Feel to know: Using emotion to measure real understanding.

The argument

We have spent decades measuring understanding by asking people to demonstrate it — tests, assessments, quizzes, surveys. The problem is not methodology. The problem is timing.

Understanding arrives emotionally before it arrives verbally. The moment a concept clicks happens in the 200–500 millisecond window before anyone can write it down, tick a box, or tell a trainer they've got it. By the time any self-report measure captures a response, the lived moment of comprehension — or confusion — has already passed.

Emotion signals understanding more accurately than any survey or test. And AI can now measure that signal in real time, individually, remotely — without disrupting the training environment that produced it.

The technology behind the talk

EchoDepth is the emotional AI platform built by Cavefish. It uses 44 facial Action Units from the FACS taxonomy — the scientific standard for emotion measurement (Ekman & Friesen, 1978) — to score Valence, Arousal and Dominance in real time during any browser-based session.

No hardware. No facility. No specialist operator. Participants join on their own device. EchoDepth captures the involuntary physiological response at the moment of stimulus exposure — before any verbal or written response has formed — and translates it into structured, actionable insight.

Deployed across pharma research, FMCG consumer insights, culture measurement, and — as this session demonstrates — human performance and training contexts.