Inside How Figma Builds AI Products | Yuhki Yamashita (CPO Figma)

One of the things that I think is really helpful for a PM is to realize that, you know, the way that they can help a designer is actually to think about the constraints that are producing a particular design. Because sometimes like a design might not make sense, but it might be because the designer is operating under the wrong constraints. Maybe they have to stick to this design system or they're limited what they're able to explore because they've been told that they have to kind of like just stick to these box or there's some feasibility thing. So that's one thing I think a lot about and encourage DMs to think about, which is like, hey, like you can actually be a partner, even ask your critiquing a design by kind of critiquing the constraints rather than design itself. So that's kind of one thing I think about. And then, you know, the other thing is like really speaking in the language of users, right? Like when I think about, you know, some PM designer interactions are very oriented around goals and KPIs, but when I think about designers, they're so human centered. And so how do you actually, you know, shift the language more towards a user is feeling or, you know, what goal they're trying to achieve or a particular user journey?

My guest today is Yuki Yamashita, Chief Product Officer at figma. Yuki and I talked about how Figma created a culture where employees love their craft, how they build AI products to solve real user problems, and how they plan to make design accessible to everyone. Yuki is a fantastic storyteller and shares an inside look at how figma's amazing product culture works. If you enjoy our conversation, please like and subscribe to support the podcast. All right, well, welcome Yuki. Really glad to have you here.

Yeah, thanks for having me, Peter.

Yeah, you know, since you work at Figma, let's start this question. You probably have managed both PMs and designers. I'm curious if you have any thoughts on, you know, how PMs can be great partners for designers and vice versa.

Yeah, yeah, I have a lot of thoughts on this. I mean, I think for a pm, you know, one of the things that I think about is, you know, when I've written a lot about this, but basically, you know, it's very easy to, it's very easy to kind of critique designs and make it feel really personal or really get detailed. But one of the things that I think is really helpful for a PM is to realize that, you know, the way that they can help a designer is actually to, you know, Think about the constraints that are producing a particular design. Because sometimes like a design might not make sense, but it might be because the designer is operating under the wrong constraints. Maybe they have to stick to this design system or they're limited in what they're, you know, what they're able to explore because they've been told that they have to kind of like just, you know, stick to these, this box or there's some feasibility thing. So that's one thing I think a lot about and encourage PMs to think about, which is like, hey, like you can actually be a partner. Even ask your critiquing a design by kind of critiquing the constraints rather than design itself. So that's kind of one thing I, I think about. And then you know, the other thing is like really speaking in the language of users. Right. Like when I think about, you know, some PM designer interactions, maybe PMs are very oriented around goals and KPIs, but when I think about designer, they're so human centered. And so how do you actually, you know, shift the language more towards what a user is feeling or, you know, what goal they're trying to achieve or particular user journey? So those are some things that I think a PM can do to be a better partner to a designer. Got it. And then in terms of a designer, I think that, you know, recognizing that, you know, PMs are often thinking about like goals, they're thinking about trade offs, right. And you know, surfacing also borrowing from some of that language. And you know, Julie Joel has a lot of great tips that she's written about that I really love about how, you know, designers can contemplate the way PMs think. And like, I think, you know, I really highly encourage people to read that as well.

I think the traditional PM role has been all about moving the numbers and that kind of stuff. But I think one of Figma's values is love your craft. So how do you kind of screen PMs for this value during the hiring process?

Yeah, I mean it's really interesting sometimes it's very self evident. Right. Like in some cases we've had PMs share some of their work or do a short assignment and kind of attention to detail they might have or even the attention to these small little things that a user might care about. The passion for that, when that shines through, it's really obvious that they really care about those kinds of details and they don't need a metric to tell them to care. So there are some subtle cues like that. And then I think the other thing is just whether or not you really get excited about building great user experiences, right? And I think, like, some people are just really, really just are thrilled to like create amazing experiences and others are doing it in service of a goal. And, you know, I think some people, you know, when I find people who just inherently want to make the experience better for the user, who are very natural at thinking about kind of like through the user's eyes, like, those are the people who tend to also, quote, love your craft. But yeah, those are some things I look for.

Got it. And you know, with your pm. Org, like, how do you balance this craft value with moving the numbers with maybe like internal optics, like having a great product review, if digital or something? Like, how do you evaluate your PMs with all this stuff?

Yeah, yeah. I mean, hopefully it's not about the product review per se, right? But I mean, I think that everyone has goals, right? And it's just that sometimes those goals are much more measurable and sometimes those goals are a little bit less measurable and high level. And so for me, it's really about are you setting good goals and are you demonstrating your ability to hit them, even if it's not always completely perfect, how you measure it? That's the thing. Even when we do things like building great user experiences, it comes from a place of, oh, I want to solve this particular problem, or I want to address this sentiment from the community. And so there are ways in which we want to have confidence that after having launched it, you actually met those goals. So those definitely are ways in which, like, I evaluate yams, right? Because what we're trying to do is solve problems and, you know, teach you, solve the problems. And it's just the case that unfortunately not every problem is as measurable as others. But that's okay. And that shouldn't prohibit you from being a good VM.

So this big AI release that you're doing for FigJam, and maybe we can start with how FigJam, like briefly evolved from Figma in the first place.

You may remember it's the height of the pandemic. We had quarantine and there are a lot of people using figma in figma design, as they call it, in really interesting ways. People were having happy hours there, you know, and just hanging out and playing games. They're running retros and meetings and standups and, you know, it was kind of this interesting way to have this third space, if you will, where, you know, it was an immersive enough environment, but it wasn't kind of, it wasn't overbearing, but you could kind of get everyone engaged because you could see people's cursors, you could react, you know, you could kind of respond in real time to things. So we just found that in a lot of our customers and people in our community were using figma in this really interesting way. And that gave us the inspiration to think, well, these use cases are actually useful for a lot of different people, not just designers. And the reality is like actually two thirds of our weekly actives are non designers anyway, right? So what could it mean to build a. Use a tool that's more tailored for them? And so we started pushing on that idea and that was the emergence of Figjam, which we brought into beta the next year in 2021. And since then it's been really cool to see people take it in very different directions. Whether it's some of the things that you would might expect, like, you know, creating brainstorms and user flows and, you know, diagrams and things like that, all the way to running all sorts of meetings that maybe have nothing to do with product design or product development, but it's just more engaging because people can do, really have light, lightweight ways to interact or engage, give a thumbs up, you know, emote, fire, whatever that might look like. So that's been, you know, the evolution of figma mod to figjam.

I will say that of the meetings that I've had with FigJam has been the best meetings because like, the product itself is just so fun to use. I don't know if you guys have fun as a core principle early on or it just, you know, you let the designer just make it fun.

Yeah, well, it was, it was really interesting because it actually was. I distinctly remember this, this was two months before we were launching like figjam to beta, where we had a board meeting where the board was asking us, you know, what is the differentiator for figjam? Because there are a lot of other whiteboard tools out there. And Dylan said, well, it's the fact that it's fun, you know. And then, you know, we actually were like, wait, it's not fun enough. We need to really lean into that. So, you know, Emily and Jenny, who are the product and design leaders on FigJam, actually ran a sprint. They called it FigJam Fun Jam and they came up with all the features like, you know, cursor chat and emoting and all these things that didn't actually exist in February of 2021 2. Two months before we launched. So you're absolutely right. Fun was the principle with the hypothesis that, you know, actually a lot workplace tools are a little bit boring. You know, you kind of get confined in your box, maybe, like, you know, you have to really feel like you put a lot of effort in to express yourself. And we're kind of saying, hey, here's an infinite canvas. They're really lightweight ways to express yourself or, you know, react to things, and you can get as creative as you want and have all the space to do it. So I think that was. It's worked out in terms of, like the hypothesis around bringing fun to the workplace. Actually, it's really helps engagement, which is a really important part of getting things done.

Yeah, I love it. I'm sure it must have been fun to build too. So let's talk about the AI features. Maybe you can briefly explain what they are and maybe how this project even got started in the first place.

Yeah, well, I'll tell you how it started. We had this big AI hackathon earlier this year and somebody built this amazing. We had a lot of actually cool ideas, but a particular one that we all really loved was this, called Jambot. And Jambot is this visual way in which you can take a sticky, for example, plug it into Jambot. And Jambot can do different things. It can rewrite that sticky, it can turn that sticky into an entire story or haiku, and then you can keep iterating on it and say, hey, actually make this more fun, or change the tone or things like that. And it was this really interesting way to visually explore your creativity. Right. And whereas before in something like ChatGPT, you're going back and forth via text, and that's super fun too. But there's something visually that was interesting, or creating forks in the road of like, oh, what if I go this direction or that direction? What could it look like? And tracing back all your steps in this kind of, like, immersive way, that was really fun. So that was kind of what started, and we got a lot of usage out of it, and people were really intrigued and it obviously motivated us to think about, okay, how do we bring these features even more natively? And then the other kind of motivation was just like, one of the thought experiments that I always had was figjam can be intimidating in a weird way because you can arrive at this infinite canvas and like, how do I actually. Like, how do I actually get started? Like, you know, it sounds cool to run a meeting in figjam, but Like, I don't know how to do that. And one concept we kind of pitched internally was there are some people in our company who are really great at making great figjam boards and that are, like, make the meetings really fun and efficient. You know, there's Emily and Jenny on my team Verve, like, product lead and design lead for figjam. And, you know, like, we were always like, it would be so great if we could just deploy Emily and Jenny to every single team out there. You know, like, and it's like, send them in a box and send them to every customer. And so Emily and Jenny in a box kind of became a meme, and we're like, wait, AI actually can help do this? Like, you know, we know some of the best practices, we know what, you know, great meetings look like. What if it could be easy to be like, here's what the kind of meeting I'm trying to run, and, like, you know, like, what could that template look like? Or, you know, how could I get a file to that state so I can go from there? So that was kind of the first thing of, like, you know, there was a distinct, you know, user need of, like, feeling like it's a lot of work to facilitate a meeting. How do we make that faster? And then there are these other obvious workflow things that, you know, people are really always, like, asking us for, which is you run this really big brainstorm, and it's great, it's generative. But as you know, brainstorms can be really messy. It's like, where do you go from there? Once you've gotten your entire team to give you 100 different ideas, right? And usually what you do is you want to kind of vote on the best ideas. But even figuring out, well, what are we even voting on? Is actually a manual process of, like, okay, these ideas seem very similar, and depending on how you cluster them, it can actually affect the vote. And that's what kind of forces me to think about, okay, what could it mean for AI to help sort ideas and stickies into key themes? And then summarization is another thing, which is, it's really great to have this amazing brainstormer session in a figjam file. But it's always, you know, every great meeting ends with, well, what are next steps? And so, you know, like, being able to summarize what happened in a pretty objective manner, like, which gives you some. A really great starting point for identifying, okay, like, here are the things that need to happen, or, here is what happened and be able to even, like, do something like copy paste that into Slack really easily. You know, those are some of keys or problems we saw in figjam, independent of AI. And AI just so happens to be a really effective and efficient way to solve that.

Got it. So building an AI product is a pretty winding path, as you know. Right. So maybe you can share a moment during the process where it was particularly challenging or it was really tough in building this product.

Totally. Yeah. I mean, there are many moments, like, we looked at some of the summaries that were being generated and we're like, wait a minute, there are 10 stickies and the summary has 10 bullets. Like, that's not really helpful. Or when we created these templates look really primitive. It's not that much better than people randomly putting some stuff down. And so I think one of the things that we really had to learn was how do you actually. And this is actually especially true for PMs. How do you actually really get involved in terms of owning quality? Because I think sometimes it's easy to look at AI and think of it as a black box and you say, okay, well, either it figures it out or it doesn't. And I don't have too much control over it because, like, you know, it's. You just don't exactly understand how it works or it's not deterministic. And, you know, I found that, you know, and even myself, like, having been new to building these kinds of features, it's really easy to kind of feel intimidated or feel like it's not really my job to debug what's happening in the box. Right. But actually, like, if you can take a much more proactive role in owning quality and saying, hey, this is not good enough, or actually, this is the kind of output that we want, you know, it becomes an exercise even with your engineers around, like, well, what are the prompts that could be, like, designed to, like, better produce that? Right. And just really understanding what are we feeding and what's coming back and what would we prefer to come back? How can we exercise more creativity around, like, what we're feeding in the first place? So I think there's just kind of this journey of realizing how to get involved in a way that I think a lot of our PMs went through as part of this process, including myself.

Yeah, I think. I'm not sure about your experience, but, like, it kind of requires, like, a lot of manual work. Right. Just, like, improving the prompt and seeing the output and doing it again. I don't know if the team had to yeah, yeah.

I mean, it's kind of like a debugging mindset in a way, but once you kind of realize that is actually the core part of the product job, like sometimes you're like, you think your only job is to define the user experience and like the golden had a happy path, but actually your core job in these cases is actually like, you know, like that trial and error.

Yeah. And did you get the users involved too to look for edge cases or try to have it go off the rails?

For sure, yeah. I mean, like, you know, we have a big culture of, you know, dog fooding internally and you know, so that was kind of the first step, of course, but we released, you know, this to a small group of alpha users too. You know, I know you got access and thank you for, you know, trying it out. But like, yeah, those are the ways in which we start to understand what works, what doesn't or, you know, even like what's good enough enough or whether people are actually encountering real aha moments.

But yeah, okay, and how do you decide when this product was good enough to ship? I think you bundled a bunch of features together for this. Right. So how do you.

General sense? Yeah, yeah. I mean, I think that the biggest thing is there's a portfolio of things, right. Like there's some things that, for example, you run this really big brainstorm and it's great, it's generative, but as you know, brainstones can be really messy. It's like, where do you go from there once you've gotten your entire team to give you 100 different ideas, right? And usually what you do is you want to kind of vote on the best ideas. But even figuring out, well, what are we even voting on? Is actually a manual process of like, okay, these ideas seem very similar and depending on how you cluster them, it can actually affect the vote. And that's what kind of forces me to think about, okay, what could it mean for AI to help sort ideas and stickies into key themes and then summarizations and everything, which is, it's really great to have this amazing brainstormer session in a figjam file. But it's always every great meeting ends with, well, what are next steps? And so being able to summarize what happened in a pretty objective manner, which gives you a really great starting point for identifying, okay, here are the things that need to happen, or here is what happened and be able to even do something like copy paste that into Slack really easily. You know, those are some of keys or problems we saw in figjam, independent of AI. And AI just so happens to be a really effective and efficient way to solve that.

Got it. So, you know, building an AI product is a pretty winding path, as you know. Right. Like, so maybe you can share a moment during the process where it was, like, particularly challenging or, like, it was really tough in building this product.

Totally. Yeah. And I mean, there are many moments, like, we looked at some of the summaries that were being generated, and we're like, wait a minute, like, there are 10 stickies and the summary has 10 bullets. Like, that's not really helpful. Or when we created these templates look really primitive. It's not that much better than people randomly putting some stuff down. And so I think one of the things that we really had to learn was how do you actually. And this is actually especially true for PMs. How do you actually really get involved in terms of owning quality? Because I think sometimes it's easy to look at AI and think of it as a black box. Right. And you say, okay, well, either it figures it out or it doesn't. And I don't have too much control over it because, like, you know, it's. You just don't exactly understand how it works. It's not deterministic. And, you know, I found that, you know, and even myself, like, having been new to building these kinds of features, it's really easy to kind of feel intimidated or feel like it's not really my job to debug what's happening in the box. Right. But actually, like, if you can take a much more proactive role in owning quality and saying, hey, this is not good enough. We're actually. This is the kind of output that we want. You know, it becomes an exercise even with your engineers around, like, well, what are the prompts that could be, like, designed to, like, better produce that? Right. And just really understanding what are we feeding and what's coming back and what would we prefer to come back? How can we exercise more creativity around, like, what we're feeding in the first place? So I think there's just kind of this journey of realizing how to get involved in a way that I think a lot of our PMs went through as part of this process, including myself.

Yeah. I think. I'm not sure about your experience, but, like, it kind of requires, like, a lot of manual work. Right. Just, like, improving the prompt and seeing the output and doing it again. I don't know if the team had to.

Yeah, yeah. I mean, it's kind of like a debugging mindset in a way. But you know, once you kind of realize that is actually the core part of the product job, like sometimes you're like, you think your only job is to define the user experience and like the golden, the happy path. But actually your core job in these cases is actually like, you know, like that trial and error.

Yeah. And did you get the users involved too, to look for edge cases or try to have it go off the rails?

For sure, yeah. I mean, like, you know, we have a big culture of, you know, dog fooding internally and you know, so that was kind of the first step, of course, but we released, you know, this to a small group of alpha users too. You know, I know you got access and thank you for, you know, trying it out. But like, you know, those are the ways in which we start to understand what works, what doesn't, or, you know, even like what's good enough enough or whether people are actually encountering real aha moments.

But yeah, okay, and how do you decide when this product was good enough to ship? I think you bundled a bunch of features together for this. Right. So how do you.

General sense? Yeah, yeah. I mean, I think that the biggest thing is there's a portfolio of things. Right. Like there's some things that, for example, in the cases around generating a template, it's great from going from 0 to 1, but at the same time, maybe some people don't need that. But what people really need is something that's more deeply integrated to their today's workflows, like running a brainstorm and summarizing that. So I think there's kind of this portfolio approach of, okay, what kinds of users can we hit and can we have a package that kind of can feel useful to anyone regardless of how they use it? So I think there was a little bit of bundling we wanted to do to have a bit of a moment versus just ship a bunch of things in succession. But, but that was really kind of more of a bundling strategy, I guess. But when it comes to feeling good about its readiness, there's just a lot of it was based on what our customers are telling us. And even two weeks before, there's a lot of feedback around quality that we were unsure of. And so, yeah, it was a little bit of a roller coaster ride. And we also acknowledged that it's actually just a work in progress. We're going to learn a lot and people are going to tell us at scale when it breaks down for them or we're going to be able to tell what kind of queries that people are trying. And as a result of that, hopefully we get better too. And so that's the hope with a feature like this.

So what would you say is success from the launch? Is it just seeing people come back to use it again or just see all the new use cases people might have?

Yeah, I mean, I think a few things. Right. One is more than just people continuing to use it again. I mean, that's like a secondary metric that, you know, show, like feature retention is a good proxy. But at the end of the day it's like, are they getting their job done? Right? And there are people who were never able to facilitate a meeting that maybe, you know, this lowered the activation energy for doing so. And that's really kind of wonderful from we activated a user, helped them understand what Fig Jam was about or, you know, like maybe that meeting where it was, you know, unclear what next steps were or, you know, like the likelihood that people run another brainstorm again because they found it an effective place. Like, those are the ultimate use cases we care about versus just like using, using AI. I really think that AI helps with product understanding and help people get things done effectively or more efficiently. But those are the things that I look for. The first category of what we want to do, but the second category is also learning, which is that our community has a really wild imagination. They come up with all sorts of cool stuff. And so when we start looking at what are people trying to do at Big Jam, that's really interesting for us too, to see, okay, what are they trying to generate? Maybe we don't do a good enough job on some of those use cases. Does that inspire us to think a little bit differently about the use cases we want to enable? And figjam at the end of the day is an infinite canvas that people can do anything on, in a sense. And so I think having more visibility into what people are trying is going to give us a lot of good inspiration for directions.

Awesome.

Yeah.

I'll tell you my personal experience. So at my current company, usually the designers use figjam to run like a brainstorming session. And yeah, usually people love it, but then people like me get intimidated, like, wow, this designer came with such a good template. It's like, I'll never be able to use figjam the same way. So I think there's a lot of upside here.

This feature, that's absolutely our hope. Yeah.

Awesome. So, something else I would have chat about is I think figma's mission is to make design accessible to everyone. Right? So how do you cater to most power users, but keep the product simple enough for people like me who don't really know what they're doing? How do you balance these two?

I mean, it's definitely a theme that we've been grappling with a lot in ostack. We have some initiatives going on that are addressing this right now. But I think that a couple things, right, you want to make sure that you balance streamlining, but also not make it feel like it's dumbed down. That's kind of the trade off I think that you're alluding to, right? And there are a lot of industry ways to solve this, right? Like, is it progressive disclosure? Are there things like shortcuts? Are there other ways in which we can kind of expose the right functionality at the right time? And I think all those are viable solutions and all those are things we think about. At least on Figma, there's a particular dimension that we look at, which is one of the ways in which we add a lot of power today is via design systems. We add a lot of powerful design systems features like we did in config our user conference earlier this year, like variables, right? And they make design systems more powerful because it gives design systems managers way more control. But also it kind of gives people these components that they're much easier to use. And our hope is, as we add more power, we differentiate between the people creating these design systems and people consuming these design systems. Because hopefully in a world where in the future someone, whether a designer or non designer, can arrive at their team's workspace and open up a Sigma file and it just feels like you're just dragging in a bunch of Lego blocks to assemble a screen and get you really quickly started. It should feel as simple as that. But on the other hand, really underneath the hood, in order to make those simple feeling Lego blocks, a lot of things needed to happen, a lot of power needed to be created. So we expose a lot of functionality and power to design systems creators. So kind of in our heads we always say, hey, with design systems creation, like maybe it's okay to add a little bit of complexity because they really want that control. But for consumption, it absolutely should feel like really simple. I'm not saying we're fully there yet, but that's kind of spiritually how we think about it. For Figma design, and then for a figjam, we feel like it should always feel simple. And that's why there are a lot of product decisions that we've made. That intentionally are different from Figma design in terms of, okay, well, maybe instead of giving infinite choice, we make more sensible defaults or things of this nature. Or maybe we make this controls feel less like there's some powerful stuff on the properties panel on the right, and rather something in line that feels right there and contextual and within reach or feels like you're directly manipulating something. So those are some things that are really important for figm as well.

What do you think is like the kind of like the magic moment for like a non designer using figma? You know, it's like, is it making something they're proud of themselves or. Yeah, yeah.

I mean, I think, you know, oftentimes a non designer is actually, you know, being brought into a file with a designer, right? And. Or opening up a file a designer made. Like that's, you know, what our data shows, right. And you know, they're able to poke around and I mean, certainly the magic moment is being able to just see things being updated in real time, right? But there is also ways in which, for example, like you open up a prototype or you open up a design file and your designer's showing you around and you want to observe them, right? You want to follow them and then you kind of like they take over your screen and like they walk you through the file. I think that's a pretty magical moment because it kind of really feels collaborative. Like, it really feels like you're looking over the shoulder and they're. Or maybe they're walking you through the hallway to show you some stuff. Like, like, so those are pretty, pretty magical ways. I think, like, another one that I really love is, you know, oftentimes as a collaborator, maybe your first view is something like a prototype, right? Maybe it's a deck, maybe it's a prototype and you're able to add some comments to it and the designer's able to see it in their view as an editor, right. And so they're just in the designs and they can respond to it immediately, even though they're not necessarily in the prototype with you. And so those are some potential magic moments as well. But our hope is, as you said, there's nothing more fun than creating and making and I think our aspirations to make it so that anyone feels comfortable doing that. And there's a lot of work that still needs to be done to make it feel that way. But that's definitely kind of like when you think of our mission, what we're excited about.

My designer jokes that I'm like, IC 0.1 designer or something learning from her. But like, yeah, like, you know, as someone who cares a lot about crafts, I wish I had better design skills. So, like, for people like me who want to learn, like, what's the best way to learn this stuff or sorry, stuff?

Yeah, well, I mean, I feel like a couple things. One is just like, I find it. I'm the kind of, you know, there are a lot of people who are really good at learning and like, can go watch a YouTube video, can go read up on things and become better. I'm unfortunately not one of those people. Like, I either I need a project and I need to have someone to emulate or copy, you know, and I think, like, those are things I always look for. So, for example, it sounds like you have a wonderful designer who like, you know, you're learning from. And like, one of the things I used to always do is just like alt drag, which is like, you know, duplicating like something that they've already made and playing around with something and seeing if like, you know, I always have ideas for like, different things to try out. It's always intimidating, maybe to start from scratch, but like, if it's just like a tiny, tiny fork off of something that exists, that's like the beginnings of something, right? And then you start wrestling with it and you're like, oh, okay, this is how you did it. Or like, wait, I don't understand how to change this. Then it forces you to learn about like, oh, this is how a component works, or, oh, this is auto layout. And that's why things flow this way. And so you throw yourself into something that already works and then you break it a little bit and then you know, okay, command Z. Okay, that was messed up, but okay, what was that? And so I think those are ways that I find the best way to learn in a sense, right? And then the great thing about figma is like, there is someone who can be in the file with you and if you have a patient enough friend or coworker, they can show you a few things. So that's the way that I learned the most just by just diving in and trying. And there are a lot of resources, for example, in the community where you can do that. And I think that's why we think there are two things about figma that I think it's really been helpful for helping people learn faster. And the first thing is just the fact that it's kind of like open source. There's a FIGMA style that someone was worked on. You can just you know, we have a community of figma files and dive into it and just like inspect how it was done. Right. Like, there's nothing more powerful than like starting with something rather than something scratch. And then the other thing is like it's in the browser, right? Like, you can. You don't have to go at it alone and someone else can be in there helping you out. So hopefully those two things make it a little bit easier. But you know, it's also our jobs to help you like kind of ramp up as well.

Yeah, I think getting my hands dirty and breaking things is the best way I learned to. So how do you. Let's close with two questions. First, I'd love to hear about how you're using AI in your everyday life, I guess, both at work and also just like at home.

I mean, it's really awesome to have FigJam have AI right now because every single meeting I'm basically in FigJam in some way, right? And sometimes there's this discussion happening or a brainstorm happening. I go often to the side. I'm like, I just want to quickly summarize what's happening. Or we run these meetings sometimes with the PM team where we give a lot of asynchronous. Or it's kind of like a multi threaded meeting, right? Where people give a lot of updates and on like wins that they've had or shoutouts that they want to give or stuff like that. And we spent 10 minutes of heads down time just silently putting stuff and reacting to each other. And as N has gotten bigger in our team with like 30, 40PMs, it's like, I just kind of want to take that and instead of trying to read everything, just extract out key themes really quickly, you know. And so like, those are some ways in which, you know, I've started to use AI even in meetings. Right. Which is really cool. So that's kind of like the first one. AI my, you know, personal life. I think like one of the things that I find in at least the creative process is I think just like nature of the way I build things. Like, I'm actually not like, really, you know, I'm not good at drawing. You know, I'm not good at starting from scratch, but I'm very good at, like, when there's some idea I can like refine from there. And so I actually use it as the first job. For example, I design logos for my softball team and the jerseys on it. And it's like, okay, there's some concept like, what if There is this object placed with this object. Well, AI can generate that and then I can refine from there and turn it into a vector or whatever. And so that's kind of a workflow I've started using. Whereas before maybe it was just like you search around Google Images and hope that a search term kind of gets you some things that you need and you cobble it together.

Oh, interesting. Yeah, that's really interesting. It's really cool that you're still spending time on your craft, it sounds like. Or at least using AI on your craft, even though you're leading teams and doing all this other stuff at work.

Yeah, well, I think one of the first questions we ask every Figmate people at Figma is, how are you a maker? And I think we all have that itch to create things and we find side projects. So in my case, sometimes it's just making decks where you kind of release that kind of. Of desire or energy.

Yeah. I've been interviewing a bunch of folks. I was talking to Scott, and yeah, I would love to just shift this PM culture back to making, as opposed to just all the internal stuff and alignment and stuff like that. Yeah, I guess to wrap up Figma is growing. You've been there for four years. How do you keep this maker culture alive in the company, man, instead of just a bunch of people immediately aligning with each other.

Right. Yeah, Well, I mean, I think a few things. One is we do try to hire people who love making inherently. Right. And so it's definitely true that as you get bigger, there's a lot of other things that you have to do. But people are here because they love to make and get their hands dirty in a sense. And, like, I'm that way, Dylan's that way. And so it's definitely a bias that we have in our culture. I think there are some ways in which you have to create that, like programmatically. So, for example, at figma, we have Maker Weekly and we say, hey, the only thing requirement is you make something. And so sometimes that means people make product ideas, but then other times people are making their craft projects together or they're doing murals in the office together or things like that. And I think those kinds of moments to pull out and be super creative actually get the juices flowing. And a lot of ideas have emerged from that or a lot of creativity comes out of that. So. So those are some things that I think are really important to give space for because it's so easy to just be like, overwhelmed by everything in front of you or the problem that you're solving and forget to take that kind of, like, you know, exercise that part of your brain.

Awesome.