Effective conversational AI in telemedicine

Project supported by British Academy Innovation Fellowship (Brandt) IF2223/230141

This project brought together expertise from Applied Linguistics at Newcastle University, and Ufonia, a market leading organisation in conversational AI, to investigate what makes interactions between humans and AI systems most effective.

Conversation Analysis was used to examine interaction in calls between NHS patients and Dora, an AI clinical assistant used for autonomous telemedicine services. This helped identify areas for development of Dora’s conversational structures, which were then implemented and trialled. Subsequent cycles of analysis assessed the success of the interventions.

News items

“Patients love it. We have had only one person, of all the calls that have been made, who said Dora wasn’t personable." - Frimley Health NHS

The project team

From left to right: Joe Tindale (Lead Software Engineer, Ufonia), Kleopatra Sideridou (PhD Candidate, Research Assistant), Nikoletta Ventoura (Conversation Designer, Ufonia), Rory McKinnon (Product Engineer, Ufonia), Spencer Hazel (Conversation Analyst), Adam Brandt (Conversation Analyst)

Exploring with Dora: identifying principles of effective conversational AI through an autonomous telemedicine service

Links to external sites