Unlocking the Process of Psychotherapy

Lyssn uses Automatic Speech Recognition and Artificial Intelligence to unlock the process of psychotherapy: making feedback faster, more detailed, actionable, and reliable. With AI-based assessment of sessions, clinicians and supervisors have a greater opportunity to reflect, discuss and improve their skills. AI can be used to augment a clinician’s expertise and extend their capabilities in dealing with complex problems.

Founded on scientific research

In 2008, before the emergence of digital assistants like Siri and Alexa, a team of university researchers saw an impactful possibility. They wondered: with advancements in speech recognition and machine learning, could an AI system be built to automatically analyze therapy sessions and provide feedback to clinicians? Back then, the concept sounded too far-fetched. The initial grant for a clinical study was denied and the team told the premise “Sounded like science fiction.” Thankfully, NIH was willing to take a chance on the idea. Since that time more than $14M in Federal funding has led to more than 50 peer-reviewed articles laying the scientific groundwork for AI-supported assessment and feedback to improve evidence-based therapy and patient outcomes.

The Lyssn founders are the academics and psychotherapists that did this ground-breaking research applying AI to psychotherapy. Through over a decade of clinical inquiry they have amassed more expertise than any other team in the world. The company was established on a strong commitment to science and continues to push the boundaries of technology to support and enhance training, supervision, and quality assurance of behavioral health care.

Empathy evaluation via AI

Empathic listening is a fundamental component of quality behavioral health. Empathy can be measured, but expert human evaluators are expensive and slow. Lyssn’s research was the first-ever to show that machine learning models can automatically and accurately estimate empathy, or a clinician’s level of understanding of their client with results comparable to human ratings of the same session.

Quantifying session data

Counseling sessions focus on relationship crises, job issues, suicide, addiction problems, and sometimes just the weather. Our research showed that machine learning can automatically capture and summarize the key content in therapeutic sessions, including psychiatric symptoms, problems, emotions, and general focus.

Scaling up fidelity

Cognitive Behavioral Therapy is the most well studied form of psychotherapy, but how often is CBT used in community settings and with what quality? We simply don’t know. Using the latest generation of neural network models, Lyssn automatically estimates CBT competence that is indistinguishable from expert human evaluators.

Automating assessment

Developments in machine learning, particularly ‘deep learning’ methods have revolutionized many classification tasks, particularly automatic speech recognition. At Lyssn, we build on years of university-based research to adapt cutting edge techniques specifically for behavioral health counseling conversations.

Advancing AI technology

Current R&D

The unique AI technology behind Lyssn is used in therapeutic sessions to enhance, support, and evolve services. The scientifically developed and tested platform is in active use today in universities and clinics and by telehealth providers. We continue to improve and add to our product offerings and customized services. Find out more about our ongoing NIH research and products in development.

Academic papers

Collectively the leadership behind Lyssn have been conducting research on the use of AI in the assessment of behavioral health care delivery for decades. Below is a list of selected papers.