Beyond the 10-Minute Interview: Scaling a Longitudinal Community with AI

When we hear ‘research at scale’, most of us picture a study with thousands of participants turned around in 48 hours. But in qualitative research, the real challenge can also be maintaining depth and rigour over time. Drawing on a recent longitudinal community project featuring 16,000 participant posts, conducted in partnership with Sparkminds Insights, this session explores how to use AI to support high-volume qualitative research.

We will look at how AI stepped in to do the heavy lifting: reading responses, acknowledging participants instantly, and actively moderating and probing to extract deeper insights. To put that into perspective, reviewing those 16,000 posts is something that would have taken a human moderator 11 weeks, if that was all they were doing for every working hour over those 11 weeks.

But making this work meant realising AI isn’t a magic button. It is not a ‘set and forget’ tool that you can just leave to ‘crack on’ unsupervised. We will share what it really takes to work hand-in-hand with the tech, from the initial ‘test and learn’ phase (including what can happen when you leave an AI to its own devices) to figuring out the ideal division of labour. Because while AI can execute the hard work of moderating thousands of posts, ultimately, the AI is only as good as those instructing it.