AI summary
- In two years, AI went from an occasional tool to an almost daily one, with Claude as my main tool.
- I don't design any faster, but my time is better spent: I focus on what brings the most value.
- Four key uses: illustrations when there's no budget, quick prototypes, desk research and survey analysis.
- The environmental impact of AI remains a real dark spot, one I haven't solved yet.
It's moving fast, really fast
My use of AI went from "fairly occasional" to "almost daily" in less time than I would have imagined. The habit set in very quickly, taking up more and more space without me even noticing. Writing a quick email here, getting an answer on a niche topic there… Today, a working day rarely goes by without my good friend Claude.
Choose your weapon
I've already given the game away. Like many people, I started with ChatGPT. I dropped it quickly because the answers were a bit off the mark and, compared to an internet search, it didn't add much.
Then Zorro Claude came along. I noticed the difference right away. Its understanding of context is exceptional. Its answers anticipate our questions. It's like a friend you tell you're a bit hungry around 6 p.m., who comes back with a platter of chorizo, cheese, cherry tomatoes and a cold beer. You've barely finished the beer and another one is already in your hand. Naturally, you want to stay.
And it doesn't stop at conversations with Claude: you can feel that same understanding in how Anthropic's products evolve: Claude Code, the desktop app, the creation of the MCP protocol… Every update is just so… logical and well thought out.
I don't design any faster
I can only speak for my own job, but I firmly believe AI hasn't made us more productive. I don't create more screens, nor faster. Designing doesn't take me less time thanks to AI. Sure, Figma's built-in AI saves time on repetitive tasks (duplicating and filling tables with placeholder text, for example), but that's a different use from Claude and the like. My time is better spent: I can focus on the most important tasks, or the ones where I bring the most value. I can explore more options because Claude saves me time on very time-consuming tasks.
What has really changed
Many voices in graphic design have spoken out to say AI is killing our professions. I'm one of those who think the market butcher who makes his poster with ChatGPT would never have spent a single euro on a graphic designer anyway.
I'm not saying it's better: we're only at the beginning, and I already loathe all these soulless copy-pasted posters (a chalkboard does that job far better than ChatGPT junk). I'm simply saying we should compare like with like. In my job, thinking and logic play a big part (UX). Even though there's basic knowledge and rules, most of the information lies with the client or their audience. Today, humans can't be replaced for that part.
However, once that data is collected, reprocessing it, sorting it and extracting its essence is quite tedious. I believe this is where AI has a card to play.
How I bring Claude into my work
After a few months of using it for everything and above all for anything, I took some time to think about how to really make the most of AI:
- What does AI do better than me?
- What can't AI do?
- What does AI do more easily, where I bring less value?
- How can I bring as much value as possible on top of what I already bring myself?
1. Illustrations
Yes, this may sound a bit contradictory, let me explain. On some projects I've worked on (Fonds Prévention Argile, for example), there was no illustration budget available (at launch). With AI (Midjourney at first, then Nano Banana and a bit of ChatGPT lately), I was able to add (sparingly) a few illustrations or images to liven up some rather dull websites. Once again, without AI these images simply wouldn't have existed.
Where this differs from our friendly butcher mentioned above is that my prompt (with my design background) is extremely precise and targeted. I know what I want, I specify a style, a format, etc. The output doesn't look like all the posters you see everywhere. It's not better or worse than our butcher, it's just different, and it actually helps avoid the sameness I blame AI for.
2. Testing quickly
By giving Claude design systems (notably at Fonds Prévention Argile), I've been able to quickly generate interactive prototypes that let me explore options at low cost before committing, and iterate efficiently, which I probably couldn't have done before. Having time to test this much at the start of a project is a luxury; it's rare.
3. Knowledge base
Probably one of my favourite aspects: drawing on reference websites to pull out UX rules, patterns… Nielsen Norman is pretty much my bible. When you start a project with no foundation and no information, you need to rely on something solid. This site offers UX/UI standards; everything is documented, justified and based on decades of user research. I set up a Claude project that draws on these resources and gives me factual answers very quickly. For desk research, I haven't found anything better.
4. Processing survey results
User surveys, whatever they are, are a goldmine of information. You still have to squeeze the juice out of them. And that can take a lot of time and, let's be honest, it's extremely tedious. Claude has allowed me, while taking care to anonymise the data, to:
- Easily process large volumes of data from quantitative questionnaires, and quickly pull out stats and trends.
- But also, and above all, analyse qualitative data and open-ended answers, create broad categories (thematic analysis) and count how often they come up, to quantify that qualitative data. This already existed, but how easy it has become changes everything. I love it.
Let's not kid ourselves
I won't go into detail on this point, as I don't know enough about it, but the environmental impact is a real dark spot for AI, and it's a problem. It bothers me a lot when using it, and I haven't yet managed to find the right balance.
To conclude: do we really have a choice?
We always have a choice, but I believe that here, if we want to keep growing, the answer has already been chosen for us. The question is rather: how do we bring it into our work without losing what makes each of us unique, while continuing to improve what we deliver?
As for me, for now, I've chosen not to just put up with the situation but to turn it into a strength. I see AI as extra possibilities; it's up to me to use them wisely.