Case study in progress

Deloitte AI Avatar.

A SCADpro × Deloitte study of what actually drives trust in AI avatars — and which scenarios they belong in.

The dissolved portrait used throughout the project — a human silhouette resolving into colour

Client: Deloitte Digital, through SCADpro
Year · Duration: 2026 · 10 weeks
Mission: “Understand why and when a customer is most likely to engage with a human-like AI avatar solution vs other engagement tools, equipping our team with the insight to strategically align AI avatars with use cases tied to customer segments’ specific needs.”
Hypothesis to test: “The human-like appearance of AI avatars will help create trust and comfort in customers as they navigate various use cases.”
Evidence: 1,021 survey responses (985 analysed), 24 paired interviews, a scan of 27 avatar products
My part: Design lead on every deck, the white paper and the visual language running through both; co-designed the interview protocol and the coding framework; survey blocks, analysis and the final report

Coming to this page

What the full case will cover

The six hypotheses Deloitte arrived with, and what the evidence did to them

Why people default to text until they have watched an avatar work

Which scenarios an avatar belongs in — and which ones regulation keeps it out of

The six principles the client left with

The research is finished and the findings are cleared to share — the write-up is what is still being built. Want to see it before it is published, or talk through the study? Email me.