
Generative UI Still Needs a Designed Sequence
AI models can now generate the interface. In AI strategy tools, the harder problems are drift, numbers nobody can trace, and a model that agrees too easily.

Longer-form thinking on leadership, innovation, and design. Lessons from 25+ years of building products, leading teams, and helping organizations find what's worth making next.

AI models can now generate the interface. In AI strategy tools, the harder problems are drift, numbers nobody can trace, and a model that agrees too easily.

Now that AI can build almost anything, designers are calling taste the moat. Taste that isn't tied to a where-to-play and how-to-win choice sets the designer apart, not the business.

A panel of synthetic CFOs matched my conversations with real CFOs because a design team built it from fifteen real ones. Synthetic users work when someone talked to real people first, and all research has a shelf life.

Design thinking lost its teeth in business because people like me made the method the point. Drucker named the point in 1954: a business exists to create a customer. A confession, and an upgrade for designers and the executives who fund them.

In product circles, systems thinking usually means the machinery that builds the product. The system that decides whether the product matters connects a customer, a market, and a company's choices. Here's how I learned to see it.

At Gordon Food Service my team moved customer accounts to online ordering 20% a quarter, three quarters running. The target kept leadership investing. It also meant every new account landed on problems we had chosen to leave for later.

AI makes screens and prototypes cheap to build, so the scarce skill on a product team is choosing what deserves to exist. That choice belongs to design leadership.

Early hallucinations spent the trust professionals had in AI tools. We earned it back by splitting the work between the model and plain calculation, and showing people where one ends and the other begins.

My first blog post, in 2002, asked why we were rushing to use Flash. The questions I ask about AI products today are the same ones, and I still get them wrong.

The design work in AI products moved from screens to sequences, prompts, and output contracts. It's still UX, and it still starts with where people actually are.

Why the UX community's obsession with user delight is killing business value.

Culture can deliver or destroy strategy. A distribution company's sales culture shows how resistance to change can undo a customer experience transformation.

When leaders refuse to choose—between legacy and future, service and scale, investment and cost—they fall into the tradeoff trap.