Can AI tell researchers what matters?
AI can read a transcript in seconds and hand back a set of coherent, credible-sounding themes. Coherence and correctness are two different tests, and the real insight usually sits in the gap between what someone said and what they meant.
Devanshi Ray is a behavioural scientist who has spent her career on the gap between what people say and what they actually do, most recently at Swiggy. In this session she walks through where AI genuinely helps, where it misleads, and the checks she runs to keep researcher judgement in the researcher's hands.
What we'll get into
Why she writes her own read on the themes before AI touches the transcript, and how easily you anchor to a confident first summary
Why frequency is a poor proxy for importance, and what gets lost when the single-participant observation gets dropped
How to push past the immediate "why" to the psychological mechanism underneath, and why that mechanism is what makes a finding usable somewhere else
What she's tried with AI and abandoned
How to hand findings to product teams so the reasoning survives the journey
Who it's for
UX researchers, insights leads, and anyone being asked to move faster while keeping the thing that makes research worth doing.
What you get
Attendees get the recording and a takeaway resource built on Devanshi's framework.
When
Thursday, 19 November 2026
10:00 PM IST · 4:30 PM GMT (UK) · 8:30 AM PST
Online. You'll get the joining link before the event