BHT2026 · Nashville · September 22–24, 2026

Let's talk about what meaningful AI support looks like in practice.

Valerie Forman Hoffman, our Head of Research, and Lukas Wolf, our co-founder and CEO, will be at Behavioral Health Tech this year. Find us at BHT to compare notes on how AI fits into real-world support.

Find us at BHT Explore a partnership

Research to discuss

Three papers, one practical question: what should the field measure?

We have work on short-term emotional-wellbeing outcomes, crisis-risk safety architecture, and how engagement patterns relate to outcomes. All three papers are available as public preprints and linked below.

Randomized controlled trial
Short-term Evaluation of a Voice-Based Generative AI Emotional Support Companion: A Randomized Controlled Trial
400 adults, four weeks, randomized against web-based psychoeducation. At two weeks, the Sonia group reported greater improvements in mood and anxiety symptoms than the control group.
Preprint on PsyArXiv
Safety architecture
Suicide- and crisis-risk detection using large language models in mental-health chatbots
How Sonia recognizes a conversation that goes beyond what AI should handle and routes people to appropriate human support.
Preprint on medRxiv
Engagement patterns
How different patterns of Sonia use relate to emotional-wellbeing outcomes
A cluster analysis of the randomized trial found no significant outcome differences between sporadic and power users. Improvements were not limited to the people who used Sonia most often.
Preprint on Research Square

Meet us in Nashville

If you see us, come talk to us.

We are especially interested in how payers, providers, employers, and researchers are evaluating AI support: what evidence they trust, where human judgment belongs, and which outcomes matter after the novelty wears off.