO. That person cares about efficiency gains across the entire organization, about how their teams are making progress with agentic coding. Maybe they care about the number of lines of PRs their teams are writing these days — I don't know if that's the most valuable metric, but if that's what they care about, how do we start speaking in their language?

So those are the kinds of things we see with a more senior buyer. They're thinking a lot about the migration. If I'm leaving tool X to come to Linear, how much is that going to cost me in terms of time and energy? And on the flip side, is that going to be worth it?

So you start to really change your messaging from feature-based — which is where we started. The keyboard shortcuts and things like that will always have a place in your product marketing. But you really need to shift into value-based messaging over time. What is the value that Linear is providing? And how do we talk about that, not just by saying "Linear offers this value," but ideally by having other customers share their own stories about how Linear delivered that value for them — some kind of social proof on top of that.

Sometimes this just takes time. At the beginning, you have no customers, so all you have are features. Then you get customers telling you you're providing value. Then you get customers who've been with you long enough that they're willing to put their logo on your website, have a customer story, and do a video — and have social proof. It doesn't happen overnight. But you have to make that journey if you're continuously going upmarket.

Carlos González de Villaumbrosia | Product School 00:27:00 Even when you talk about the customer logos, they have to be the right customer logos. If you go to a Fortune 500 with a bunch of startup logos, it might even be scarier. So I want to double click on your point around migration. When you are selling into companies that do not have a tool already, that's one thing. But if they already have their teams and their context — even if the tool is not great — sometimes just the switching cost is enough to say, "You know what, I'm going to stay." So how do you navigate those types of migrations, especially in the large enterprise segment?

Enterprise Migration Strategy

Cristina Cordova | Linear 00:28:00 The first piece is you have to understand why it's worth it to switch. What is your story for whatever product you're trying to sell, and why is that switch worth it? What is this additional value your product serves and provides to your customers that goes beyond, at the beginning, "It's like a nicer, prettier version of what I have today." That's not gonna move the needle with a CTO or VP of engineering.

So you talk about what is going to move the needle. How about 10X your building capacity? How about moving fast as an organization without necessarily losing control, which you might care about if you're a CTO or CISO? How about double the number of active users? If you want this to be a place where all of your context lives in your organization so you can utilize it for agentic workflows, then you care about being a tool that people actually want to use. How do you go from bug reported to bug closed and resolved as fast as possible? If we can 3X or 4X that, how much value does that bring in terms of increasing your product quality and your experience serving customers? Those are the things you have to explain — why is the switch worth it?

Then really talk about the companies that are using your product to pull ahead. Ideally you have customer stories from customers who are large enough to matter, and who are also in different spaces — healthcare, fintech, SaaS, consumer. Because every customer sometimes tends to think they're a special snowflake. "That B2B SaaS business isn't like my consumer business." Often the workflows we touch are pretty similar regardless, but I still think they like to see customers that look like them.

Then you have to really talk about how the switch actually works. We tend to talk about piloting Linear so you actually have a sense of whether it's working for some teams. You try out with three or four teams, run a successful pilot, and then we give you resources for how you're fully going to migrate your team and make that a really seamless transition. And we have tools to sync your Linear issues with your legacy product, making that really easy for your team.

It's really a combination of saying, "We're here for you." There's a customer success portion of it, a tooling portion and things you need to build to make migrations really easy. And then you need to make sure the value is there so that regardless of how easy or not easy you think it's going to be to migrate, you know in the end it's going to be worth it.

Carlos González de Villaumbrosia | Product School 00:31:00 It seems like these days even getting through procurement and checking all the boxes is not enough. Successful implementation into the client and making sure they adopt and see the value fast is critical. I'm seeing more and more companies having their own teams — you can call it forward-deployed engineers, you can call it something else — to go in and ensure this is actually working. For large enterprises, is there something that you think can really move that needle for them to feel really good after the pilot is approved?

Cristina Cordova | Linear 00:32:00 We have solutions engineers at Linear, and they're a really key part of our workflow. I've never been at a company that has extended to the full forward-deployed engineer model. In a lot of ways it's about showing you how easy a switch can be by just demoing a switch — that's really helpful. There's also the other piece, which is really ensuring that you have the guidance — what are the materials this customer needs to make this switch? If I know I need to switch out my incident management software and switch out a deep API integration or what have you, you just need to be able to check those boxes or prove that doing that switch in terms of APIs is not gonna take that long.

So sometimes our solutions engineers will demo that type of integration and build it in less than a day, and say, "Hey, see? We built it for you. Look how easy it can be. We'll actually demo it for you so we can prove this is real, this is live." We're not typically embedding with a customer and saying, "We're gonna touch your code." Typically that's not what we're doing. 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.

Carlos González de Villaumbrosia | Product School 00:34:00 That's a great point, because I think product-led growth as a motion can work well at a certain company size, but it gets to a point where you also have to do certain things maybe a little more manually without overdoing it. Because ultimately, if you have to do too much, this might not be the right value prop.

Cristina Cordova | Linear 00:34:00 Right. Yes. And we're not a services company. I think there are a lot of software companies that do services, and that might be in our future, but we don't wanna become a services company. For every company you have to figure out what that line is. "Okay, we'll do this for you, but we're not gonna be touching your own code" — which is what we always said at Stripe. "We can build a demo for you, we can show you how easy it's going to be, but we're not gonna build the integration on your own stack." Figuring out what that line is is important for any organization. There are a lot of really great businesses out in the world that will do all those things for you. You just have to figure out the style of company. It's really hard to have a more constrained team while doing services. It can sometimes not match the ethos of the company.

Carlos González de Villaumbrosia | Product School 00:35:00 It's funny, we do that as a business. We do the training and the implementation as a service. We don't have our own product, and I think it's like doubling down on your strength is something I'm getting from your answer. I want to cover the final theme, which is product. I remember when I had Karri on the pod like a year or so ago, the headline on the website was something like "the system for product teams to build." And I checked, and it says something similar, but it's "for product teams and agents." So I want to know what that means in practice.

Product Positioning: "For Teams and Agents"

Cristina Cordova | Linear 00:35:00 So it's the product development system for teams and agents now. When Linear started, it was "the issue tracker you'll actually love to use," something of that variety. It started out with this very narrow use case — issue tracking at its core. We focused on what we thought we were great at, which is building a product you're actually gonna love using. We kept it really simple and really narrow, which is actually what I advise a lot of early-stage companies to do. Not everyone's gonna be your customer, not everyone's gonna wanna take a bet on you, but the early-stage startup that thinks Linear might be interesting — that message will appeal to them. That's who you're trying to go for on day one. You're not gonna win Fortune 100 companies with that message.

So eventually your message has to evolve. The best way to evolve messaging when you're a company like ours — PLG, very much focused on going upmarket, but you don't wanna forget the early-stage startups — is to find what broadly appeals to a wider audience. I think the idea of a product development system makes you go, "Hmm, what is that?" Is it an issue tracker? A system where I can do my roadmapping or product planning or bug intake? The reality is it's all of those things. We're creating what we feel like is a category for what we're doing that feels right for the moment.

The "for teams and agents" part is really about taking Linear into this new era, the AI era. A lot of people look at a business like ours and say, "Oh, this is a seat-based business. I know what this is." And yes, it is. Slack, lots of other companies operate this way. At the same time, we know we're moving into a new era, and it's important we make that clear to people. We don't wanna shove AI in everyone's faces — people are tiring of that in a lot of applications these days. But it's clear that our audience knows the power of agentic coding in particular, and it's important we want them to know that agentic workflows are the default in Linear. That's the change in the messaging. As you start to grow, you tend to get broader.

What we're seeing is a sea change in the market with AI and agentic coding workflows. We're ultimately trying to say, "We are the right tool for that sea change." If you as a company today say, "The way we build products is so different from just a year ago," I want you to look at Linear and say, "Actually, Linear is the tool for this new era."

Carlos González de Villaumbrosia | Product School 00:40:00 I want to break this down with you, because back in the day a lot of SaaS companies would try to replace Excel — that seemed to be the enemy. And they would all position themselves in the middle and say, "We have integrations with everybody else." Now I'm seeing a lot of SaaS companies saying, "We are the agentic platform," whatever that means. But they say agents and platforms and put themselves in the middle and say, "We integrate with everybody else." So as we look into the specifics of your product, I'd love to learn more about those details around integrating with coding agents, setting up the right context, and really understanding how this handoff process is really different now.

Building for the Agentic Era: Triage Intelligence and APIs

Cristina Cordova | Linear 00:40:00 We started our AI exploration internally with what feels like a very difficult data problem — typically intake for most companies. You have a lot of things coming in: bug reports, feature requests. They're coming in from Twitter, from your support channels, from shared Slack channels. Ideally they're all coming into a central source of truth so that you're able to say, "Hey, we've gotten 10 bug reports for this thing in the last hour. It seems to be something we need to urgently fix." You're not gonna know that if you have a pretty big team and different people seeing this across different customers.

So we built a feature called Triage Intelligence that helps people figure out where a new bug report should go, who on which team within the organization it belongs to, what labels should be attached, whether it's a bug or a feature request or something else, and what the right status is. We have all this data around where things have gone historically in your organization, done manually by humans — and we can take that data and say, "Actually, we know where this should go going forward." That's a pretty big shift for a lot of customers. Some want to click Approve on every one of these suggestions. Others say, "Linear, I just want you to do it all for me."

And that's the spectrum we see in AI and agentic use cases — what we call self-driving SaaS. Some people are in the Tesla age of AI, where it's like, "Ooh, my car can park itself for me, but I'm still driving it. I'm still in the driver's seat, in ultimate control." And then you have the Waymos of the world — I'm in the backseat, not even looking at the road, not paying attention at all, just letting the AI do the work for me. We have to build products for every single one of those people.

We started with Triage Intelligence because we felt it was this universal use case — everyone has too many bug reports and too many things to manually triage. And then we saw the rise of agentic coding workflows, and the power they had just in building our own product — how much better they were becoming as models improved. So we built a platform. I feel like I'm making your joke for you from earlier — "We're a platform for agentic..." Anyway.

By that I mean we shipped a set of APIs. A number of companies building coding agents said, "Hey, we wanna build on this so you can be within Linear and automatically delegate an issue of work to an agent and have the agent get started on that work within Linear." And then once they're assigned to that issue, you can be very much hands-off and come back when there's a PR ready to review. That was really powerful. This was in the middle of last year, so not everyone was ready for it. But now more and more customers are ready for it every day, and it's becoming a core part of how people build products on Linear.

Then we came to the conclusion that a lot of companies are also building their own agents internally. We wanted to make sure our APIs work for those totally bespoke, custom-to-your-organization use cases too. Companies like Ramp and Coinbase, who are on the bleeding edge of these workflows without necessarily being AI companies at heart. They existed before AI existed in the way it does today. Turns out the same APIs that work for Codex and Cursor to build their third-party coding agents also work for Ramp and Coinbase to build their custom agents on our platform.

What's coming next for Linear is giving you access to every kind of popular agentic coding model and letting you build with it on Linear. We can integrate with all of the most popular models and ensure that you can code directly from within Linear itself. We're not becoming a frontier model company — we want to partner with everyone our customers want to use and make those agentic coding models available to them from within Linear.

We also launched Linear Agent a few weeks ago. That's a great way to start using our agentic workflows without necessarily using them for code. Coding solves the problem for a lot of users we have. But you have a lot of PMs in Linear saying, "Hey, I just had this meeting. Here's my transcript. Can you just create all of these as Linear issues for me?" And we'll do it. Take all my notes from this meeting with my team and create a PRD. Here's my skill — I now have a skill for creating a PRD in Linear, so you know exactly what a PRD should look like for my team and my company. We're starting to see more and more of these kinds of workflows arise from the agent launch. And it's a really powerful way that different people who are maybe not your typical product builders — not PMs, designers, or engineers — can participate in the product building workflow as well.

Carlos González de Villaumbrosia | Product School 00:47:00 I think the shared context piece is huge. Allowing people that maybe didn't create the skill to benefit from it, or see live the evolution of a feature or something that is being worked, and contribute to it without having to be the ultimate expert in setting up the environment — that unlocks a lot of non-technical people to start building.

The other piece you mentioned about UI is super interesting, and I want to double click on it. We're seeing how it's possible to use agents within your own product, and now it seems like it's also possible to use similar agents within another interface that you might not own — like from Claude or from OpenAI, or even Slack or Microsoft Teams. So how do you think about the role of the UI? Where do you want to play? Are you trying to be the place where people go and use the agents there, or is it more about being able to use your agents in other places?

The Agent Ecosystem: Where Linear Plays

Cristina Cordova | Linear 00:48:00 I think it's all of the above. Our first homegrown agent was actually the Linear agent for Slack — we didn't even build it for our own product surface area. We built it for Slack's, because we use Slack a lot and we know that a lot of our users use Slack a lot. You've been in one of those 20, 30 message threads in Slack, and at the end someone can say, "@Linear, can you file all of this as a set of issues and assign it to me?" And we'll do that. Then we built one for Gong, and then Intercom and Zendesk.

We know we're not gonna be the tool for everything. And I think it's a mistake for anyone today to assume that everyone has to do everything through your product. That's just not how ecosystems and technology tools work. If anything, there's a proliferation of tools now, not a consolidation. So we want this agent workflow to exist in all of the tools you use to run your business. And at the same time, where it makes sense for us to build agents into Linear to do core Linear workflows — whether it's agentic coding, or the core Linear agent to write your PRDs, triage for you, and all of those things — we'll do that too.

So we're gonna answer to our customers when they say, "It would be great if you could just do this for me automatically, and I could just ask the agent, and it'll do it." And I'm like, "Great, we'll do it." You might be able to take some of these same things and do them in Claude Code or do them in Codex or other tools. I think we're all gonna have agentic workflows, and maybe you do them in one tool and maybe you do them in another. But I think where you miss out is not having them at all.

We want Linear to be the place where all this context and data exists, but also where you actually get to do something useful with it. The only way to do things at a really high level of scale is a lot of these agentic workflows. I can have a list of 30 different things and say, "Change the label on all of these, assign them to this brand new team that we've created to handle it." These are things that used to take hours and hours for a lot of teams that just don't take that amount of time anymore. We'd rather teams focus on building rather than managing Linear.

Carlos González de Villaumbrosia | Product School 00:51:00 You've got range. It's been a pleasure to go through how you're scaling people ops, go-to-market, and product all in the same conversation. Thank you so much for your time, Cristina.

Cristina Cordova | Linear 00:51:00 Awesome. Thank you, Carlos.