CEO & Co-Founder of Mentaily Shares Insights From Sheba’s LIV
From Global Gap to Sheba-Born Solution
Shtein opened the session by addressing the scale of the challenges facing mental health care in Israel. In her words, even before the war, “last in terms of, in comparison to the OECD countries, in terms of number of psychiatrists, number of social workers, number of psychologists, and last in terms of how much time people need to wait to meet a professional in behavioral health.” Without intervention, the average patient faced wait times of up to two years for a first appointment. She traced the problem back to the COVID-19 pandemic, when a surge in symptoms of mental distress overwhelmed a telemedicine system that could increase virtual availability, but not the number of clinicians. The bottleneck, she noted, was never the channel of care but the supply of trained professionals to deliver it, a crisis not limited to the Israeli context. Shtein went on to describe how LIV grew out of her work at Sheba’s ARC Innovation, where she led the hospital’s telemedicine hub alongside Dr. Asaf Caspi, now Director of the Department of Psychiatry at Sheba. After October 7th, anticipating a surge in trauma-related need, she approached Microsoft to build a solution together with Sheba and KPMG, with the explicit goal of creating a clinical tool rather than a commercial product. She noted that Mentaily later spun out of Sheba as an independent company in April 2025, while retaining close ties to the hospital. As Shtein stated, “The fact that we were created inside of a hospital, through a collaboration, through a development of a real solution for a real need that was from the field, is part of our DNA.”How LIV Works, Validated by the Data
LIV conducts a full psychiatric assessment at home, by voice or text, generating a clinical summary of diagnosis, history, and recommendations for the clinician ahead of the visit. This functionality positions it as a decision support tool, not a replacement, cutting intake time from up to two hours to roughly 30 minutes while keeping the full transcript and reasoning transparent to the clinician.
Shtein illustrated this for the webinar participants with data from two studies: a blind comparison showing LIV matched psychiatrist diagnostic accuracy in over 94% of cases, and a second study with the Israeli DARPA and Ministry of Defense of first responders, finding over 90% accuracy for PTSD and over 96% for risk detection. She noted the model is tuned toward sensitivity, and that some patients, particularly trauma survivors and first responders, reported greater comfort with the AI than with a human clinician.
Deployment at Scale
LIV can be used for four core deployment purposes: population-wide screening, such as assessing evacuated populations for PTSD, initial triage to determine urgency and the right level of care, full clinical intake based on DSM-5, and ongoing monitoring between sessions to flag deterioration.
The platform’s current reach stands as evidence of that scale in practice. LIV holds commercial agreements with Israel’s Ministry of Defense and National Social Security, and is deployed with the VA and Walter Reed Military Hospital in the US and Sunnybrook in Canada. Shtein noted the platform received the first Israeli regulatory approval for AI in behavioral health, maintains HIPAA and ISO compliance, and was named by Microsoft among its top 10 AI for Good initiatives.
Answering Real Clinical Needs
Shtein closed the briefing by returning to LIV’s origins. Built inside a hospital rather than as a separate startup, the platform emerged from a specific clinical need identified by clinicians themselves, first during COVID, then accelerated by the demands of war. She framed that origin as core to its credibility: a tool developed with psychiatrists, tested against psychiatrists, and refined through real-world deployment rather than commercial ambition.
For mental health care systems facing similar gaps between demand for treatment and clinical capacity, Shtein positioned LIV as evidence that AI can extend the reach of scarce clinical expertise without displacing the clinician’s judgment. This model is already part of clinical workflows at Sheba Medical Center, but its relevance and applicability range far beyond.


