An arrow icon
back to all posts

Conversational AI for Mental Health: What We Learned from Real Users

We recently announced the results of our first study, which offers insights into how people experience and interact with Clare, a conversational AI designed to provide mental health support. Conducted in cooperation with Charité – Universitätsmedizin Berlin (Charité – Berlin University Medicine), the study examined 527 users over eight weeks, exploring their characteristics, motives, expectations, and the development of a working alliance. The results were published in Frontiers in Digital Health

Key findings:

• Users deal with high symptom burden: anxiety (69%), depression (59%), severe stress (32%), and loneliness (86%).

• Key drivers behind using Clare: avoiding embarrassment (36%) and self-consciousness in face-to-face therapy (35%)

• Expectations: emotional support (35%) and a safe space to express feelings (32%)

• A strong therapeutic alliance formed within just 3–5 days (WAI-SR M = 3.76).

The findings underline the unique value proposition of AI mental health support, namely its ability to provide a stigma-free environment people turn to for emotional needs and self-help reasons. Furthermore, the rapid development of trust highlights the potential of conversational AI to engage users early and meaningfully.

While Clare might offer promising potential in mental health applications, the study addresses the ongoing ethical and safety concerns surrounding AI in therapeutic contexts. It also suggests that future research should explore long-term outcomes and consider clinical integration, raising questions about how AI, like Clare, can complement or integrate with traditional therapy. Such exploration could help to clarify the role of AI in the wider treatment landscape.

Stay in the loop!

Don't miss out on the New Blog Posts! We cover mental health topics with valuable insights and tips from Clare and our experts.

Get an email notification straight to your inbox.