Do Synthetic Users Actually Work?
Synthetic users promise research answers in minutes, with no recruiting and no scheduling. Can you actually trust what they tell you? And if you can't fully trust them, what are they good for?
In B2B, every interview slot is expensive. The customer base is small, the same names get called again and again, and the business asks more questions than that pool can absorb. Synthetic users look like a way out. The real question is when they help, and when they're just telling you what you want to hear.
In Episode 2 of The Insight Loop, Looppanel's live conversation series for research and product teams, we sit down with Bruno Duarte, Manager, Product Design at LearnUpon, to work through it honestly.
🔍 What we'll dig into
Why a synthetic user built from your own personas, ICPs and past studies behaves very differently from one a model makes up
What they're useful for: stress-testing a concept, catching dead questions in your discussion guide, and checking an idea against what you've already learned
Where they break, and the decisions you should never make on synthetic input alone
The Pre-Flight Check: a simple way to decide which questions deserve a real participant
The short version: the output isn't an insight but a better interview.
👥 Who should attend
UX researchers weighing AI participants for their own studies
Product designers and PMs who need a fast directional read without burning a research cycle
Research leads at B2B companies with small, over-asked participant pools
🎁 Every registrant gets The Pre-Flight Check, a short kit with a readiness checklist, a question triage grid, and a way to test synthetic users against a study you've already run.
🎤 Speaker: Bruno Duarte, Manager, Product Design, LearnUpon
📅 Date & Time: Wednesday, 21 October 2026, 10 AM Pacific
💻 How to Join: Join live from anywhere through a link we'll be sharing closer to the event