Hi, I'm David.

Engineers were amongst the early adopters of AI in their daily work, and designers followed. User research came later, and mostly for the last step, summarizing what people had already said. That gap is where User Sage began.

The question was how research actually starts using AI, from the first draft of a study rather than the final summary.

When User Sage was just an idea, most teams only used AI for one part of user research, the analysis at the end. Transcripts went in, summaries came out. A few teams were experimenting with AI-moderated testing. Useful, but AI stayed out of the decisions that mattered. It showed up after the fact, in the summary, instead of helping shape what to test or what to ship.

There was a reason for that. Researchers didn't quite trust AI yet. Nobody wanted to walk into a product review and quietly wonder whether the findings were real. So teams split: some trusted everything the AI produced, some wouldn't touch AI at all, and both were mistakes. What was missing was a clear line between when AI help is enough and when a decision needs real humans.

Underneath that was a second problem: even when teams wanted real human input, getting it was slow. Finding the right participants, keeping a panel warm, returning to people you've already talked to instead of starting from zero with every study. Access to real people has always been the hardest part of research. Fixing that is the other half of why User Sage exists.

I wanted to solve the trust problem first, which meant bringing AI into the whole research process instead of bolting it on at the end. It started as my question and is now a team's work. The whole company rests on one sentence.