Rewriting how we work

Inside a cross-functional working group defining an AI-native product development lifecycle: showing up both as practitioner and skeptic.

01 - What I can share

Companies don’t get AI transformation by buying a tool. They get it by redefining, at close range, what design does, what engineering does, what research does, and where the seams between them sit. That redefinition is a design problem in its own right — and it’s the one my working group is running.

I sit in the design seat of that group. What I contribute is not policy — it’s evidence, framing, and working examples of what the new PDLC actually looks like when the team runs it end-to-end.

Cross-functional working group. Every role in the product development lifecycle has representatives at the table. The mandate is company-wide: propose, pilot, and roll out the new ways of working that AI tooling makes possible — and identify what it removes from the table.
Design’s contribution is upstream of policy. I don’t write the policy. I bring the working examples the policy has to reconcile with — the actual prototype built in an evening, the actual skill saving a team hours, the actual gap between what an AI tool can do and what a product decision needs it to do.
Design × engineering vocabulary. A meaningful share of the work is translation. Design speaks in intent, edge cases, and journey; engineering speaks in interfaces, invariants, and tests. AI tooling sits on the seam. Bridging the two vocabularies — carefully, at the level of concrete artifacts — is where a lot of the value shows up.

Confidential, Company-specific direction under NDA

02 - How I show up

Not officially — informally. These are the three modes I move between across the working group.

Role 01

Practitioner

I use the new tooling on real work every day. When the group debates whether a workflow could work, I can pull up the version I already run — and, more importantly, name where it breaks.

Role 02

Translator

Design and engineering do not share a vocabulary for AI tooling yet. I sit on the seam and rephrase — turning an intent-language design concern into an interface-language engineering constraint, and back.

Role 03

Skeptic

A working group in an AI transformation is under pressure to declare early wins. I hold the line on evidence — running the new workflow against a real project before declaring it standard, and flagging what regresses when a new practice is adopted.

03 - Where it stands

This is work in progress — the playbook isn’t finished, it’s being written by the work we actually ship. What I’m certain of, and what I keep pushing for in the room, is that trying, failing, and iterating beats holding out for the perfect approach up front.

The perfect approach, if it turns up, will be the version we’ve already run three times and edited twice.

The transformation is not that the tool is new. It’s that a single designer, in a single evening, can now do what used to take a squad a week — and the organisation has to catch up with what that changes.

Running the same conversation?

If your org is figuring out what AI-native product development looks like, I can help. My experience — what’s landed, what’s regressed, where the design seat actually moves the needle — is portable, and I’d bring it to your team. Company-specific detail stays under NDA.

Get in touch Back to AI Work