
A sculptural still-life experiment.
AI Makes Us Reappraise How We Scope. Whiteboards. Working together in a room. Meeting new people, shaping ideas, challenging each other's thinking. Eventually realising it's time for a coffee break and checking the missed emails and calls - even if it's only your partner saying they'll be late home.
I miss those working sessions.
They were enjoyable, time-consuming and often incredibly rewarding. And I'd started to wonder whether AI was going to see this kind of collaborative activity slowly dwindle away.
I'm beginning to think the opposite might happen.
AI needs context. One of the most important lessons emerging from our early use of AI in the design process is just how important context is.
Much has been said about the ease with which AI can create convincing mock-ups, provide apparently sound rationale, iterate endlessly and increasingly turn those designs directly into code. Give a competent person the right tools, a reasonable prompt and some reference material, and they can produce something remarkably convincing. But convincing isn't necessarily good.
We're already seeing the downside: blandness, sameness and a lack of meaningful differentiation. AI can help us answer how to build something remarkably quickly, but knowing what to build, why we're building it and what should make it different requires much more context.
What does your brand stand for, and how should that be experienced?
Who are your users, and what will genuinely connect with them?
What information, content and tasks matter most?
What are the real problems we're trying to solve?
What should make this experience different from the competition?
And perhaps most importantly: What does good actually look like? These aren't simply things to add to a better prompt. They're things we need to understand before we write the prompt.
The brief becomes more important, not less. This changes how I think about the role of AI in the design process.
As AI takes on more of the production work, the value shifts towards defining the problem, establishing the context and setting the ambition. The brief matters more and the thinking that creates the brief matters more. And the collective experience, judgement and imagination of the people creating it matter more.
AI therefore amplifies the need to front-load scoping and alignment. Before we ask AI to create, we need to give it the right world to create within.
Brand. Users. Priorities. Problems. Constraints. Evidence. Intent. Ambition.
This becomes the context against which everything AI subsequently produces can be generated, challenged and judged. Because when AI can present us with almost infinite options, our challenge is no longer simply creating a solution.
It's determining which solution is right. That's where experienced people become incredibly valuable.
Move the human effort upstream. There's a bigger implication here.
For years, a significant amount of design time has been consumed by production. Creating wireframes. Exploring variants. Moving into high fidelity. Producing specifications. Responding to feedback. Making revisions. Preparing handover. Then making another round of revisions. AI is beginning to compress much of that activity.
The obvious conclusion might be: We need less design. I think that's the wrong conclusion.
Instead, we should ask: Where should we reinvest the time AI gives us back? And increasingly, I think the answer is upstream. Spend more time…
Understanding customers.
Framing the right problem.
Challenging assumptions.
Exploring possibilities.
Understanding the competitive landscape.
Defining how the brand should be experienced.
Deciding what should make the experience distinctive.
In other words, use people where human judgement, experience and imagination have the greatest value - and use AI where speed, iteration and production can amplify them. That starts to reshape the design process itself.
The traditional process often looks something like:
Scope quickly → Design → Produce → Refine → Deliver
An AI-enabled process could put much more emphasis at the beginning:
Understand → Align → Imagine → Define → AI-assisted creation → Judge → Refine
The important shift isn't simply that the second half gets faster. It's that we now have the opportunity to make the first half better.
From producing options to judging them. This also changes what expertise looks like.
When creating five different design directions took days, the ability to produce them was valuable. When AI can produce 50 directions in minutes, production itself becomes less scarce. Judgement becomes scarce.
Which idea actually solves the problem?
Which reflects the brand?
Which will connect with customers?
Which should we discard?
Which unexpected idea is worth pursuing?
And which seemingly polished solution is actually completely wrong?
AI can dramatically increase the number of answers available to us. That makes our ability to ask the right questions - and recognise the right answers - even more important. Experience isn't removed from the process. It moves.
Maybe AI gives us permission to be creative again. There's another consequence of this that I find much more exciting. Perhaps AI gives us time for ideas again.
Too many businesses, products and digital experiences have become accessible, usable, compliant and reasonably effective… but bland.
The moments of delight have disappeared.
The unexpected details.
The anticipation of a user's needs.
The little interactions people remember.
The pure expression of what a brand stands for.
In our drive to create scalable, consistent and efficient digital products, perhaps we've optimised some of the character out of them. AI gives us an opportunity to put some of it back. Not because AI itself provides the creativity, but because it reduces the cost of exploring possibilities.
An idea that might previously have been dismissed because it would take too long to prototype can now be explored. A slightly ridiculous suggestion can be made tangible. Three competing ideas can become ten. We can see them, experience them, challenge them and throw most of them away. That creates space for experimentation again. And perhaps the efficiency AI gives us shouldn't simply be converted into doing the same work faster.
Perhaps some of it should be spent on being more ambitious.
Bring back the whiteboard. Which brings me back to those workshops. Let’s put people in a room together and interesting things happen.
> Someone challenges an assumption.
> Someone asks the question nobody else thought to ask.
> An idea from one person triggers a completely different thought from another.
> A customer insight gets connected to a technical possibility.
> A ridiculous suggestion turns into something surprisingly good.
> People disagree.
> Someone draws something terrible on a whiteboard.
> Someone else sees something in it.
And eventually, somewhere between the Post-it notes, coffee and conversation, an idea begins to emerge.
That collaboration brings different perspectives together, challenges established thinking and creates the conditions for innovation. Perhaps the workshop doesn't disappear in an AI-driven design process. Perhaps its purpose becomes clearer.
We don't need ten people in a room deciding where the button goes - AI can help us with that. We instead need them deciding:
What we're trying to achieve.
What problem is really worth solving.
What customers actually need.
What should make the experience different.
What we believe in.
What we're prepared to challenge.
And, ultimately: What does good look like?
Then we give AI that context and let it help us make those ideas real. AI may dramatically change how we produce things. But perhaps its greatest contribution to design won't be allowing us to spend less time together.
It will be giving us the opportunity to spend more of our time together on the things that matter.
