
Data collection, admin and decision
support are key areas
Replacement isn’t the answer,
making therapist’s lives easier is

Through usability research study with clinical professionals we uncover critical needs and considerations when building mental health AI software for therapists, in a NeurIPS 2022 accepted poster.

Broadly (and globally) adoption of mental health technologies has been poor1,2,3 and despite the promise, a trend is emerging: solutions are failing to deliver real-world impact4,5,6. Sustained adoption is part of the problem.
User-centred design is a cornerstone of software development and critical if digital tools are going to deliver value for the populations they serve. But in mental health it’s rarely practiced as often, or as completely, as it should be7,8. Mental health digital tech is not fulfilling its potential. A lack of end user engagement in the development process is a contributing factor.
In order for the mental health community to trust, value and adopt AI software tools they need to be involved in the development process. This usability work demonstrates methodology for collaborating with end users and reveals clinical perspectives on how AI should be deployed to meet their needs.
“The design and the colour of the chatbox are warm and soothing,
the tone of the bot is personable, friendly also professional.”
- participant, usability study