This is brought to you by Forthought. Building AI agents for every customer moment.
Hey, welcome. Thank you. How you doing? Good. Thanks for coming. Congrats on a great day. Hi all. Hey. Okay. Our last session of the day.
You what? You've like doubled the number of products you had at Config. Can you just, I don't want, you know, we'll talk about them, but like quickly the four you had and the four you added.
Sure. I can give the commercial. So that's why I'm saying short, not a commercial, but just so people know what's going on. Abbreviated version. So we had Figma Design. We launched that in closed beta 2015, GA release 2016. Hopefully some of you are familiar with it, or your teams are. And then from there, we did FigJam. So basically we saw this behavior of people treating Figma Design like a space. Pulled that out of Figma Design, made a dedicated product for it. And then we realized that a lot of the people that were in Figma Design are developers. So we've got 30% of our sort of active users in Figma Design that are identifying as a developer role. Looked at the NPS for those folks, not as happy as the designers. So then it was okay, how do we make a better product for them? Introduce Dev Mode, a dedicated place for all their workflows. So that way you don't have to like draw giant arrows to your frame in Figma. And instead you're able to have a very clear path for handoff, integrations with other tools we use, etc. We launched Figma Slides when we realized that 5% of the files created in Figma Design were slides. Same story, pulling it out. And then this year at Config we did four launches. So, Figma Draw, better vector illustration tools. We did Figma Sites, a way to go directly from Figma Design and have a publish flow to publish on the web. And we also did Figma Buzz. Figma Buzz is basically a production workflow for assets. So if your brand designer has a template, how do you as a marketer or anyone else in the org take that template and then, you know, you could bulk create for example with a spreadsheet and instantly you've got 80, 100, whatever number of assets you need. That is something that's been really well adopted and it's been awesome to see that workflow in enterprise especially. And then finally Figma Make where you're able to go from prompt to app and you can also copy and paste your Figma design in and I think that with Figma Make we're still at the very beginning but this is one of the top priorities for the company and something that we're just incredibly excited about.
And you get a bonus too because after Config we launched our Dev Mode MCP server.
Oh yes, nice. And we will talk about that because it fits into the agent story.
The, I mean, we titled the panel 'Escaping AI's MS-DOS Era' and in some ways Figma Make and sort of these no-code text-based tools are the MS-DOS era, right? Or like how do you think about, you know, you have your core product is so intentional and then Make and all the sort of coding products are by their very nature right now fairly like chaotic and text-based. Like why go into this space and like how do you think about it when clearly sort of like a work in progress in terms of delivering like polished work.
Yeah. I mean, so I guess zooming out, Figma's platform is trying to get you from idea to product and we have plenty of holes in there. Those are opportunities for us to partner, to build, buy, whatever. But I think as we look at the opportunity for us, there obviously is so much you can do with AI right now in terms of putting a prompt in and then going from that all the way to a working app or working prototype. Working prototype is I think, you know, really the realm that most of the results are in right now. If you ship them to, you know, live open web, you might find that there's a lot of security issues with that so, you know, be warned. And at the same time, I think that if you look at the way that designers work, they're very free-form in nature, they're visual thinkers. We have a way to edit code and make people use it, but also I think people want more in terms of that free-form visual side. And that's just the start when we think about what it will take for our users to take AI and integrate it into the applications, the software, the experiences that they're building.
I just have a hard time believing that a prompt is where we stop, right? If you look at just the evolution, the arc of any interface and any software ever. You know, it's not like you get rid of the text-based part. We saw the terminal. In fact, I love the terminal. And I've invested in a terminal company, Warp. You know, they're awesome. But at the same time you layer on top and there's a reason GUIs exist. Humans are fundamentally very visual creatures. You know, look at our culture today. It's a very visual culture. Any media you consume is almost predominantly visual at this point. That's changed a lot in the last 15 years alone. And so I think that's the arc we're already on. And to think of it as like, yeah, the prompt box, that's it. That seems wrong to me. It's clearly we've got the Make product and the Design product and over time you'll figure out how they can sort of communicate and interact with each other more. You already have it so you can use Figma's assets so it doesn't just see an image, it sees, okay, here are the rules and everything like that. But there's a lot sort of you can imagine down the road.
I mean, do you think it's mostly like product managers prototyping right now? Or I had, I think Yuki on your executive team was sort of showing me almost like an Uber Eats sort of competitor but it was clearly, you know, you couldn't build a fully functional Uber Eats. It's still here. We're not competing with Uber Eats to be clear but, you know, they're spinning up and no, but yeah, I mean do you think product managers are the main users or who do you think really is getting value out of something like Figma Make today?
Oh, we see it across the org truly, but designers, product managers, I think when we look at our user base, we're really focused on people that are part of that product development cycle. And, you know, right now it's, you know, you ideate, you visualize solutions, you build them, you ship them, but I think after that it's measurement as well. There's an opportunity to create a loop around that and different phases, different surfaces, they'll split apart, they'll merge together just like the personas will. Even before AI was a big part of the conversation, I think that many of us saw that, you know, product managers, designers, engineers, researchers, their roles were blending more and more and the things that they were doing, you know, everyone was becoming a bit more generalist anyway. That was already the arc we were on before everyone started talking about AI non-stop. And now I think with AI as a part of the conversation, there's even more feeling of I need to be a generalist. And so how do we empower people to go do that from whatever context they're coming from.
And do you think your design community has embraced that? Like do you think, I don't know, you know, my wife is a documentary filmmaker, I go to some of their parties, that is not an enthusiastic about AI crowd. How is your community, the design community, feeling about AI and how much do you think they want to be coders and sort of say, 'Oh, I can be everything now.' versus are sort of hesitant like, 'I like design.'
Well, I think that the maybe bigger point that is really important to internalize is that, you know, you've had, I think I heard in the last session, you know, what is the differentiator? Is it, you know, the liquidity in your marketplace for example? Liquidity is very sticky, it's an important differentiator for a marketplace. But for most software, I truly believe that design craft, taste, point of view in your product, in your brand, that is the differentiator. And so the more that people realize that, the more they want to lean into it. And if you're anyone in the org, you're probably going to need to find a way to contribute to the design process if design is how you win or lose. If you're a designer, you should be welcoming that. Not because, you know, you shouldn't be coming from a mindset of like, oh gosh, you know, other folks are touching pixels, my job is at risk. I mean, I know that folks sometimes feel insecure and I understand that and empathize with that, but design is a lot more than just touching and creating and drawing.
In some ways you're saying, haven't you heard the word taste is like the buzzword of the moment in AI. So if you're a designer you're the one appointed to have taste. So in some ways and also to guide the process, I mean you're doing this wide exploration. You're considering all these different paths that are possible and then, you know, you then have to converge and you have to lead your team through that. And I think that you're going to see a lot more designers in leadership roles, a lot more designers in CEO roles. And I really think it's how people are going to win or lose. So design needs to step up too to that leadership position.
Do you worry about copying? Like, I mean, you know, Figma is so much about creating but I do when I play with these create, you know, no-code tools often what I do is I copy and paste a website I like and I upload it in there and I say make a website sort of like that. You know, do you, yeah, I don't know, what do you, stop that? Like how do you think about sort of rote copying in your tools?
Yeah, if you're using frontier models obviously that's something that you have less control over. And so perhaps worry is not like the right thing to have. It's already happening. You can't do much about it. As you do your own post-train, your own model training, that is something that you want to get away from and you definitely don't want regurgitation, right? At the same time, like as a user, no offense to you, but if that's what you're doing with like the amazing tools we have now, perhaps challenge yourself to think of it bigger.
Yeah. Yeah. I'm not the most skilled. Certainly.
Yeah. You could build your skill, right? We all can. So, yeah, I think that, you know, I actually when people used to ask me how do I learn Figma? My first response was often like take a screenshot of some app you like and then just figure out how to use the tool to draw it. And that's not to say go make that app and like, you know, launch it, but rather figure out how to go use the tool with something that you care about. And I think that there's opportunity for people to learn for sure and any way to get people in and learning is good.
But yeah, the entire reason for starting Figma is to get more ideas out in the world, more unique ideas. That is sort of what I'm getting at though. Like it feels like the old school Figma, if I were to copy, you're sort of like you're learning the craft. And so then by the end of it, you know the craft and hopefully you decide to take your own path. Whereas like the no-code, it does feel like am I learning the craft when I do it?
I think, again, my mental model is more of we're creating prototypes with something like Make right. And for me the most beneficial moments have been when I've been able to prototype a lot of different directions very quickly, throw them away and then go build what I want to actually go in the direction of. Right. And so I think it's actually a way to kind of explore the possibility set, this idea maze. Even if you get to the most obvious things that are cliche, okay, it's fine. You know, it's cliche now. Think outside that box. Right? I do the same thing with LLMs sometimes. Like I was writing a wedding speech. I didn't use the LLM to write the wedding speech. I said, tell me the 10 most cliche ways to start a wedding speech so I can avoid them.
Yeah. Yeah. I've heard some LLM wedding speeches. Oh, they're terrible. Worry. But it is a good way to like, oh, what are the key beats or like what's your...
No, I wasn't trying to structure with it. I was trying to get away from the cliche, right?
The talk about like sort of, I'm going to butcher this. The MCP server. I mean, it feels like so you're making it possible for basically external organizations through open-source to use the context that Figma has on my design sensibility and sort of externalize that potentially even to agents. So, walk me through the thought there and also like sort of the willingness to open up what seems very valuable and having sort of context on your customer.
Well, so first of all, I mean, we've had Dev Mode as a product for a long time. And the entire goal of Dev Mode is help developers translate designs into code. We've had a REST API since, coming up on a decade now, right? So you're not hauling everything off. Intention is to be more open that way for a long time. MCP is, you know, many people would call it an API. You can call it what you like but to me it's just another evolution in that so that people that are working with MCP supported environments whether it's your IDE with something more agentic, you know, whatever, a developer can then pull in the context they need in order to go build out a design. And what I've been super inspired by is this is really so much just the start of what we're doing for MCP. I was kind of nervous to launch it honestly because it's early and despite that and almost no promotion. I think we did like a blog post and a few social posts, maybe one live stream this morning that I missed when I was asleep after that. But yeah, I think that the just sort of natural buzz about it and how much people are engaged in this workflow and trying to figure out how to fit it in because they really want to figure out how to use these latest tools to speed up development for them has been inspiring and exciting.
And also another great surprise that I'm personally very excited about is a lot of people want to write back to Figma via MCP too. And so that's something that, you know, we're looking into both the fundamental need of how do we actually enable that need through Figma Design but also how do you make it so that with MCP you can write back as in I've learned something about my design somewhere else that I want to sort of input into Figma. Sure. Or maybe you are using a model somewhere else or you have other context then you want to be able to translate it back into design.
