In 30 seconds
- AI can summarize research and generate prototypes fast.
- It can't replace talking to real users.
- Faster prototypes mean more ideas tested, if you test them.
Useful applications
- Summarizing interview notes and support tickets
- Clustering feedback into themes
- Generating clickable prototypes quickly
- Drafting hypotheses and experiment designs
The risks
- Synthetic users or personas mistaken for real evidence
- Summaries that smooth away surprising details
- Building faster than you can validate
Staying grounded
Use AI to prepare and synthesize, but keep real users at the centre: interviews, observation and experiments. Cheap prototypes are only valuable if you put them in front of people. See product discovery.
Myth vs reality
Myth · AI personas replace user research.
Reality · They can't tell you what real users do.
Myth · Faster prototyping means faster learning.
Reality · Only if prototypes are tested with users.
Myth · AI summaries are objective.
Reality · They reflect what they were given and how they were asked.
FAQ
Should we use synthetic users?
Only to prepare questions, never as evidence.
How do we check AI summaries?
Spot-check against raw notes.
Who owns discovery decisions?
The product trio, as before.