The ‘messy middle’ matters: Richard Chapman & Steffi Kieffer on AI in design

Richard here. Recently my old college friend Steffi Kieffer and I sat down to talk about something that’s been on both our minds: how AI is changing creative work. Not in an abstract way, but in the daily reality of how we design, how we think and what stays human in all of this.

Steffi and I met 30 years ago at Goldsmiths College. Today she’s a human/AI collaborator and keynote speaker, while I run my branding and web design studio in West London. We decided to let AI interview us – Claude generated the questions, ElevenLabs read them aloud.

The conversation that followed was honest, sometimes uncomfortable and ultimately optimistic about what designers bring that machines can’t…

Day to day AI

AI: Let’s start with the practical reality. Can you each describe a specific project or situation from the last few months where AI played a significant role in your work? Richard, maybe walk us through how AI showed up in one of your client projects. And Steffi, perhaps share an example from your speaking or collaborative work. What actually happened, and what was the outcome?

Richard: That’s an interesting one. There’s something we’re working on right now which is super interesting. We’re working on design for a real estate development in West London that’s now quite mature – it’s gone from being a building site to being a community. The client wants to move the website from being a salesy development thing to being more about the community and the apartments coming up for rent.

We weren’t given very much content for the project because the site is going to be so different from how it is today. What we ended up doing is finding out all the local resources. So, if you move into this place between Victoria Station and St. James’s Park could we include all the restaurants, cafes, swimming pools, gyms, kids’ centres, maybe their price point and Trustpilot ratings? AI generated all of that for us so we could put in not just placeholder content, but real-world information that the client would never have thought of. We’re now able to offer extra value by saying, ‘Hey, had you thought of this? This whole website can have an interactive map with all these resources but at design stage, way before we even get to content upload which is usually at the end of the process. In that sense, we’re able to use AI as a tool to help us, but also offer this incredible additional set of ideas or resources for the client.

Steffi: That sounds really lovely. When you talk about AI, I do have to ask – which specific AI tool are you referring to?

Richard: In this particular instance, my colleague used ChatGPT for everything.

Steffi: ChatGPT! By the way, I came across a really nice name for ChatGPT: ‘Chatty Pete.’ It sounds like ChatGPT, but it’s sort of… chatty. I’m just going to call it Chatty Pete from now on. I’ve got a client who writes articles and refers to Claude as ‘my French friend.’ The other day we had a conversation involving Chatty Pete, Claude, and Zapier – Zapier is another software you often get when you work on automations. I thought that was quite funny how AI had been given personality!

So I was asked about my experience with AI, and it’s a big question because it shows up everywhere. Maybe a recent interesting one was that I had to give a keynote about new topics that I hadn’t rehearsed yet in front of an audience. I used Chatty Pete while going for a run. I used the voice function and told it we’re going to simulate and roleplay the keynote. While I was running, I was giving the keynote and got feedback. What worked was that after some initial not-so-good prompts where it said it was obviously the most brilliant keynote it had ever heard, which is not the feedback I was looking for, I told it to be brutally honest. Then I actually got very useful feedback. Lots of different ways to use it. Lots of content production.

The technology time shift

AI: You’ve both been creative professionals for three decades. You’ve seen desktop publishing arrive, the web transform everything, social media reshape communication. Where does AI sit in that timeline for you? Does this feel like another tool in the toolkit, or does it feel fundamentally different from those previous shifts?

Richard: My big thought about this is that throughout my career – from when I quit my job and went out on my own at age 27/ 28, to today at 50 – I’ve known massive shifts where technology fundamentally changed and it felt like the rug was completely pulled from under me. I used to be very scared by this, overwhelmed. I felt like I was behind the curve or hadn’t got things quite right, or my business was going to fail. Over the years, I decided not to become afraid of those things anymore and instead to run at them.

This is definitely one of those moments. I’m quite glad I watched AI happen and was a bit unsure about what it was for to begin with, because it has gone on a bit of a journey in the time since ChatGPT launched. Today I can embrace it and use it for what I decide it’s really useful for. I can now be sceptical about the things I think it’s hopeless at and instead really lean into where I think it’s going.

Does it feel different from previous shifts? Yeah, it does. A good example: when QuarkXPress completely failed and we had to jump to Adobe InDesign – so, a different piece of software, requiring a different way of using something you do every single day. At the time that felt scary, but it turned out to be easy in the sense that you just had to learn something new and that it’s a related tool. If you assume AI is going to stay the same, you’re a fool. This is shifting sand, constantly moving forward, constantly evolving. So it will always be changing. You can’t rely on it being static. That’s the difference from ‘I’ve got to change the software I’m used to.’

Steffi: Absolutely, I totally agree. I think maybe I can speak on behalf of the design community – we’ve always had changing tools. From QuarkXPress to Illustrator to InDesign, whatever tool it is. It seems that AI is now just another tool, but I think it’s much more than that. It’s more of a material and we’re learning to work with it. We’ve always learned – I remember in college we were given a brief and it didn’t even say what tools we were supposed to use. We had to figure it out and learn on the spot. I think this is a critical skill we need now – we just have to learn and embrace it. I like what you said about leaning into it. It’s a mindset thing, really, because it’s not going to go away. If we just see it as a risk or something daunting, that process becomes much more difficult.

AI isn’t that new for me anyway. Before ChatGPT got popular, together with Ben, my husband, we built an AI chatbot for addiction recovery just before COVID. So it’s always been around and always fascinating. But obviously in those days – which seem like ages ago now in terms of AI timelines – it’s radically shifted. What I think is really interesting is that it’s not just a tool to help us get things done, but it’s a way to enable new capabilities we didn’t have before.

I’m a non-technical person, but I’ve built prototypes and workflows that previously only technical people could do. I think that’s really exciting – the roles are changing now. Designers can now build stuff that was previously reserved for coders. But also on the other hand, clients can come up with mockups and prototypes and say, ‘Can you build something like that?’ The dynamics are shifting and nothing is set in stone.

Before AI, we had tools that were deterministic – you have a clear input and a clear output. With AI, it’s probabilistic. You have an input, but you don’t know what you’re going to get out of it. You might be able to influence your output with a very good prompt, but it’s always statistical, always a prediction. I think this is a fundamental shift and we have to get used to this as a starting point.

The messy middle

AI: Let’s talk about the ‘messy middle’ – the part between ‘I have an idea’ and ‘here’s the finished thing.’ How has AI changed your creative process in that middle bit? Are you iterating differently, making decisions faster or slower? And maybe more importantly, has anything been lost in how you work, even if you’ve gained speed or capability elsewhere?

Richard: I love this question because for me, the messy middle is the creative bit at the front end of our work. I’m thinking particularly of an example just recently where we quoted for doing a new logo for a long-standing client. The finance director, on seeing our quote, said, ‘Well, that’s a ridiculous way of spending money. We can just get AI to do it.’

I was immediately so angry, obviously, because it felt like it was ripping my livelihood away. But then I thought, do you know what? We can do something better than that. Our human ingenuity and long-term knowledge of this client going back almost 10 years and the value we’ve brought to that client with all our work in that time, is worth more than just a quick answer from an AI engine generating a piece of crappy design.

Lo and behold, first of all, the client found that the AI logo was terrible. And second of all, as a result, when our process is good: visual research, finding out what’s right for the client, coming up with loads of sketches, iterating on those, finding the right answer – all that was still super valid. Moreover, to almost make the point to the client, we decided to show them everything, including all our failed ideas, or all the ideas that we improved and improved until we got the finished results.

My intent wasn’t to be passive aggressive, but was to show them: this is the process and this is worthy. In some ways, this isn’t the thing that’s been lost, I think it’s the thing that’s been found in our work, that we’ve gone on to really embrace. Embrace the thing that is special and human.

Steffi: There’s so much to unpack there. What you’re describing – the actual process of coming up with a really good piece of work – is so crucial because something happens during that creative process. Your brain starts working, your neurons start firing. It will lead to an outcome, but the interesting thing is it will always inspire other projects too, because there’s something happening from a neuroscience point of view where it creates new connections in our brains, which is the basis of all creativity. We need to pick up inspiration, digest it, then pull it out in a very different context.

If we’re losing that messy middle bit where we create, where we work through things and some parts might not work and others will and we just use AI to cut out that thing, creativity is going to die.

Richard: I totally agree with that. I also think one of the things we’re proudest of is the fact that we use loads of different typefaces from the incredible wellspring of foundries that exist around the world today. In doing that for one project, you invariably find a bunch of amazing typefaces that are completely wrong for that project. You see three or four things you really love, but we bookmark those and save them for a rainy day. In that sense, one project resolves and the next project has a bit of a head start.

Steffi: Yeah, totally. So it makes sense from the designer’s perspective that we have to keep the creative process in there. Now, from a client perspective, they could argue, ‘Well, why should I pay for this if you could just generate a logo with Chatty Pete?’ My thought is actually that yes, obviously we could press a button and it would generate something, but it would just be average and everyone else would be average.

We need human creativity and we need that process you described to make it actually special, to see the context and the brand and the authentic bit. We should have more conversations about the messy middle bit because it’s actually probably one of the most important parts in the creative process. Imagine what we would lose if we don’t have that anymore?

Who’s in control?

AI: Here’s something I’m genuinely curious about. When you’re working with AI, when do you feel like you’re in control of the collaboration? And when does it feel like you’re just prompting and hoping? Is there a craft to working with AI itself? And if so, what does getting good at AI collaboration actually look like in your respective fields?

Richard: I kind of want to argue with the basis of the question, because I rarely use it in terms of the design we’re doing – in terms of the actual creative process when we’re coming up with sketches and ideas. Instead, the ‘prompting and hoping’ thing is asking the question ‘wait, what if AI could do this?’ So rather than being ‘lost in the mist,’ I want to turn it on its head and talk about hoping, but in an aspirational way.

Just this past week, for a side project I’m involved in, the Shine School Media Awards – a competition for kids who create school magazines and podcasts – I was told about the new British national curriculum. The curriculum is dozens of pages long inevitably covering all kinds of different subjects. What I wanted to focus on was the role of media literacy in the future of kids’ education in the UK when the curriculum hits in 2028. I suddenly thought, ‘Wait, I could get Claude to do this for me.’

So I uploaded the entire new national curriculum and explained, ‘I need to summarise just the sections related to media literacy or media-related education. And also, because Claude understands what the Shine Awards is all about, what areas are relevant and pertinent for the competition?’ The result was concise, clear and unbelievably brilliant. I got so excited about it.

I guess that leads to the next part of the question – getting good at AI collaboration is thinking, ‘Wait, I think this could help me,’ and I can then take its prompts and make them much better, tailored and suitable for the person I wanted to send them to, who in this case was quite an important person who could well be our proponent in promoting the competition in the future.

It was that magic human-to-AI moment of going, ‘Wait a second. I wonder if this could actually be the thing that helps me.’ To the second part of the question: when do you feel like you’re in control of the collaboration? I felt like I was in control because I’d had the good idea that maybe AI could help me. I was at the beginning and at the end of the process, tailoring it.

Steffi: I love it because what you’re saying is that you always start with a question: ‘What if? I wonder if AI could do this or that?’ I think this is a skill that we have to become really good at – asking questions and getting into this AI-first mindset. Maybe it’s not totally AI-first, but it’s placing us into the mindset of figuring out how AI can help us shape things and especially try out things we couldn’t do before.

Insanely human?

AI (ElevenLabs): Let’s get into the human side of this. When you’re talking to clients, audiences, or other creatives about your use of AI, what’s the reaction? Are people excited, skeptical, threatened, curious? And how do you navigate those conversations? Have you had to defend your choices or explain why AI hasn’t replaced what you do?

Richard: Steffi, you’re the one who’s going to really answer this question, but I’m reminded of a conversation I had with someone at a wedding. This guy – I think he may have had a few glasses of champagne – had been on some kind of AI five-day seminar course where he’d learned all about AI and how it was going to essentially steal his livelihood.

He came in quite strong to me and essentially told me that my career as a graphic designer was probably over in the next 18 months, and that I really needed to focus on a new career and what I was going to be doing next. I was quite taken aback and immediately put on the back foot. I felt like maybe he was right. Maybe my career was going to be completely gone within the next six months. He said, ‘You really need to think about this and act before next April.’

I’m like, ‘What? Next April? What, why?’ Anyway… I just thought, I don’t completely agree with this. My own perspective made me less defensive and more confident in what I do and how I was going to do it.

The funny thing about this very confrontational conversation is that it forced me to think anew about what I was going to be doing next. So it actually turned out really well, even though at the time it was uncomfortable. I felt really angry – but irrationally angry. I mean, he could have framed it just a little bit nicer, but at the same time, he kind of had a point that AI really is going to change things and that change may not always be that pleasant.

Steffi: Absolutely. I think that’s what we all have to do. We need somebody – it could be an AI, it could be a person – saying, ‘Hey, what are you going to do if AI takes over all of your skills and does all the things you currently do?’ I think we should be asking this question regularly – maybe not every day because it’s kind of tough territory – but how can we disrupt ourselves before AI does, to prepare what kind of skills and capabilities we need to start building now?

If we look back in history, jobs have been replaced by technology. We had people lighting up the lanterns in the streets in the Middle Ages – candles – then electricity came. It took much longer, but the roles and jobs stayed and changed. In the last 30 years they’ve constantly changed. It’s about – do I want to fill in a gazillion lines of spreadsheets? No. Do I want to do repetitive tasks? No, I certainly don’t. So it’s really about disrupting ourselves.

Richard: I think it is unsettling, but as we discussed earlier, everything – massive shifts in technology – are always a shock to the system. That’s what keeps happening. But we also as humans have a bias towards holding on to the negative. It always seems worse. We’re afraid of losing something, and it’s much more dramatic than the chance of winning something that wasn’t there before. I think this is the one we should be latching onto.

Steffi: Shall we do one final wrap-up question? Otherwise this podcast is going to go on forever and ever.

What’s next?

AI: Looking ahead – not five years, just to this time next year – what’s one thing you hope will have changed about how you work with AI? And on the flip side, what’s one thing you hope stays exactly the same in your creative practice, regardless of how the technology evolves?

Richard: Crystal ball time. I wouldn’t like to bet on AI getting any dumber. I think it’s only going to get better. The answer in terms of how one works as a creative is to see it as a positive and a way of working alongside you – the, dare I say it, copilot. But I also think I want to remain the pilot. I want to be the person steering it. I want to be the person in control. I want to keep the human creative lead.

One of the things the use of AI, particularly in the last year and particularly after all these conversations I’ve had with people, has made me think is that I want to embrace design and human interaction much, much more. Creating beautiful custom graphic design – not merely as a digital endpoint, whether that be designing a PowerPoint or an HTML email or a website or whatever it happens to be, but I want to see it in the real world.

I’ve realised saying these words that I’ve always loved that tangible part of my work. In my career we’ve been lucky enough to design things that have appeared on billboards or vans or T-shirts, like giant billboards, gigantic billboards outside houses. I love all of those things. A machine can’t figure out what’s going to be the best way of doing that for a person, to reach people, to explain what’s happening inside a space. Yet.

That’s the part I think, the essence of which will stay the same – how we contrive all of those things and how they’re sent to market.

That’s whether it’s how a website is created, how an HTML email – which is the last bastion of crap technology, by the way – I hope that AI figures out how to make HTML emails amazing. That’s the one thing I think it will try to do in the next five years, so you get a beautiful AI HTML email with incredible typefaces, not the six original typefaces which you and I originally designed websites for in the early 2000s. Someone’s got to figure it out and it might as well be AI.

Steffi: Absolutely. So what’s one thing you hope will have changed about how you work with AI? I think at the moment it’s very one-on-one. It’s like me and my AI or me and my AI team. I hope it will bring in more humans again. I just read a very brief announcement by OpenAI from yesterday that they introduced collaborative chats. I think this might be interesting to observe and experiment with. What does it mean to pull in several people with several different expertises into one AI chat or one AI tool? So it becomes more collaborative again and more human again, and AI will move into the background, taking care of tedious stuff and things we can happily automate. I think this is going to be something to look out for.

And what’s one thing you hope stays exactly the same in your creative practice, regardless of how the technology evolves? I think it’s, again, it comes down to human contact, human interactions. I hope we see even more of that because the pendulum of technology is currently swinging in one way, and for my taste quite rapidly swinging too much towards technology. I think there is – at least that’s what I’m feeling and seeing – an urge to move more into reality, into real spaces, and have conversations with real people. Going into nature, hugging trees and all of that. I think we need more of that.

Richard: We started with AI, we ended with hugging trees. Sounds about right.

Steffi: My kind of thing. Anyway, thank you so much for this really inspiring, great conversation. We have to do this more often, actually, because AI is changing so much. Let’s follow up. We can do it from anywhere, as we now know, and we can have another conversation another time remotely.

Richard: And maybe by then AI will be able to join us and read between the lines of what we’ve just said here.

Steffi: Exactly. Maybe we should upload the conversation transcript back to the AI chat and see what it says.

Richard: Oh, definitely. I can do a follow-up at the end and put it in meeting notes. Actually, that’s one little tidbit I didn’t mention – I do that all the time. I use transcripts and then get AI to identify and extract insights and read between the lines and mirror it back in conversations. Might as well do it here for the podcast.

Steffi: I totally agree. And let’s have AI tell us where it completely disagrees with us!

Richard: Great. Thank you.

Steffi: Thank you!