I use AI as part of a practical design method: framing problems, generating routes, and building things people can react to.
The point is not to hand work over to a machine. The point is to move from a loose idea to something visible, testable, and worth discussing — faster than would otherwise be possible. The human part stays in the framing, the taste, the editing, and the final call.
I define the question properly, review constraints, gather references, and use AI to widen the field early without losing focus.
I use AI to explore multiple concepts quickly, compare routes, and surface patterns worth developing rather than chasing novelty for its own sake.
I turn the strongest directions into visible prototypes, interface concepts, or experience flows that people can respond to.
I review what holds up, cut what does not, and refine the strongest idea into something more coherent, useful, and convincing.
I shape the final concept into a clear story: prototype, rationale, system thinking, and next steps for development or further testing.
Synthesising references, patterns, and precedents faster than manual review allows.
Producing multiple directions quickly so the best ones can be identified and developed.
Getting ideas into a visible, interactive form early — before too much time is spent on weak directions.
Comparing options side by side and identifying what holds up under scrutiny.
Structuring the reasoning behind choices so the rationale is clear to everyone involved.
Starting question How might a complex data dashboard be structured for non-technical users?
My role Used AI to generate 12 layout directions. Selected and refined two. Built one interactive prototype.
Outcome A working prototype the client could react to within the first week — before any visual polish.
Starting question What are the key tensions in this brief and what are the strongest design angles?
My role Used AI to map the brief, surface contradictions, and identify three credible directions before the first workshop.
Outcome A structured framing document that shaped the entire project — produced in half a day.
Starting question What would this service look like if we designed for the edge cases first?
My role Generated speculative scenarios, mapped edge cases, and built a lo-fi concept for stakeholder review.
Outcome A provocation that shifted the client's thinking about their core use case.
AI helps me move quickly, but the thinking, editing, and final decisions stay human. The strongest version of any project is still one where a person takes full responsibility for the outcome.
I don't use AI because it sounds current. I use it because, in the right place, it genuinely helps — widening the field early, generating comparisons quickly, and surfacing patterns worth developing. It is a tool inside a disciplined process, not a replacement for one.