Overview
Our AI Therapist Agent platform is not yet in market—but it’s built, functioning, and now backed by new independent clinical research that supports its foundational concept. A recent randomized trial by researchers at Dartmouth College found that AI-powered therapy can significantly reduce symptoms of depression, anxiety, and eating disorders. While our system is not yet licensed or deployed, this study affirms our direction.
What sets our approach apart is the integration of user sensing. Unlike traditional chat-based therapy bots, our AI Therapist Agent is designed to actively interpret cognitive signals from the user—detecting when their brain and body are resonating with a statement, or showing subtle signs of dissonance, doubt, or emotional conflict. This capability opens the door to more precise, personalized, and effective therapeutic interaction—especially when grounded in cognitive behavioral therapy (CBT) principles.
The Study That Supports the Direction
In March 2025, Dartmouth College researchers published results from a clinical trial using an AI chatbot named Therabot. Over an eight-week period, 106 participants with major depressive disorder, generalized anxiety disorder, or eating disorders engaged with the bot for CBT-informed therapeutic support.
The results were significant:
Depression symptoms dropped by 51%
Anxiety symptoms fell by 31%
Eating disorder symptoms decreased by 19%
Participants described the experience as emotionally engaging, accessible, and beneficial—many said it felt as supportive as working with a live therapist. These results validate the effectiveness of AI in delivering therapeutic care, even in a fully text-based format.
Our Unique Differentiator: Cognitive State Sensing
Our AI Therapist Agent goes a step further. It’s not just a conversational chatbot—it’s a sensing agent that interprets moment-to-moment cognitive patterns in the user. This is made possible through optional neurotechnology integrations (like EEG or fNIRS) that allow the system to detect:
Resonance: When a statement reflects internal truth or self-alignment
Dissonance: When the user experiences emotional discomfort, doubt, or resistance
Cognitive friction or disengagement during specific statements or reframing attempts
This sensing layer enhances traditional CBT by allowing the AI to not only deliver interventions but adapt in real time—focusing on statements that activate meaningful reflection or reveal avoidance. The result is a conversation that not only sounds therapeutic—but actually responds to the brain’s real-time feedback, even when the user isn’t consciously aware of it.
Customization Through Licensing
The platform is not yet licensed, and certain features—such as specific therapeutic flows, biometric triggers, and escalation protocols—are intentionally undeveloped. We are seeking a licensing partner to define these with us, based on:
Clinical scope or target population
Integration with their existing health or wellness stack
Level of sensing used (with or without EEG, for example)
Preferences for tone, escalation logic, or optional human-in-the-loop design
This partner would help shape the final version of a neuroadaptive AI therapist agent that moves beyond words—and begins responding to the deeper signals behind them.
A Collaborative Path Forward
With validated AI therapy models gaining public and clinical credibility, now is the right time to bring sensing-based agents to market. We are looking for a partner—healthcare system, research group, university, or wellness platform—ready to:
Customize the platform’s capabilities
Lead pilot deployments or clinical validation studies
Collaborate on data strategy, IP positioning, and ethical frameworks for sensed feedback
References
Sarda, P., Bottrill, E., et al. (2025). AI chatbot Therabot found to significantly reduce symptoms of depression, anxiety, and eating disorders in randomized trial. Dartmouth College. ScienceDaily
Fitzpatrick, K. K., Darcy, A., & Vierhile, M. (2017). Delivering cognitive behavior therapy to young adults with symptoms of depression and anxiety using a fully automated conversational agent (Woebot): A randomized controlled trial. JMIR Mental Health, 4(2), e19.
Bickmore, T. W., Trinh, H., et al. (2018). Automated support for self-management in serious mental illness: A virtual health counselor for behavior change. ACM Transactions on Interactive Intelligent Systems, 8(2), 1–28.
Mishra, V., Bajaj, V., Pachori, R. B. (2020). Detection of focal EEG signals using spectral-temporal features for seizure prediction. Biomedical Signal Processing and Control, 57, 101731.
Ellamil, M., Dobson, C., et al. (2016). Identifying introspective and extrospective states using fNIRS in naturalistic environments. Frontiers in Human Neuroscience, 10, 329.