Design used to stop at the handover. Mine doesn’t.

Graham Powell-Wood

Early in my career I spent late nights drawing redlines for developers in Photoshop. A plugin took the job, then Zeplin removed the task entirely, and its value went with it. I’ve been watching what tools take, and what they leave us, ever since.

For fifteen-plus years I designed for large organisations: as a founding member of Deloitte Digital, then as creative director at Somo and across EMEA at CI&T, where the design team grew from five to thirty. The work ranged from Audi’s design language and Tesco Bank’s app to cinch, which went from nothing to launch in six months. The lesson that stuck: the experience is downstream of how the organisation works.

In 2023 I wrote that AI would keep climbing the value chain, and that a designer’s job would become judgement. In 2024 I left to test it. I co-founded Alternative Company, a home for AI-native companies: small teams, agents doing the work that doesn’t matter, people owning the calls that do. AltCo helps other organisations work that way; Laid & Wove and Cartogra are what we build for ourselves.

It’s the same job it always was: designing how an organisation meets the people it serves. Only the material has changed: screens agents, brands context, services judgement. I still think like a designer, about the person on the other end, about what to leave out and about whether the thing is any good. The difference is that I now build what I design, from the first prototype to the system in production.

How I work
Show, don’t claim.A working prototype settles more arguments than a deck.
People own the decisions.Agents do the work; someone accountable makes the call.
Knowledge decays.What I knew about this field a year ago is already out of date, so I keep building.
Leave things out.The best design decision is usually the one that removes something.
Talk to me about

AI-native companies, the end of the leverage model, designing for agents, and the ethics of all three.