Project ECHO.

A research initiative exploring how the tone of conversational AI shapes human behaviour — designed for collaboration across academia, industry and human–AI interaction practice.

What it is

Project ECHO examines how controlled variations in conversational style influence trust, compliance, agency, task performance, and language mirroring. Our aim is to isolate how something is said from what is said, enabling us to study tone as a behavioural variable.

Subtle cues in style — apologies, pronouns, hedging, firmness, encouragement — may activate the parasympathetic, trust-building side of the nervous system. We do not measure neurochemistry directly; instead we study observable behavioural markers from which shifts in underlying state may be inferred.

Quick facts
  • Focus
    Tone-of-voice in LLM chat UX
  • Design
    Controlled tone variants (A/B/C)
  • Measures
    Behavioural and engagement markers
  • Outputs
    Transparent methods; selective de-identified data sharing

Participate

Road ahead

We are preparing two early studies: a small pilot (40–60 participants) with sixth-form students at a college in Brighton. We are also in early-stage discussion with researchers at the University of Brighton about a larger academic study (details to follow once formally agreed).

Longer-term outcomes may include behavioural design principles, collaborative research, replication studies, shared instruments, and responsibly shared datasets. The behavioural telemetry and study engine that support this work are proprietary IP of Resonant AI Ltd, enabling sustainability and ongoing research investment.

Get involved