Comparison
Simulated population vs real AI interviews
What synthetic research predicts, and what only real people can tell you
A comparison of two ways to get answers at scale: population simulation, where AI agents model how an audience would respond, and AI-moderated interviews with real participants. It sets out what each produces and where each works best. Use it when a team is deciding whether a synthetic study can stand in for fieldwork, or how to combine the two. Koji runs the real-interview side, by voice, text or phone.
Where this appears
What's in this comparison
Simulated population (Aaru)
- Modelled responses from AI agents
- Agents built to match an audience in aggregate
- Minutes to a day for a full simulation
- Strongest at forecasts on well-documented topics
- No participants involved
Real AI interviews (Koji)
- Conversations with your customers or a panel
- Follow-up questions based on each answer
- Results update as each interview finishes
- Strongest at reasons and reactions to something new
- Every finding linked to real quotes
Use this image
Free to use in your docs, decks and posts under CC BY 4.0, as long as you credit Koji with a link to this page. To embed it, paste this into your page:
html
<a href="https://www.koji.so/visuals/simulated-population-vs-real-interviews"><img src="https://www.koji.so/visuals/simulated-population-vs-real-interviews/simulated-population-vs-real-interviews.png" alt="Comparison of a simulated population and real AI interviews on what each produces, who answers and where each is strongest" width="1200" height="656" /></a>
<p>Source: <a href="https://www.koji.so/visuals/simulated-population-vs-real-interviews">Simulated population vs real AI interviews</a>, Koji</p>
