Thinking

Synthetic users give you the same headline as real ones and none of the answers you were paying to find.

Each one was a sensible decision. That's the problem.

Every time you set an AI agent to work, it hands you back some time. It won't tell you how much.

For twenty years design was the accusation. Courts and regulators have flipped it: when they want to fix a behavioural harm, the remedy they order is a settings screen.

Uber's self-driving car was built to hand control back to a person in an emergency. It had no way of telling them an emergency had started.

Most brands try to be excellent everywhere. The distinctive ones pick a few moments and design them until they're unmistakable.

Everyone is asking what their AI should refuse to do. The people with the best answer already wrote it down, years ago, and called it a brand strategy.

Air Canada's chatbot linked to the very page that contradicted it. The link changed nothing.

Why distinction is the point of customer experience, and why the answer is sometimes a slow lane.

Trust is what the customer gives you. Trustworthiness is what you are, and unlike trust, it can be designed.