On this episode, I speak to Hugo Alves, co-founder of Synthetic Users, about one of the most controversial topics in modern product development: using generative AI to simulate users for research and decision-making. Hugo has a background in clinical psychology and product, and has spent the past three years building a platform that generates synthetic qualitative interviews to help teams reduce risk and make better decisions.
Episode highlights:
- What Synthetic Users actually is - generating in-depth qualitative interviews with AI-powered "synthetic" participants to help teams reduce risk and accelerate discovery
- Most companies don't do enough, or any, research in the first place, and they need as many tools in their locker to help with the ultimate goal: making great products.
- The pragmatist's view of AI - why Hugo doesn't care whether LLMs are "conscious", only whether they produce useful outputs
- The agentic "swarm" approach - using specialised sub-agents (planners, interviewers, critics) instead of one giant prompt to improve quality and reduce drift
- B2B vs B2C - why synthetic research works well in B2B contexts, and the harder (future) problem of modelling organisational dynamics
- Bias, sycophancy and realism - the technical concerns around LLMs and how to validate responses with pilots and human comparison studies
- How to use synthetic research in practice - filtering ideas, informing human interviews, and treating it as an accelerant rather than a replacement
- "It shouldn't exist" - the moral argument against synthetic users, reacting to UX thought leaders and their objections, and why some of those objections aren't really about evidence
... and much more.
Contact Hugo
