SonderMind, a hybrid mental health provider, has unveiled a suite of AI tools to save providers time, support the provider-patient relationship and engage patients between sessions.
Among the tools now available are features that streamline notetaking for providers and allow patients to set goals and log reflections. Others help patients and providers prep for sessions and support treatment planning for providers. The tools were co-developed with licensed therapists and released only after clinical review and workflow testing.
The suite of tools, all developed in-house, is governed by SonderMind’s AI Constitution, a clinical, privacy-first framework that guides development, deployment and oversight. The framework is approved by SonderMind’s AI Governance Council, made up of clinicians, compliance experts and product leaders. The tools will undergo annual bias and safety reviews.
“We're very focused on ensuring compliance with various state and federal regulatory changes happening with regard to AI,” Mark Frank, co-founder and CEO of SonderMind, told Fierce Healthcare. “We wanted to make sure we maintained the utmost levels of privacy and security, and we felt that doing that ourselves was one of the clearest ways to make sure we weren’t putting any of our patient data at risk.”
For years, SonderMind has been thinking deeply about how to enable a different kind of care model with tech. It already uses AI in its patient-provider matching algorithms.
“The right provider isn’t just, ‘Do they have availability, or are they licensed in my state, or do they take the insurance?’ The right provider means, ‘Do they actually treat the demographic I'm a part of, do they treat the issues that I'm having?’” Frank said.
In 2021, SonderMind acquired Qntfy, a predictive analytics platform that uses mental health biometric data to suggest potential treatment options. This helped SonderMind deliver personalized care using data without loading that extra work on providers. In 2024, SonderMind launched AI-driven journaling, available to anyone who downloads the SonderMind app. More recently, the company rolled out AI-assisted clinical notes for providers, including an ambient scribe. SonderMind has seen this improve provider note-completion time by up to 80%.
The company has wanted to expand patient engagement with its digital tools. The data show that 77% of patients who engage with SonderMind digital tools sustain therapy improvements, as measured by patient-reported outcome questionnaires, for two years or longer. And those who engage with its AI tools see an 85% reduction in symptoms across the therapy journey, the company said.
Those outcomes are of interest to its health system partners interested in managing the total cost of care, Frank said. Today, SonderMind has at least 74 digital tools in its app. They include games, neuro-tunes, resonant breathing exercises and more. The benefit of having an AI-driven tool is to capture the complexity of the mental health space.
“In mental health, you walk into your session and you go, ‘I don’t really know what’s wrong, I just know I don’t feel good,’” Frank noted. That’s very different from having a standardized protocol for how to treat a broken arm at an urgent care. AI can make connections for both patients and providers that otherwise might be missed, per Frank.
With features like journaling or goal setting, AI can also ingest a massive amount of data and then help providers tailor their treatment plans.
“You’re starting with so much more information at your fingertips as a provider,” Frank said. While providers are not obligated to accept any AI recommendations, “it just empowers them and makes them that much better.”
SonderMind plans to continue expanding its suite of AI tools. It is exploring possibly introducing a chatbot. SonderMind providers regularly practice measurement-based care, with 80% of sessions using patient-reported outcome questionnaires to track progress, per Frank: “We try to get as much information about progress being made as possible.” This can be supported by AI in the future, he suggested.
AI may be able to discern progress in a more qualitative way, for instance, from a chat than via a more rigid questionnaire. Ultimately, there is room for “material” improvement to patient-reported outcome questionnaires, in Frank’s view. The PHQ-9, for depression, was developed in the 1990s with the support of an educational grant from pharma company Pfizer. “We can come up with better things that help us understand how people are progressing in their depression treatment,” Frank said.
Regardless of what’s next, “our point of view is that these tools need to contain a provider in the loop, they need to have some level of clinical engagement,” Frank stressed.