Insights

Notes on trust, adoption, and the behavior of AI users.

Trust · Coming soon
Why users abandon AI features after one bad answer
A single confidently wrong answer costs more adoption than ten quiet failures. Here is why, and what to design instead.
Reliance · Coming soon
The over-trust problem nobody budgets for
When users trust an assistant too much, the errors it does make land squarely on your customer. Calibrated reliance, and how to design for it.
Regulation · Coming soon
What the EU AI Act means for your product's design choices
Manipulative and deceptive patterns are now an enforcement matter in Europe. A plain-language read on what changes for product teams.