Limbic, a clinical mental health AI company, has published a randomised, double-blind study in Nature Medicine showing its AI therapy system outperforms human therapists in delivering cognitive behavioural therapy. The system combines leading large language models with Limbic's clinical reasoning layer. The study evaluated 227 participants across licensed therapists, standalone LLMs and LLMs with Limbic's layer. Results showed 74.3% of AI-powered sessions scored higher than the top 10% of human sessions, with AI agents scoring 43% higher than standalone LLMs on the Cognitive Therapy Rating Scale. Clinicians preferred the enhanced AI agents 82.7% of the time. A real-world analysis of 19,674 therapy transcripts showed users with highest exposure to Limbic's system achieved a 51.7% recovery rate, compared to 32.8% for lower exposure groups.

NHS Confederation and Limbic explore AI use in mental health. Rebecca Gray, director of the NHS Confederation's Mental Health Network (Credit: NHS Confederation) * The NHS Confederation and Limbic have partnered to explore how AI can improve mental health services * The project will identify the barriers and opportunities facing NHS systems related to AI use and long-term adoption in mental health care * Limbic was announced as the first-ever winner in the scale-up category of Digital Health Rewired Pitchfest in 2023 Former Pitchfest winner Limbic has partnered with the NHS Confederation to explore how AI can improve mental health services and access to care. Limbic was announced as the first-ever winner in the scale-up category of Digital Health Rewired Pitchfest in 2023 for its AI mental health chatbot. The NHS Confed Mental Health Network will work with Limbic to identify the barriers and opportunities facing NHS systems related to AI use and long-term adoption in mental health care. Rebecca Gray, director of the NHS Confederation's Mental Health Network, said: "The 10 year health plan sets out a vision for a health service that embraces digital and AI to improve access and the effectiveness of care. "The rise of AI-enabled tools presents both opportunities and challenges. Issues of safety, effectiveness, accessibility and a lack of guidance mean that there is a real need for support in this area. "Digital Health is delighted to be working with Limbic, whose wealth of experience and expertise will help to ensure that science and evidence underpin the adoption of AI in mental health services. "We need to give members the confidence to make informed decisions focused on the biggest areas of need, sharing practical examples of what works, and ensuring that AI strengthens services rather than adding unnecessary complexity." The project will also provide practical guidance on AI adoption, supported by case studies and examples of use in mental health, and support systems looking to expand their use of AI tools and create an environment in which they can thrive. Ross Harper, chief executive and founder of Limbic, said: "Digital Health know that AI has enormous potential to scale mental healthcare, but realising that potential requires collaboration, evidence, and a problem-first approach. "Our partnership with the NHS Confederation positions us at the heart of this critical conversation, helping shape how clinical AI can genuinely improve outcomes for patients and clinicians alike." Kathleen Henrick, UK general manager at Limbic, said: "Mental health services are under enormous pressure, and AI offers tangible solutions, but only when implemented safely and effectively. "Digital Health is delighted to partner with the NHS Confederation's Mental Health Network to ensure mental health leaders have the practical guidance and evidence they need to deploy clinical AI confidently and responsibly. "By sharing its experience from deployments across the NHS and learning from mental health leaders, Digital Health can ensure AI adoption is guided by clinical need and patient safety. "Together, we can move beyond the hype and focus on regulated clinical AI solutions that deliver measurable value across the care pathway."