We're sunsetting PodQuest on 2025-07-28. Thank you for your support!
Export Podcast Subscriptions
cover of episode #140 Designers Coding with AI - Inside Intercom’s AI x Design Revolution w/Caitlin Rich

#140 Designers Coding with AI - Inside Intercom’s AI x Design Revolution w/Caitlin Rich

2025/6/25
logo of podcast Honest UX Talks

Honest UX Talks

AI Deep Dive AI Chapters Transcript
People
C
Caitlin Rich
I
Ioana
Topics
Caitlin Rich: 我认为Intercom公司正在积极转型,通过AI技术来提高研发效率和整体生产力。我们鼓励员工在所有工作中首先考虑如何利用AI来简化流程、加速工作并提升质量。这种文化转变意味着我们不再将AI视为一种作弊手段,而是鼓励大家拥抱AI,如果现在不使用AI,反而被认为是落后的。这种转变正在推动员工更开放地进行实验,并分享他们的经验,从而共同进步。

Deep Dive

Chapters
Intercom is undergoing a cultural shift to leverage AI for increased productivity, particularly in R&D. The company encourages AI use, contrasting with previous secretive practices. This change is impacting teams across the organization.
  • Intercom aims for 2x productivity increase through AI in R&D.
  • A cultural shift encourages open AI usage.
  • Previously, AI use was seen as 'cheating'.

Shownotes Transcript

Translations:
中文

A big thing recently is Dara Curran, who's our CTO, just released a blog article on this, but we're thinking about how can we change the company to be like 2x more productive, like the whole, especially in R&D, using AI. So it's really that cultural shift, like before we do anything, we need to ask ourselves, like, how can AI make this easier? How can it make it faster? How can it make it

better. And like that mindset is really showing up across the teams. And I feel like for a long time using AI has been seen as sort of cheating. You kind of do it in secret and you don't really tell anybody about it and hope that no one notices and think that you're just doing all this work yourself. But actually we're kind of being given the opposite. We're actually, if you're not using AI, you're doing it wrong. That cultural shift is really, really starting to change.

Hello, designers, and welcome to a new episode of Honest UX Talks. This is a very special episode because it's part of my limited series, How to AI, a series where I invite guests from juicy tech companies, from intercom to perplexity, to unpack how

the work they're doing in the AI space, how they're designing for AI interactions, how they're hiring designers in the age of AI if they're looking for AI fluency, how are they infusing AI in their team, and many more very exciting topics. Today I'm joined by Caitlin Rich. She's been in the design industry since 2008, which is 17 years now, so she was doing design before it was

cool. And now she's the group design manager for the growth team at Intercom, focusing on ensuring new customers have a great experience, including the adoption of Intercom's AI agent called Finn. I stumbled upon a post on LinkedIn a couple of weeks ago where Emmett, the VP of design at Intercom, shared a screenshot from their Slack channel where designers from the Intercom team were celebrating their first

full requests, they were fixing things in the live product by vibe coding with cursor and it really hit me. Oh my God, the future is happening now. And in that screenshot, many people were thanking Caitlin Rich, who is my guest today, for enabling them to use AI to understand how to use cursor as an interface to promoting real changes, real fixes in their product and how to get accustomed to this new world of vibe coding.

So I've decided to reach out to Caitlin and invite her to unpack what are some of the instruments and frameworks that they're using in their design team and how we can all kind of adopt these new technologies in our own design work day to day. This is going to be a very exciting conversation and we will unpack a lot of topics.

from screening for AI literacy when you're hiring to how their team, the team at Intercom, is using Cursor to do small UI tweaks in the real product. So that's really, really great. Before we jump into it, I want to take a moment to thank our sponsor, Wix Studio. It got me this series, and the series I'm planning here, already started to record, has made me reflect on how can designers start embedding AI in their workflows.

One of the reflection points I landed on is that we have AI in our design tools and I think that's the best entry to this new world of AI possibilities. A very good example is Wix Studio. They have a bunch of AI capabilities

But one that's been around for a while now, so it's probably even gotten better, is the possibility to ship quicker with an AI code assistant. Essentially, they present us with a generative AI model that was augmented by Wix Studio. And within their VS code-based environment, you can get tailored scripts.

It's troubleshoot, retrieve product answers. And this is just one of the ways in which you can expose yourself to this new technology in a product that you get because it's a design product. You can generate AI images on canvas and apply certain styling with the help of AI. You can enhance design.

imagery in one click with the help of AI. And so you can start implementing AI in your design process by going in your classic design tools, Wix Studio as a website creation platform, or you can go to Figma and experiment with Make. With Wix Studio, you also get the

possibility of actually building the product and seeing your changes, your AI experiments live. So I think that's a great opportunity and we tend to run for the juiciest new AI tool out there when in fact the tools we use are presenting us with these opportunities for learning and we should give them more attention. So thank you Weak Studio for bringing smart, meaningful AI in our lives. I'll talk about responsive AI. It's life-saving in the next episode.

But until then, let's go to our episode of today and see how Intercom is doing AI in their team. So, Caitlin, thank you for joining us on Honest UX Talks. I'm honestly very, very excited for this conversation. I feel that even if we don't record it, we don't publish it. This is one of the highlights.

of this couple of months for me because like I mentioned in the intro, I was fascinated by the work you, at least from the outside, I think I'm going to be more fascinated when I learn more about what you're doing from the inside. But it was fascinating to see your work and contribution and culture shaping at Intercom. So before we dive deep into it, I would love to learn more about yourself, your background and your current role at Intercom and then we can dive into the AI stuff.

And thank you for accepting my invitation. My God, of course. Well, thank you so much for having me. I really appreciate it. Yeah. So I guess in terms of me, so obviously Caitlin, I've been in the design industry since 2008. So I guess that is 17 years now, which is kind of crazy. I started my career as a graphic designer. So I used to work in agencies back in Sydney, Australia, which is where I'm from. And I used to be very much a jack of all trades. I used to originally do like

branding. I did a lot of print work. Like I'm talking financial reports, like big long booklets. And then I started working a lot in web as well. So I actually used to spend a lot of time in Flash, which was my thing back in the day.

which definitely feels like a different era now. But yeah, over time, I shifted into product design and I really found my footing in the startup scene, especially when I moved to London, which was almost 10 years ago, which is also a long time. Yeah, since moving to London, I've worked at startups of all shapes and sizes, mostly within fintech. And yeah, there's a lot of fintech here in London, so that's kind of naturally where I gravitated from.

a while and yeah it's been a brilliant place for me to grow my career as both like an IC and now I'm kind of more on the management path which brings me to my role here at Intercom. So I am one of the design managers here and I look after the growth group. That means I spend like a lot of time making sure that

All of our new customers have a great experience when first using the product. That's really thinking about like our trial and making sure they get set up successfully, that they're activating correctly. And more importantly for us now is also the adoption of Finn, which is our AI agent. So how do we make sure that our customers are understanding the power of Finn and getting set up?

So yeah, so that's kind of me. I think like what drew me to Intercom was obviously like, you know, the complexity of the product, the strong culture of ownership here as well. But then I think for me, it was also the opportunity to kind of just help shape how AI is changing the way that we build an experience software. And that's something that I'm really passionate about.

Thank you for sharing your story. And I think something that my audience definitely doesn't know about me, and it just occurred to me right now, listening to you was that for many years, I was a huge fan of intercom product culture. So I'm still following the CEO and I think it's a CEO, CPO. I always like Paul. Paul Adams. Yeah. CPR. Okay. Okay.

So I'm following him on LinkedIn. I think he's sort of a visionary. And I was a fan of Intercom because they were making their design culture public for a long time when not many companies were doing that. So the career ladders at Intercom were something that I kept referencing for many years when people asked me about, how can I grow in my Oregon? Hey, here is a good career ladder.

that will give you insights into the skills you need to hone in order to go to the next level, whatever that was, staff principle or whatnot. So I'm a fan of Intercom. So yay, I'm happy to have an interview with someone from this company that I admired for a long time. But what really made me reach out to you recently, like the push I needed, was seeing screenshots shared on LinkedIn by, I think Emmett is the VP of design. And he shared this screenshot that was taken from your design Slack group.

And it was essentially designers celebrating pushing their first PR. So pushing real code into the product, as far as I could read into that screenshot. And it felt like this

This is such a shift that I'm seeing for the first time with my eyes. I kind of had an intuition that we're going to go in this direction. We feel like it's happening, but I haven't seen it really until that point. And then people were thanking you.

In that screenshot, everybody was thanking you for enabling this and showing them how and teaching them. And oh my God, I have to talk to Caitlin. That was the immediate reaction. And then I reached out to you immediately. And here we are today. So can you share a little bit about what went on there? What was the AI experiments education? How are you involved in this? And why is everybody thanking you?

Absolutely. So yeah, that screenshot actually came out of an initiative that we've been running across Intercom. So our plan is to get every designer here or product designer here set up with a developer environment in Cursor so they can start making like really basic UI changes themselves. And it's been such an amazing experience for many reasons, like not just in terms of like what this means for designers, but also just seeing people's confidence grow in working with AI and code, which is awesome.

And I can't like take all the credit here. So Ellie Short, she's one of our amazing designers here at Intercom. Like she's really been kind of like helping lead this change. So she kind of off her own back started working really closely with the engineers on her team to get her set up

with a developer environment and start looking at Cursor. And she started sharing like what she was learning. And this kind of sparked this whole sort of wave in momentum. So like I partnered really closely with Ellie as well as Alexia, who's an incredible engineer here. And we began organizing just like really small groups of designers to kind of like set them through these training. And that included, okay, let's walk you through getting a developer environment set up. How do we train you to use

cursor and how do we help you make your first like really small code edit and then how do you create a pull request that goes through to an engineer to approve and every step along that way was so painful and

It was such an amazing process. It was so iterative. So what we do is like we have that small group, we kind of go away and think, okay, where were people getting stuck? What was really hard? How do we try and simplify this? And I think that was awesome to have like engineering as a great partner in that to really, really help make sure that we can do that.

But what we're trying to get to is to a point where any new product designer who is joining Intercom, they can just follow a document that helps them get set up with a developer environment super easily, cursors installed on their machine, all the guardrails are in place so they don't start making any major mistakes that cause engineer a lot of headaches.

and then start contributing without needing like an engineer to handhold them through that process. Yeah, it's been like amazing for us to get to this point. And I feel like we've come so far and we still have so far to go, but it's just been amazing. Like designers who've never touched code before, who've never really seen our code base, suddenly able to fix things that have just been bothering them for a really long time. But I think like there's another angle to this as well, like where, you know, making small UI changes is great, but then it's also like what happened

for designers, they just have access to our code base. And like Cursor has become this really great, like intelligent layer between designers and the code base. A great example of that is Katarina, who's like another brilliant designer here at Intercom. Like just last week, she was looking to improve our messenger settings. So the messenger is like for our customer support tool, like the little messenger that pops up in the corner of your product or website. It's actually, there's so much that you can do with this messenger. The settings are so complicated and

And, you know, she could have screenshotted every single page of the settings and she could have noted like all the functionality that she had to consider. But instead she asked Cursor. She used the agent. She asked her questions and it gave her like a list of all the functionality that it supports and all the functionality she had to keep in mind for her to like redesign this from the ground up. So it's also just been interesting like...

how do we just use Cursor for other things we're not even thinking about? You know, originally when we started thinking about working in this way, we're like, what are designers spending a lot of time on? And how can we try and kind of like help them get around these tasks that seem like maybe they're a little bit redundant? And that whole process of like QA, which is such a big thing for designers to spend so long like

looking at what engineers have built, screenshotting it, marking up screenshots, putting them into a GitHub ticket, and then like hoping that an engineer is going to pick up and actually do it. What actually we still had a journey is like building a bulk front end and like building the bulk of the features. But is there a world where designers can come in and then

polish it directly in code rather than having to do that workflow that was quite cumbersome? How do we like tighten that loop? And, you know, we're not trying to turn every designer right now into a full stack developer, but like, how can we just start removing some of that friction and give people more agency as well? I have so many follow-up questions. And I'm just, while you're talking, I'm thinking, how do I prioritize my questions? What's the best order to start asking them? So I'm just going to go ahead and throw them out there. One of the first question is,

to your point about engineers, not replicas in them, of course, and then having them get rid of that kind of small asks where they have to just polish some UI and not necessarily do something sophisticated, which they might better use of their time. What's the sentiment in the engineering team? Like, I know there was this, like historical friction between designers and engineers. Sometimes there are many teams where things are going very smoothly. But I've also seen that kind of vibe where there's some sort of rivalry, right? And what's the

sentiment? How are they experiencing this new superpower that designers have? That's a great question. And it has been generally pretty positive, like absolutely a lot of questions coming from engineers of, you know, how do we not get to a position where designers are just breaking things that they're committing actually like really bad code and

What happens if someone submits a change and a bug occurs from it? Like whose responsibility is it to then go and fix this bug? So there's absolutely a lot of those questions that we're still trying to kind of figure out. And I think it's really for us to just to make sure that designers are obviously still working really closely with the engineers on their team. Another thing that's really helped us here is just to try and like

strengthen those guardrails and this has definitely been a learning experience. How do we stop designers from making changes to parts of the code that they shouldn't be touching? And one of the things that we've tried recently, and this is working relatively well, is having a actual rules document for Cursor that's just aimed at designers. And part of designers in our rules in Cursor, it's like, please always refer to this document and we're always asking it to refer to this document. But it can essentially say what you're trying to do is too complicated.

and you shouldn't be doing it. And that's great. If Curse is turning around and saying, love your idea, but no, that's great. And that says to us like, okay, what I'm trying to do is way too hard. I get it. I'll leave that to my engineers. We'll go back to the old workflow versus actually, I just need to change this padding. It's eight pixels. It should be four pixels. Or there's a typo here. Things that are like pretty...

low risk, then how do we then just try and allow designers to be able to do that? And that's what we're still trying to get that right with engineering. But we also have other engineers who've got similar motivations to us. So at the same time, we've got engineers who are also like, how can we help you at the same time from a more company-wide level as well?

And my second question was around tooling. How did you choose Cursor? Because I know there are a bunch of options on the market. And I'm curious if you did like this kind of very deep research of which one is the safest or which one is the easiest to plug to your environments and so on. And how did you choose Cursor? Why Cursor? And then also why?

What did the designers get? Are they essentially just prompting through text and then they see the code implemented? What's their interface like? Yeah, definitely. Yeah, it's a good question on Cursor. Honestly, the reason why we chose Cursor is because that's where our engineers were, especially when it comes to, for example, getting like the developer environment set up and Cursor running. Like we're really trying to simplify that process, but that's just not for designers. That's also for like new engineers who are joining the company. There's like a shared sort of goal for like

how do we just make this so much easier and how anyone can be able to follow this in the future? So just using Cursor because that's what our engineers were already using just made sense for us rather than us trying to implement some other tool that engineers weren't as familiar with, where we just, it'd be much harder for us to get that support. But in saying that, like, well,

If another option comes along, and I'm absolutely sure looking into the future, many options will. We're very open to that. And these things that we're actually talking to our engineers at the moment, they're using Claude Code more and more and actually should be using Claude Code instead of Cursor. Cursor makes sense for now, but that'll change in the future. And I definitely see on the horizon so many products arising that we're already talking to and starting to look at.

that, you know, how do we make that interface better? And I can go into the interface with cursor, but you know, there will be an interface in the future. That's more like point and click on something you want to change and you prompt it. And it does everything in the background, such as creating the branch with the change, being able to do that for you, actually create the pull request and all the information that goes through to an engineer. Like that's our

designers are having to kind of get used to the engineer's workflow but how does that change and I'm sure that's going to change like a lot in the future absolutely because yeah the interface right now you know we're trying to kind of find what the best practices are and as I said like that rules document that does a lot for us already like it kind of creates the branch and it creates the branch and the name that the engineers sort of expect but then from our side what we found really helpful is you know first off giving cursor the url of the page that you're trying to make

the change on, giving it a screenshot and then just providing a lot of detail about what you want to do. So this is very much, and I kind of find this many AI tools, like natural language, don't give it like three words and expect it to be coming back to you with what you wanted. Like

go through a whole bunch of descriptions, like almost just ramble into it and give it all the information. Like don't give it too much information because that's a lot to work with. But at the same time, like you can then kind of see like what code changes are going to be made. You then have to then obviously preview it in your local environment and then actually like build a preview as well for you to be able to actually see what it looks like in production before pushing the pull request. And that's quite lengthy.

but even still, it's been like an amazing process to even just understand how our code is structured and where it's making these code changes and what else it relates to. So yeah, it's been super interesting that way. It really feels like I'm talking to someone from the future. Yeah.

Five years ahead of everything I've been exploring recently. How did you get here? Do you have an AI manifesto or is there like the AI KPI in your team? Is it incentivized or...

Was it, let's say, leadership AI program? How did you become so curious, interested and really carving the time for that? Because I think most companies and any designer out there is interested in new technologies and infusing their workflows, their thinking with these new instruments. But there's also the problem of they don't have time, they don't have the

autonomy or agency. They're not encouraged. They're not supported. And so how come Intercom has this and how does it look like? Did someone just say you are in charge of the AI revolution inside of Intercom or how did it work from just planning, if you want?

Yeah, absolutely. So I guess there's two sides to it. There's like how I became really interested in this space and then also in terms of the sort of culture at Intercom as well and how that's changing. So I guess for me personally, I think when the penny really dropped for me over the Christmas period, actually, I got hold of Windsurf and I had a lot of time over that break and I just started, which became the term vibe coding. I mean, back, this is not even six months ago, but I started building iOS apps and I

just found it just so liberating. I feel like in the past, you know, I've worked for so many startups and it's always like, what's the growth strategy? How are you going to monetize this? What's the market size for this product you're trying to build? Where I was just like, I just really would like this app and I'm just going to build it and see what happens. And for example, I got a tarot card deck for Christmas and I was like, ah, I

going to build an app around this? Like, how can I use AI so you can like pick which spread you would like, and then you can ask it a question, and then you take a photo of the spread, and then it uses AI to then send your question in the photo to come back with a reading. And like, that was just a really fun app to build. And I'm still just tinkering away with that in the back

and started building like you know I built an Apple Watch app where it's like a to-do list slash Pomodoro where you can just like pick any task view at random and like it's just a good way for me to like not have to pick up my phone or something like that that's distracting if it's on my watch so

That's when I just became really fascinated with what's possible and started to really try and kind of change that culture a little bit within the teams, like sharing more what we're doing on the side and these side projects. But then we come back to Intercom and kind of zoom out a little bit. This is not just design at the company.

So a big thing recently, so actually Dara Curran, who's our CTO, just released a blog article on this. But we're thinking about how can we change the company to be like 2x more productive, like the whole, especially in R&D, like using AI. So it's really that cultural shift, like before we do anything, we need to ask ourselves, like, how can AI make this easier? How can it make it faster? How can it make it

better. And like that mindset is really showing up across the teams. And I feel like for a long time using AI has been seen as sort of cheating. You kind of do it in secret and you don't really tell anybody about it and hope that no one notices and think that you're just doing all this work yourself. But actually we're kind of being given the opposite. We're actually, if you're not using AI, you're doing it wrong. So...

that cultural shift is really, really starting to change. And yeah, it's definitely seeing like people experimenting more, like sharing out in the open about what's working and helping each other like evolve together. So that's definitely like a big part of what I'm doing here as like one of the managers is working with the ICs, like understanding what's working for them.

What are they trying that's working? What's not working? But the ones that are doing things that are really interesting and really productive, then how do we then scale that? How do we then help all the other designers do this as well? Like, do we actually need licenses to tools? Do we need training involved? Do we just need people to have the curiosity to try as well? So that's been a big part of it too. This is, I think, the dream workplace at the moment for a designer trying to figure out how to AI their work. And yeah,

I've heard from what you said, what stood out was this idea of personal autonomy and agency and just driving this on an individual level within the company support and encouragement and like cultural shift that you were mentioning. But also it's very much.

I wouldn't say responsibility, but it comes down to this individual curiosity and experimentation and also not gatekeeping, but sharing with other people and in a way becoming a champion of how we can transform the way we work together. And I think this is the key here, right? How do you give people that space to feel enthusiastic about trying this out and talking about it and creating that conversation and let's say vibe designing?

the AI workflow in a company. The last follow-up question would have been, what are the ways in which you decide what tools to use for AI, what parts of the design process to automate or infuse with AI and so on? But I've heard, and maybe correct me if I'm wrong, that it's more of a just experimenting, finding out what works, then

keeping those things, seeing if there's a license that's needed. And I really appreciate this flexibility and freedom and the safety of the environment, right? Just yesterday, I was giving a talk at a company. And of course, it's in the financial industry. So you know very well that it's highly regulated. There's a lot of anxiety around data privacy and so on. And they were just

scared of trying out any AI tool. Like there was this safety concern in their mentality that was very hard to overcome. And the company was going at great lengths to try to encourage them that, look, we've made sure that you're safe to try it. Please try it. It's safe. I think that this is another thing that you managed to create for the teams, the safety to play around and just experiment.

What about someone who might not have that in their company? Did you ever reflect on what it's like to be in an environment that's not? There was this study recently by the Design Fund, and they showed that 7% of companies are supporting their teams in

improving their AI literacy, experimenting with AI. So it's a very low percentage. And I think you're among the lucky companies out there doing a great job. How can someone who's not supported by their environment start to experiment with these technologies? Like they don't have access to cursor, but they want to see it in their work.

Did you ever reflect on that? Is there anything that someone who's not in this privileged environment that they can do? - Yeah, that's a really, really good question. And yeah, especially spending so much of my life in FinTech, it's obviously highly regulated. There's a lot of concerns there. So it's a really good question.

I mean, I guess there's like one side of it, which is, you know, all of our careers are changing as designers. And I think it is really an element of needing to skill up and level up. So I definitely recommend if, you know, you can't do it within your company to try side projects to get access to like, you know, I prefer Windsurf or Cursor or whatever, but

There's lovable, there's even Figma make, there's different things where you can use AI for your own things. But then also trying to kind of figure out, I guess, like what your company is comfortable with. There's certain things and obviously when it comes to like sharing customer information and putting that within an LLM, like that's quite scary. And definitely I'd say things such as like the Financial Conduct Authority here in the UK would probably not be so keen on that.

But I'm sure there's still parts of your workflow that are not as sensitive, you know, parts of it, such as like, especially around the like ideation and prototyping. And I think that's one angle that we're also looking at. Like we've talked a lot about changing a typo, like, you know, changing a small bit of

padding. But on the other side of that, it's like, how can AI really help you with ideation? How can it help you come up with new ideas where you're sort of that vibe coding, but like getting to a point where it's like helping you create better interactive prototypes? Like, I think that's things where it's like not as sensitive that hopefully companies would be a little bit more open to that. You know, especially if you have like

Figma and already using that as your tool, like you hopefully soon can get access to Figma make and, you know, that's helping you kind of like work with AI in a different way. That's not, not as sensitive. So I definitely like, I'd say for those companies as well, things are going to change in the future. This is going to become so normalized that even if your company is not okay yet, hopefully their culture will change. But you might also have to help lead that change a little bit by doing things in the back

around like trying things on your own and trying to see what works for your team and then yeah trying that way I'd say if I'm in the happier place where my company wants to start doing AI but they just don't know where to start from your experience what would be like a short time

simple recipe to get started? Like just get a vibe coding tool and that's where you begin? Or what would be your one, two, three tips for getting started with an internal AI program in a company? For example, we had a big design onsite where we got all designers, you know, we've got designers

everywhere. So we flew them all to London for a two day onsite. And one of those days was vibe coding. We gave everybody a whole entire day to say, build whatever you want. It should be intercom related and everything was kind of intercom related, but it was like just an opportunity for people to rip off the bandaid and, you know, they could use whatever they wanted. They could use lovable, they could use cursor. They're a little bit more familiar with actual like working with code and the complexity of that.

It was just great though, to kind of just say like, just go nuts and see what you come up with. And at the end of the day, we shared what people had built and it was amazing. It was like so impressive. And we're not just talking about like customer facing features, but talking about, Hey, writing SQL, like a lot of the time we need to go to data scientists to do this. Like, can we build a tool that writes SQL for us that we can just put into Snowflake and we can pull our own data? Like what would a product like that?

look like through to like, yeah, so many other ideas. So many cool things came out of that from, from so many different angles. So I definitely recommend just like forcing people, which sounds really bad, but like forcing them to rip off the bandaid and just get experimenting because it's

I think it's that curiosity, which really makes that big difference. And, you know, right now, like I was saying with us giving people access to cursor, we're like, oh, the QA process could be better. And yet designers are now using it for all sorts of things that we just didn't even really consider before. So yeah, I think just like making sure people understand that it's not just acceptable, but like required. I think that absolutely helps is going to change that culture. You're a manager. I have two questions left.

and one of them is relating to your role as a manager. And I'm wondering whether in the future or even the present of hiring at Intercom, do you screen for AI fluency? Is this something that's required when you hire new designers, having a level of AI literacy or some experience with AI? Or is it implicit? Do you screen for that when you hire currently? It's a really good question. Not like a

but it's definitely something that candidates I'm interviewing now, it's a key question of mine is just to really understand like where their curiosity is, especially if they're kind of saying like, Oh,

I'm not really interested or yeah, I don't think this whole AI thing is really going to take off. Then for me, that's a big red flag because I just don't think they're going to be a good cultural fit for us. So yeah, definitely not something that we're screening for in terms of like from CVs or portfolios yet, but I can definitely see that changing in the future, especially as that becomes more ingrained in our culture. And that's like happening quite rapidly right now that I think in the future,

yeah, seeing like that appetite, hunger to try and figure out like what this new role is that we're moving into. And I think that's really important, really important. To continue the top on the topic of change with, which was essentially the entire topic of our conversation. But just on a last note, how do you feel that our roles will change? Will we still use Figma? Or is it like a thing of the past already? We're just gonna...

prompt our coding tool to build whatever we envision? How do you see our roles evolving, being at the center of this revolution? Definitely. That's a really, really good question. And I think that's it. It's trying to look at where we are right now and what's possible. So there's two sides to this. One of them is the vibe coding, kind of these throwaway prototypes. The output looks like wireframes a lot of the time, but

It can help us obviously like communicate better interactions and that sort of thing. That's been really hard to do in Figma where you've got, you know, a hundred screens and the spaghetti that comes with prototyping, you know, that's working in a very different way. The other side of that being the, you know, how can we get closer to code and what's that sort of difference between our relationship with engineering and how does that start to blur and become a little bit closer? But I see those kinds of two ends of that spectrum that'll start to get closer over time. So yeah,

like, and this is sort of the things that we're looking at the moment, like how can we potentially train, you know, different products on our design system? And you're not just vibe coding using, you know, Shad CN, which is the kind of default design system component library that all the different vibe coding tools sort of use. How can it use our design system? And actually, how does it get to a point where that

code is strong enough for us to be able to like maybe it's not as simple or maybe it will be this simple in the future it's just a one click and it goes through into our code base it integrates perfectly and AI is helping with like every step along that way for intercom we are like an almost 14 year old company so our code base is very complex and maybe that'll kind of slow us down in some ways but at the same time like we're trying to change everything from an engineering point of view as well to kind of help with that so I don't know I kind of see how like

our roles in the future being almost like, and I say this is kind of the way that I've been working with like these sort of five coding tools is it's almost like working with a junior designer where you give it a brief, it goes away, it comes back, you give it feedback, it goes away, comes back, you give it more feedback, you're kind of more directing it. But there's like a huge amount of creativity in there. And I think designers, we're creative people. We understand like

what feels good and what works really well. And we understand our customer and what they're motivated to try and do. And I think we'll be helping kind of use these different tools to sort of orchestrate the experience that we want. It might not necessarily be, you know, finessing every pixel in Figma anymore, but who knows how Figma is going to change over time as well. So yeah,

Hard to know what Figma is going to be in the future to whether they're going to be taken over by something else. But I think right now, like everything is transforming and shifting so much. So we'll see where we get.

Both excited and scared. When this AI revolution started, I had this sentiment that was just mixed feelings and it hasn't changed. Like we're learning, we're changing. It continues to move and be super dynamic. And I have the exact same feeling I had three years ago or four years ago. Anxiety combined with so much enthusiasm and curiosity and just an appetite for seeing what this all means for us and the world.

So Caitlin, thank you so much for coming here and disambiguating some of the things that were on my mind. And it was really a boot camp into how to bring your teams around this AI curiosity and enthusiasm and just foster that within your design environment. And it's just great to see that companies are doing it and making it less intimidating and

even helping other cross-functional partners understand that this is an opportunity for all of us to kind of shape-shift and try to explore new possibilities and come together and figuring out what can we do with all of this? What's going on, really? So thank you so much, Caitlin. This was really interesting, really, really interesting. And I appreciate your time and accepting my invitation with it.

enthusiasm and if there's anything you want to say as a closing thought yeah absolutely I mean just like thank you so much this has been like a great conversation and yeah I really appreciate the chat yeah I guess it's like the future I'd say is for the curious so please like get out there play with all the wild and wonderful tools that we now have available to us and yeah I think it's

it's a really amazing time to be a creative. So yeah, let's see AI is very much something help with that creativity rather than being fearful of it. Thank you so much. I think this is the perfect mantra to close our conversation with. And thanks again. If anyone wants to submit topics for future episodes, feel free to reach out to me directly on LinkedIn or Instagram or anywhere. We want to make sure that we're talking about the things that feel and

scary or interesting so submit topics questions your concerns and see you in the next episodes bye