Designing for calibrated trust
The goal of an AI interface is not more trust. It is the right amount. Two decades of human-AI research, distilled into the design moves we reach for first.
Read the essayJournal
Notes on AI products built to endure.
Essays from the studio on designing, building and running AI products that people trust.
Engineering / 4 min
Look at the data: evals for AI products
Most AI features fail quietly, in ways no dashboard predicted. The fix is unglamorous: read real outputs, name the failures, and measure only those.
Strategy / 3 min
Start with a workflow, not an agent
Agents are the most exciting pattern in AI products, and often the wrong first one. Why we begin with the simplest system that could work, and when we graduate.
Design / 2 min
Design is the interface to the model
Two products can call the same model and feel nothing alike. The difference is almost never the prompt. It is the design around it.
Engineering / 1 min
Build for the next model, not this one
The model you launch on will not be the model you run in a year. A few architectural habits keep that change cheap.
Strategy / 1 min
When a feature should not use AI
Part of designing AI products well is knowing where a model does not belong. Four signs we look for before reaching for one.
