How Product Designers Use AI to Build Apps: Petar Ceklic

I'm looking at two designers, one that uses it, one that doesn't. And I get a lot of flack for saying this, but I'd find the one that doesn't. Because the one that does is way more productive.

Welcome to Beyond the Brief. Today, I'm joined by Petar, a prolific product designer who's been working in digital since the early days of the web. He's using AI every day to sharpen his research, review wireframes, explore product flows and even build and ship his own app to the App Store in a single weekend. We get into what AI means for designers, why Figma isn't dead and how designers can move beyond static screens into becoming builders. So let's get into it. Alright. Hello, Petar. Nice to have you on the first official episode of whatever this is. I haven't got a name yet. We're seeing how we go. So I'd like to probably start with just doing a bit of an introduction from yourself. So for those watching, who are you? What do you do? And once we've gone through that, we can start to talk about how AI is sort of impacting your workflows and your sort of design career.

Yeah. Look, I'm Petar and I've been doing this sort of in this space since 96. Like, even though I don't look that old, I am actually pushing 50 nearly. So, yeah, because I keep saying I've been in the industry for 20 years and then I was like, hang on, I decided in 96. But 30 years almost sounds too much, you know what I mean? So, like, you know, minor tangent aside. But I remember when I first started doing web design at that time, like people, web was not really a thing. And so why did I choose this? I couldn't tell you exactly why I chose it, but I seemed to be attracted to it. And just been in, you know, did my own thing when I first started, like straight out of school. I learned a lot, but like anything, too immature to actually be planning the future, right? You sort of, you do your work, you sort of finish your work and then it's like, okay, well, I need to find more work. And then you have a lull in your income. And, you know, that's great when you don't have a mortgage and all that. But I ended up a couple years in, ended up getting a real job and working for an agency. And in that agency, created a SaaS product, which it wasn't called SaaS at the time, right? It's just like a product you sold on the internet. It's like a timesheet product that was quite, I guess, you know, everyone needed timesheets at the time. People were doing it by hand or talking systems. And then just developed that and just was one of those lucky opportunities where you got timesheets.com.au. And then ever since then, the SEO was always great. Just, I guess, the domain and the actual age of the domain. So I helped develop that. And I was there for like a long time, over 10 years. And I just watched that sort of, you know, the two owners, you know, from doing websites and all that. I could sort of see that the product is where it was at. Because, as you know, like a project website, you know, you've got two or three months max. And then it ends. And it's started over with a new client. So whereas product, there's always things to do, always new features to create. So I was drawn to that. And that's how I sort of got into what I got into. But then after that, I sort of jumped around a few agencies. And basically, when COVID happened, was made redundant and thought, okay, well, now is the time to go freelance. And like anything, I'm kicking myself why I didn't do it sooner. But so it's been, what, five, six years since then, as busy as ever, sort of really focusing on SaaS apps and startups and all that sort of stuff. But, yeah, that's a quick intro of how it all started.

So I remember in the early days when responsive websites started to come out. And that's sort of when I really started to move towards digital design. But, yeah.

I guess fall into it in a way. Back then, you know, you weren't just a designer. You had to build. So early days... So early days, it was like literally tables and, you know, like doing it yourself. Like I look at the designers now that are coming into it and there's so much, there's little barrier to entry right now. Back then you sort of had to like, like I'm not a natural coder, but you had to do it. Like you would do the design and you had to execute on it. And, oh yeah, home site before that. And yeah, and like, you know, you've got the 12 fonts that come installed in people's computers and that's what you'd use. And, you know, every sort of five years on, there's always a new thing. And like, thank God it's got to a point where, you know, we can choose any font we want and we can, there's no sort of limitations right now except your creativity. But so to answer your question, sort of fell into it in a way. My first website was a scrap metal website and they had a massive catalogue. So it was less design, more just like smashing out tables of product. And I got quite good at HTML. Like, do I use it anymore? Not really, but it's nice to have sometimes. But yeah, that's the kind of intro. And then where I really sort of started finding my mark is when you're working with teams that have developers and you can go, okay, what do I actually like in my role? And it was never any of the coding or any of that. So when I could actually just focus purely on design, it was actually really good. Because I'd often limit, when I was doing the coding myself, I'd often limit the design based on what I could execute on, which is not nice, right? So, you know, make it a little bit simpler. You wouldn't have things off on the side or anything. Whereas when I could handball it to other developers that were doing it, I'm like, great, now I can make their life difficult. Yeah.

So. Yeah. That's really cool. And then, like, the product side of things is just something, as you say, through experience learning that, you know, a website is sort of a two-year product. You know, it gets done. Generally, it just sits there for a couple of years. And then the client might redo it after a couple of years. But, like, the app just always gets worked on, right? So I think, as well, your move to freelance sort of been a little bit of an inspiration for me and for many. So just to see from afar the sort of success you've been having has been great. And, you know, for people watching at home, Petar and I have known each other for some time. And more recently, we'll be chatting a lot about AI and how the technology can be sort of used in the design space. And I think you really get it. You seem to pick it up really, really quickly. And you share with me almost every other week some new idea that you've got. You've built it out really, really quickly, right? So I finally want to talk a little bit about your – now that you're in freelance and your product specifically, how you're using this technology now to sort of improve your own workflows. And what was that kind of journey like? First sort of playing with it and then seeing what it can do and then moving toward, you know, having not built anything, you know, like a production app to then having one. You know, how did that feel and how do you start to think about the technology?

Yeah, like to answer your last question. So I've helped design and develop – I've helped design many apps for clients. So I never thought I'd actually be able to execute on my own, right? Because at the time as a big barrier to entry was the actual coding of it. So I always wanted to get something on the App Store especially, like considering, you know, I've helped many clients do that. So then when I started playing with Claude, I was like, actually, this is actually really easy. And like, I mean, you understand coding way more than I do. So I'm on bypass or permission is just go because initially it's asking all these questions. It's like, I don't know, you know better than I do. Like, you know, it's asking, should I do this? I'm like, look, bypass all, just go. So far, so good. But then again, I don't know what it's doing. It's not like I checked the code. But so, I mean, that's one thing. And we can speak about the app I developed over a weekend. But how I'm using it for my day-to-day work, like it's integrated. Like it's, you know, when Claude went down a couple weeks ago, it just felt, you know, thankfully you can go to another, you can go to OpenAI, right? But if they all went down, it feels like I'm hamstrung, you know? Like it's speeding up what, yeah, it's speeding up everything I do. Not to mention that my written English is pretty bad. So that for one, it helps all my written communication, right? So specifically how it's helping, I think it's just, it just feels like you've got someone in the room to check over your wireframes. Am I missing something? It's just those sort of questions that, you know, I worry that it's making designers or people using it a little bit lazy. Because whilst we could think about those things and we could get to the end point where we're both happy with or we're all happy with, I just feel like it speeds up that whole process, right? You get to somewhere where maybe it would have taken a couple of days. You get there much quicker, maybe in the afternoon and you're sort of like really happy with it. And plus you've got other sort of considerations that you probably wouldn't have thought of. So, you know, for everything in the process, like from research to wireframes to even like, you know, like design patterns and how should I handle this? It just feels like you've got someone on your shoulder going, look, have you thought of this?

Yeah. Okay. Yeah. So, I mean, I know the feeling, right? You start using the technology and it's taken away and like, oh, you mean I have to use my hands? It feels like it slows you down for sure. Now, you're using it as like a sort of companion to chat to. So you talk about research. You talk about what else am I missing? Is that something where you're just chatting with Claude or you're just chatting with GPT or are you giving it additional context somewhere? Are you setting up any sort of structure or project with your research in one place? Like what's kind of a practical workflow that you would use that gets good results?

With Claude, I basically set up a project for every one of my clients, give it as much context before I start chatting about it. So it just knows if there's any briefs or anything like that. And any of the research I've already done, I feed that in. And then every sort of element I'm looking at, if I'm looking at a component base, I'll create a new chat for that. And so I'm keeping all my chats quite clean for each of my clients. So then I can go back and go, look, this is the component. If I need to work on it again, I jump in onto that one. And all the while, it's obviously got the full context of the project. So I guess as a practical, you know, like every designer is different. Some people name their layers, some don't. I find it's good to be neat with, you know, setting up a structure within, like any tool can work, but it works better when it's structured. Especially for this sort of stuff.

I think you're right. Yeah. And it's good that you're sort of setting it up as a project, right? I mean, you're putting in everything. You're doing the manual groundwork first. You're meeting with the client. You're understanding their problem and their need. And you're the one collating all of that and then feeding it into the agent as, I guess, a co-working companion, right? So it needs to have everything there in order to be effective. So that's kind of interesting. Yeah. So you're using it sort of just to guide you through and you're doing all the groundwork there in Figma. So I guess a blunt question. Do you think Figma is dead? Do you think design programs are sort of going to go away? Or do you feel that, you know, this is just the next evolution in terms of how we do our work?

Look, if you listen to design Twitter, yeah, Figma has died. Like, you know, it's died three times in the last six months. I would argue that a lot of designers are using it less, right? But I don't know. Like, I just feel like there's always going to be a need for hand-designed elements and products, et cetera. Yeah. I've used so many products that are completely generated by AI and they just don't feel right. But yeah. But, yeah, to your point, like, I'm in Figma every day still. It's something that's my main sort of tool. Yeah. But I'm also, like, every meeting I record, I put that into Claude, right? And it's got full context of all the meetings. And so then every time I'm meeting with a client or it's a stand-up or we're looking at a component, it's always got a good understanding of what's going on. So then when I ask it a question, it's not just going, oh, generally speaking, this is what you should use for this component. It's got full three or four meetings before that. And, like, what I most like is just give me a rough couple of wireframes of potential solutions to this. And sometimes they're pretty good. But they're never perfect. And that's that whole element of, like, you need someone in the seat. Will it be perfect in a couple of years probably?

Yeah, that's a good point. I've had people say this to me. Well, surely the model capability would just get good enough eventually that it will do it all anyway. And my argument back to that is, yes, but then the designers just have to get better, right? And you will have the models get more capable, which means that more people will have access to that level of quality. And you're seeing this across, to your point, you're seeing this across the quality of what gets generated now. You have the left-hand side. You have the little pills with the dots in them, the same metallic fonts. You know, you see all the trope of the design capability. And that's on a UI level. And I agree with you. I think where they're really bad is actually good UX. And I'm the same as you. I've tried every tool out there. Every single time a new design tool comes up, I give it a go. And I just cannot get them to think properly about UX. So, yeah, I use it like an iterative, supportive tool. So, you know, whether you're using Figma, pencil, paper, magic path, whatever, you know, if you give them context, they're good at sort of throwing things on the board and showing you ideas. But then you really have to spend time thinking it through and designing it properly. So, yeah, I kind of think that I don't think these things, I don't think design is going away. And I just see to your point, how you're already using it. It's really supportive. You know, it's a great thing to bounce off. And also it's good to hear from you that you're capturing all of your meanings and everything you do as context too. I do the same. And, yeah, I mean, I share that with other people. It's like you really need to just capture it all and store it all and give your agent that kind of context. And then when you start to work like that, everything gets a lot easier.

So, I mean, that's really cool. You know how it is. It's like you're also like a leader in the AI space, right, I'd say. But what I'm saying is when you see other people use it that aren't in our field and are just your mum and dads or your siblings. And I watch my brother do a prompt and it's the bare minimum, right? Like tell me like three or four words. And it's like, oh, that's shit. But any context, like when are you going to Europe? What do you want to know? Like it's just like anything, the more context it has, the better. And that's why I love using Whisperflow and not typing prompts anymore. Because you speak to it more naturally like a human, right? And you generally throw in more words and generally it's quick to do. And like my prompts have just gone like they're huge now. And the output is awesome. But yeah, so Whisperflow for just for having more context. Because everyone can type relatively quick, but I can certainly speak it quick.

Yeah, I'm with you. Yeah.

But you should say like or even the little terms, right? You railway, like all the little tech terms. It's got that inbuilt. Like you say it once and you correct it. It's like, okay, I'll remember that for next time.

Yeah, it's a great tool. And it's great if you're a freelancer and not working in an office environment. So if you're working with others and everyone's talking, it's yeah. I haven't quite worked out how to do that yet. But yeah, I agree with you. It's easier to speak than it is to type. And it just kind of shows that natural evolution of this being like a co-worker, right? You know, when you're using, you know, Hermes or OpenClaw or whatever it is, whatever your kind of agentic system is, the better you set it up, the more it can do. And you become more confident in just sort of talking with it and letting it do its thing. So, okay, great. Cool. So the other thing I really wanted to do was the way you went from, you go from an idea to actually producing something. So I'm really keen to see like dive into a project and sort of share the process from like having an idea to actually getting something out. Because I think a lot of designers probably get stuck in that mode of like, I'm a designer. I just designed the theme in Figma or whatever the graphics program is. How do I use this tech now to actually break out of that and design systems and flows and the entire product itself? Like I have the designers need to be builders now. They don't need to be developers, but they need to be builders. Because if you build things, you see the edge cases. You see all the little things that go wrong that you oftentimes forget in Figma. You know, you might design a few happy flows and maybe one or two error states. But then when you actually use the product, you know, I miss this, this, needs. And so I think it makes designers better to be built. And I think you're a perfect example because you've got so many great things you've created now. So I'd love to see that if you could share something with them.

And like you mentioned that designers need to be builders. I don't think it's even that. I think it's like it's so much more rewarding because for 90% of my career, it's always like, it feels like it's half the job. You've done the design, you hand it off. If it's a good developer, it gets an okay execution. I think that's right. Execution. But, you know, you don't have control of that. And I'll be honest, most of the time it's not executed the way you'd want it to be executed. So for me now to be able to go, you know what, I can build these things and I can get it exactly right. It feels so good. It feels so good to do that. So just as an example project, I guess to encapsulate this, I'm always travelling. Seems like every two months we're going somewhere. And so like I've just got a note in my Apple Notes to just remember all my packing stuff, right? There's always a template I use. And then every trip I go on, I always seem to like, there's always one or two products you add or remove. And so I'd have that in my notes. And I thought, well, this is a good project. Seems small enough to create an app out of. And so then I really wanted to do it over a weekend, give myself a limited amount of time because I don't know, I'm one of those people that seems to do better with less time than like this infinite block of time where I can just do whatever I want. So I had the idea of, okay, well, why don't I just digitize this and make it something that is available on the App Store? And I just didn't want to make, I intentionally didn't want to make any money out of it. I just wanted to be a completely free project. So should I share my screen now? Let's probably see how this works. Let me just go out.

Great.

I can't see you, Jarrad, just so I can't see your responses. But so I guess like anything, I just like to start off with a few ideas of colour scheme, you know, whether it's dark or light or have some bit of pop colour. So, you know, like anything, either design Twitter or Dribbble, just get a few screenshots. And then because I use the app on or the notes pad, notes app, I sort of had an idea of what I wanted. So I really wanted to celebrate the idea of like, you know, going on a trip is exciting. And I really wanted to sort of count down the timer to it. Because to me, that's the fun part, right, is every day you wake up, it's like, how many more days? So that was, these are the early ideas for it. You know, you've got your travel passport stuff, your sort of airport arrows, et cetera. So I thought about orange. But anyway, so these are the sort of initial ideas, like whether we have an emoji for the country, et cetera. But as you know, you progress. I'm a messy designer in terms of my boards. So they're always not laid out great. But it's just the way I work. But then I started thinking about it. I was like, oh, this is a bit too strong, like the actual percentage pack. So then I got to, you know, progressing on. It's still got it. It's got some light mode ideas. I'm I sort of was happy with this sort of texture, with the sort of like passporty texture. And the name, I love a misspelling for a name. You know, like short, but it's still misspelled, said the same, et cetera. I wanted to have an overall temperature for the period of time I'm there. So anyway, so I've landed on something like this, using these bars instead of the sort of big, strong, you know, like rounded element. And so then, so keep in mind, this is all over a few hours on the Saturday.

Yeah. This is your manual exploration that you're doing here. Yeah.

Correct. And so then I wanted to be a little bit more streamlined in terms of the actual, this is the money of the app, right? This is where most of it is. You're in here. But I thought in my own notes, there's always stuff to pack, but it's stuff to prepare. So I wanted to differentiate them up top straight away, right? So packing stuff is putting stuff in the bag. Preparing is maybe getting money out, you know, visas, all that sort of stuff. But then I also thought, well, look, if, if you're doing the flow and it's a short holiday, you're only doing carry on. I wanted it to be somewhat smart and detect if you're packing three pairs of shoes, hey, that's going to be a challenge to fit that. So there was some smarts there. So I had this amount of screens and then I thought, okay, well, that's probably all I need. And I went into Claude, started its own sort of discussions, et cetera, and started plan mode and said, look, this is what I'm thinking. This is all the aspects for it. Did a massive plan, which like I hardly read, but it sounded good on paper. And then just said, look, go. And then... Pretty quickly on... How do I do this?

So it's sort of fairly trusting in the architecture of what I was creating for you. And it's kind of a low risk gap, right? Like it's not, this is just something that you're creating for yourself to experiment with. So you can have that.

Yeah, but I never used these tools before. I never used Xcode. I never used any of it. And, you know, I was asking all these questions that I don't know about. Like, and that's the thing I love about AI. If you don't know, you just ask it, right? Why would I use that method over this? So whether it told me the good information, I don't know. But we ended up with something like this where, you know, it's very close to my design, right? There's, this part I probably didn't design, but this is all using that new sort of glass element. The weather's pulling for every country. Like there's, you know, mostly on those days it'd be hot and wet. You go in there.

So it turned all those from the design, right? So did you tell it that it was, is to do that? Or did it just look at it and say, I think this is what you mean?

Yeah, it would ask, how do you want to get the weather? Do you want to get, do you want to pay for weather and get it for every, every city, et cetera? And like, to me, no one's going to trust my app's weather. It's just an indicative. So I said, look, give me, download all the weather around the world. Give it indicative ranges, indicative sort of, you know, most days for that period, this. And I feel like that's like, you know, if it is a commercial app, I might have paid for an API to, for the weather to be absolutely current for, for that. So I figure, you know, like you've got to respond accordingly. And to me, that was a adequate response in terms of that weather element. But you can see that because it knows we're going to Tokyo. So it's. It's done a lot of this other stuff where, you know, tourist tax, power adapter. This is all sort of like.

That's all generated.

Generated. It's created.

Yeah. Because of Tokyo.

Yeah. And I guess the onboarding is like, okay, well, straight away, just where are you going? And let's say it's. It's Bali. Dates. So this, I just wanted to be super simple, right? So it's, it's, assuming you're going to fly. Of course. I don't think many people would need packing assistance for not, not flying. Carry on's default. I guess it's a short trip. It's sort of thinking it will. These are all sort of pack, templates that based on the distance and, sort of the length of the trip. But you can turn on or off these other ones. Plus you can create your own. So let's just go with the defaults. And then you may not like wearing hats. You can delete that one. So you can check each one is correct for, for what you do. So you may delete a few. You can see this is a bug here. I've got two toothpastes. So I've got to think about that. And then you, you basically said, look, do you want your preparation tasks as well? There's a question. So you can skip that or you can add it as well. So then you've got the insurance and all that sort of stuff to create. And that's it. I mean, that's the app demo in like a, a 30 second flow, but you know,

How much of that was, like one shot, right? So you, you, had some conversation back and forth about it. You gave it the Figma files. I'm presuming you connected by MCP, or you give screenshots.

Well, it's only three screenshots, right? So whilst, a mega app would, I would do that. Like I originally connected it through, but then it wasn't, I was getting some funny results. I think Figma's MCP not so great at the moment, but hopefully that changes soon. But yeah, then the first iteration of it, like, you know, the background is, was fixed for a particular size phone. Like, and then you change to a smaller, bigger phone and it's not quite working. So then I said, no, no, you know how it is. It's like, the developer tries the easy way first and then you go, look, no, make it, make it more of a canvas and make it always feel, you know, that canvas, et cetera. So,

So there was an iteration where, where like it, it spat something out and you'd, like you said before, you'd never used Xcode. So it guided you through that. You're running, you're now running the simulator and you're actually seeing the builds and testing them, you know, within hours of actually giving it the design, right? Yeah.

And just, yeah, put it on your phone. Like, you know, I didn't know you could do that with Xcode. Like I've never really, like I said, never had the install on there, but to actually run it on your phone, you're like, wow, this is actually quite powerful because like, it's not even in the, in the App Store. But so yeah, those first, first few runs of what it thought was good was not quite there like any, any developers. So then it's like, no, no, the corners need to be X mount, give it a bit of shadow. You know, some of these things like just as small things like the padding, spacing, et cetera. And you know, that's a little bit painful because you're going back and forth, you're sort of describing your, your screenshotting. And I know some of the tools you use are probably better at that where you actually can actually almost do it on canvas. But once I was relatively happy, tested it a few times, I thought, well, okay, well, now's the time to put it on the App Store. And that, that was kind of exciting. And as you've never done that before, right? You've never published that. Not, not, for, where is it?

And, and there's a whole bunch of, systems now that trying to do that for you, right? So the rec, for instance, but you, you did this with Claude, just asking, kind of brute forcing the approach, right? Is that what I'm hearing?

Yeah. Basically, like, what, what do you, what do we need to do to, to make it go on? And I think you sent me one of those gits that sort of, look, checks every bits of code of what, App Store, Apple App Store looks out for and fix it. And that's a great thing about, AI is you can just say, look, this is what to look out for. Pick it up. If, fix it, if it needs to be fixed, fixed or not. And then Sunday Arvo, it was to a point where I was happy with it, submitted it. And then it's the waiting game. Yeah. And from memory, I think, I was totally expecting it to be, sent back of something's not right, but it was on there straight away. And I was like, oh my God, that, that seemed a bit too easy. But then obviously then I got a few friends and family to test it. There is, I'm like, oh shit. And so that's why, version, what are we at? Yeah. You probably know where the version is more than I do actually. But, but, yeah, there's a few versions, a few versions in now.

So, yeah, I guess that's probably the next rough direction is, is because this is just like a learning experience for you, right? You're not too worried about the outcome. Like you want it to be great, but you know, are you, and it is, are you, how did you go about testing this? You know, did you, like you say, you've, you've kind of pushed it up and you're getting feedback from others. You've used it a little bit yourself. And the way that you're using the, the, model is just sort of make it better, fix this, change that. Have you, if you went back and was to do this again, how would you approach that side of it, the testing?

I'll probably test it more, to be honest, but because I had that challenge of let's make it over a weekend. But, but then as soon as it's out in the wild, then you start feeling kind of like this added sense of pressure of like, look, I don't want someone to be using this and then it to, fail for one. Like they're putting, you know, it takes a little while to create your own pack list. You're using it for a trip. You're excited and then it not to work. So then very close after launching, I think I, I did three or four updates in the, in the next sort of week or so. And it's all bug, bug fixes. So yeah. So, to answer your question, yeah, more testing, is what I would do differently. But like I'm, it's also a very simple app, right? Six or seven screens max. But yeah, so I'm sure a app developer would look at it and go, my God, it's so like, bad code wise, but then you can't blame me. You have to blame code, Claude.

No, again, it's, it's not a, it's a simple application, right? So it's fine. You're doing it to learn, which is great. Do you know how it's built it by the way? Do you know if it's used to framework or if it's building it in, in Swift or. Yeah. Yeah. Okay. That's great. Yeah. I guess that's the beauty of it because you're using the technology and it's, it's almost like the stack doesn't matter.

Can you, can, can you tell quickly with just looking at this screen?

I might say it's all a bit small for me, but, that's the case. Yeah. Yeah. There you go. It's a, it's a native app. So that's great. Yeah. And, and that's another thing with this technology is like, if, if you're a designer and the stack doesn't matter that much, you know, the, I'm sure engineers would disagree and, and there are, you know, it does, you, you do need to make technical considerations when you're working on a large scale software, but something like this is perfect just to experiment with, right? Yeah. And I think you did, you've done a great job of just like setting yourself that two day limit to go from absolutely nothing to a published app on the App Store. And I bet if you told yourself that, you know, a year ago that you, you could do that.

I'd be like, yeah, maybe not.

Maybe not. So yeah, I guess like this is a really good use case of sort of using a technology to sort of further yourself as a designer. What was really interesting there for me is, is you still got, you still got your approach. You still got your manual approach of like, you've come up with the idea you're experimenting, exploring and, but you haven't actually designed every flow there. You know, when you, in the board, there's only a few core concepts. It's not like you have every single different use case and every flow matte, with annotations and detail, everything as a handover file. So you're just, it's how you're just dumping screenshots into core and just saying, Hey, this is what I want to build. Obviously with more context than that. And it's sort of working with you to get the job done. Yeah.

And it's almost like a, you know, like if you're a senior designer somewhere, you may hand off the initial hard work, which is initially come up with this is in my opinion for, for most projects is like the harder work, what's kind of look and feel we're going to do. And then you can hand this off to a junior or mid designer and go, yeah, okay, I get it. Now I know the sort of colour scheme and the concepts. I can take this and, and do the other pages. And that's kind of what Claude's done here, right? So it's still got the background. It's still got the sort of look and feel of what I did initially, but I didn't design this view. So, and just like in terms of buy a coffee, I've had a few people buy a coffee, which is, you know, I really just wanted to put it out there for free, but it's funny. People actually like appreciate it. And, you know, you, you, get a few kickbacks with that. You wouldn't retire off the money, but, but it's also, once again, you're pushing back to your own website, a bit of SEO there, as you know, the game, Jarrad.

Yeah. That's, that's, that's really good. I mean, it sort of shows it. I mean, you, you, a point that you raised there that I kind of want to touch on for a bit. So firstly, thank you for sharing that. I really appreciate it. I think you've done a great job. I think it looks great. I've used it. It works great. And I kind of just go to show how you can go from idea to something real now as a designer very, very quickly, like two days. Amazing. And to be approved straight away in the App Store. I mean, that's generally unheard of for me. So, well done. But you did raise something there where you said like, I mean, you would have an art director designer who would create a few core screens and then you'd handball that off to a sort of a junior or mid tier to do that grunt work. What happens to the, what happens to those designers now in this world? Like if you have this standard studio set up where that's how it works, the tiers of designers, right? What, what do they do now? You know, if they no longer can learn from seniors in that way, how do you think that changes the structure of agency or how designers should approach their work in this, this, with this kind of technology?

Like, I don't know about you, but like, I just feel like AI is everywhere. Yeah. But it really isn't yet, especially not in Perth, right? So like there's lots of agencies that aren't even using it. The developers are still hand coding. So to answer your question, I think they've got a little bit of time in Perth, or Australia. But I'd be worried if I was them. And I feel like they're only, oh, I'd really strongly recommend that they get into AI and using it as a tool themselves because I know if I'm a, HR person and I'm looking at two designers, one that uses it, one that doesn't. And I get a lot of flack for saying this, but I'd fire the one that doesn't because the one that does is way more productive. So I just think, you know, everyone needs to sort of somewhat adapt, but I just look at it as a positive, a net positive over a net negative.

Yeah. Yeah. It's interesting, isn't it? Like. The, the detractors, you know, everything is, it is a tool, which, which changes the way you work. Right. Like if, you know, you take Figma out of the picture, it was Photoshop for that. You know, you remember doing responsive layouts in Photoshop with the different PSDs that you'd have, and now you have something that works. It allows you to, to work a lot faster. You know, no one's, you know, yelling about Figma, and saying we must work in Photoshop, you know, that ship's sailed. And I think this is the same. I think this is something that is just an additive to an existing sort of industry. And if you adapt to it and you learn how to use it, yeah, you thrive. And I don't think it's going to replace design. I think the more I use it and the more teams that I work with design actually becomes a lot more valuable. Because it's thinking in, it's thinking in systems and flows, especially in the product space and all the output, you know, the output will get better. No doubt, but yeah, like the best of the best of these generative designs in my view, still pretty bad. And designers like yourselves, you know, can, can far outstrip the quality of what gets created with these models. But I feel like your mid tiers and your juniors, you know, I've made the case before that they should learn the fundamentals as well. Like learn the real fundamentals of design, you know, all the things that make something great, you know, typographic hierarchy in the use of white space, you know, use apps and understand how, how they work before what makes a good flow. You know, if you can get all that right, then it doesn't really matter what the tool is, you know, you, you're going to be a good designer and that it does take years of doing that. So I'm not saying it's going to be easy for juniors. I think that it's really challenging, but you know, I remember when we were younger, you know, we would just make stuff, right? Like you would just, you just jump into Photoshop and you'd start creating things and you just have a good idea, I guess. So it is, I don't have the answer. I don't know the way it's going to go, but you know, I've heard on a podcast before, a good example of this was, you know, radiology and radiologists like five years ago. The AI technology is really great at being able to understand a scan and diagnose disease. And I got to the point where it was better than radiologists at diagnosing disease. And the argument was, well, radiologists are average. The machine can do it. Well, what happened is quite the opposite. The amount of radiologists exploded because the job wasn't the task. The job was saving lives. And now they can do that more, with the technology. So you need more radiologists. So you kind of unblock, you know, you clear the backlog of people waiting to get scammed because it just happened.

Yeah, I've seen that trend lately. I don't know if you've seen some of the big companies that initially may have had a mega reduction because of AI, but now they're rehiring again, right? And they're going, actually, whilst we thought AI was going to save us this much, workforce, we actually need the humans back, right? So now they're rehiring again. They're rehiring people that know how to use AI as well because it's, they're more productive. But like, I forget some of the big companies that are doing it now. I'm thinking about doing a LinkedIn post about it because it's not all doom and gloom, you know, whilst those, those articles and posts do get a lot of hits. It's not like that in reality. I feel like it's, we're still okay. We're still going to have a job. We just need to keep up with the technology, especially our field. Like, you know, there's a few designers I know that like, even like, like auto layouts in Figma, they're like, nah, I don't like it. I'm like, okay. If you keep pushing back on things that changed and guess what? Three, four years down the road, you're like so out of touch.

Yeah. Don't calcify your thinking. Yeah. I mean, I can't not use auto layout. It drives me nuts to not use it. But then also, you know, when, like you say, you're in MCP, you know, the results haven't been great, but, you know, you, you, shift to building things, rather than, you know, creating complex Figma layouts as well. I, I flipped my approach to do that. I just, I just actually go straight to build and then I refine, from there. And once I've got great requirements and Figma is fantastic for concepts. But you know, I, I'm not the kind of guy to create really, really intense design systems. I'm not saying that they're, they're not useful. I think like a really great design system actually means that I can do its job better. So maybe I need a helping hand, but yeah, to have that full, full circle does work be a really strong design system. Now I can use that to build things, a little bit faster and better.

Yeah. I just remember, you know, how it is like to me, designing, creating a design system for a big client that may take months. To me, that's not interesting work. Right. I've said no to many, many jobs that are like, oh, we just need to make a design system. I'm like, yeah, sounds great. Not for me. So, but nowadays it just seems that's a perfect task for AI, right? These are the rough sort of nuts and bolts. You sort of give it a first pass. I'll, I'll look over it and go, look, improve here, improve there. But, to sit there and just have so many man hours of just creating, iterations of forms, et cetera.

Yeah. I, yeah. I mean, at some point the, the economics just stopped working, right? You, you pay someone, you know, 80, a hundred dollars an hour or whatever it is to do that, where you can get an agent to do it for, for dollars. Right. And, and it's pretty good at that. It's, it's actually pretty good at creating components in Figma and naming them accordingly and all that sort of stuff. So I do the same. That's exactly how I use it. Yeah, great. So on, on, the, on the labour side of things, we sort of talked about that a bit. So, beyond design, where do you see, where do you see this going? Like the models get better, capability gets better. You know, we as designers, we create interfaces, right. And experiences. We create a UI for A human to interact with. Does that go away? Do, do we have agent to agent where, you know, we, we're not looking at screens so much anymore. Is our work, do you think going to be designing different types of systems? What do you think is going to be like, I don't know, vinyl or something? It's, it's, always going to be there.

Look, I love future speculation and I love all things technology. That sounds great. The agent to agent aspect. But you know, there's still people that watch free to air TV and we have to, you know, as good designers, we have to accommodate the, for everyone. And so to me, there's always going to be an interface where it's a human interfacing with a machine. Whereas agent to agent, that's easy. They don't need an interface, right? They just need a, a port.

Structure the API really nicely. So, yeah. Yeah. So I guess it's just, it's always going to be there and perhaps, you know, the, the pie gets bigger and, you know, our sort of work grows with that pie. Right. So maybe there's another way that people use services. Like, let's say you want to book a doctor's appointment and you no longer need to use the website. You just use your agent to speak to the doctor's agent and it will do it all. You know, you can say, Hey Mia, book this doctor's appointment and it will check what, what days would you like? And you say, that sounds good. And it'll book it. And then you've not even used that interface. And so that, that as an, as a, as an interface doesn't exist at the moment. But I see that as being like a, a pie, a piece of pie that's going to grow rapidly. But my view is that that pie, the pie just gets bigger. More people are using more technology and, you know, we'll, that interface piece is still really, really important. I don't see it going away. I just see it changing.

Yeah. And to that point, like is chat interface, is that a good interface? Like if you being objective and like thinking about the future, like, you know, that interface has been the same for the, what, since text messages been around. That's a long time, right? Surely we can do better than that. What is it? Yeah.

I mean, you're raising a good point, especially with design. As you said, within that, workflow of creating Part Pack was, oh, then I had to start iterating and change this board. And, and I've made this case before. It's not the interface for designers, the back and forth chat box experience. It's really difficult. And you, you're, it's much easier as a designer to get in there and do it. Yeah. So I don't think we're there yet. I'm seeing some improvements in that where you have these, it generally works really well for where there's a bunch of systems that have been built. But you can lay it on top of, the output and you can make those changes yourself. But for mobile apps, yeah, there's definitely nothing there yet. So I think you're right on that is that it is a painful experience. And I, I can imagine there'd be a lot of designers thinking the same thing as like, oh, this is just not fun. The output's rubbish and I I have have to to sit here for hours talking, you know, change that text colour or whatever. And so I guess my response to that is yes, it is awful. But yes, it will get better. And it feels like we're in this state of flux at the moment with the technology that, you know, it's like the early days of gaming when WASD and mouse control wasn't there yet. And every Wolfenstein clone had like a crazy different control scheme showing my age. And then, you know, it, it coalesced into something that worked, and that became the standard. So, I'm thinking for designers, we'll get there eventually. I'm really hoping that the Figma agent, capability is great. And you can use it as a coworker on the board so you can work naturally and just talk to it to experiment and expand. So I haven't got it yet. I have applied for, for beta access, but we'll see.

Yeah. And, and like, you know, with Claude design, the people that are really impressed are non-designers. Anyone that actually has a design sense, like, no, it's not great. It's not really isn't. I don't know. Like to me, like all the AI generated design stuff looks very, very similar. And unless you're doing some really bespoke kind of extensive prompting to make it look, smash it out of that purple, you know, blurry kind of look. Like it looks very similar.

Yeah, you're right. It, again, it's that point of like, if, if the tooling becomes, if that becomes accessible to everybody, then you see a lot more of it. Because it's quite cost effective to, to produce. Right. It was, it's like, this might be quite an obscure reference, but in Illustrator, there's a default path brush. It's like a little paint brush swatch. And, I've seen that in so many logos all over the world. Because it's the default one, you know, it's just that. And, and no one bothers to like, maybe you should, you know, brush it yourself on a piece of paper and scan it in and create your own custom asset. They just click the button. And I think you're right. I've seen the same thing with claw design. I was, somewhat critical. And yeah, just general AI generation for design in the entire list is if people can just type in a box and get a result and they don't have the decade experience. They're just going to say, that looks great. I think it's an opportunity for design. Right.

You know, core design is awesome. You know, like, and like Adrian said, we're back in 2016 guys. Like it's, to me, this is not impressive at all. Right. Terrible. But some people will, these are the people that say, say that your job is numbered, right? Design is dead, et cetera. And if this is what we're fighting against, I think we're safe.

Yeah. Yeah. I think, I think, I think, I think so too. Capability gets better, but yeah, I think it's got a long way to go. And, and again, we can use it too. So, we can, we can work faster and better too. All right. Well, look, Pete, thanks so much for your time. I really appreciate it. We'll, we'll wrap it up here and, I'll let you know how it all goes. You, you process, is there anything else you would want to let designers know? Junior designers getting in the industry, being scared of AI.

Look, I think it's one of those tools that is probably the easiest tool to use, because once you're confused, you just ask it to explain itself. Like what other tool could you do that with? Like if you told me 10 years ago, we'd be learning tools to do that. No, it's, it's, the easiest tool to use and it's so rewarding. I would think that But as a junior designer, you have to get into it. Otherwise you might be changing careers.

Mate. All right. We'll leave it there. Thanks.

On that note.