Corti Introduces FactsR™, AI Clinical Reasoning Layer for Healthcare
Corti, a leader in building clinical-grade AI for healthcare, has announced the launch of FactsR™, a groundbreaking real-time agentic reasoning system designed for clinical consultations. This innovative system addresses the common issue of AI-generated "note bloat" by a significant 65 percent and aims to minimize the need for post-visit edits, thereby transforming passive documentation processes into active clinical intelligence. FactsR™ is engineered with ambient documentation in mind, setting a new standard for ambient AI in the healthcare sector and paving the way for real-time decision support at the point of care.
Unlike conventional LLM pipelines that are often retrofitted for healthcare applications and tend to produce verbose, error-prone summaries, FactsR™ is built upon Corti's proprietary recursive fact-first reasoning loop. This purpose-built engine is designed to meticulously surface, validate, and structure clinical knowledge in real time as conversations unfold during consultations. Delivered as a modular API, FactsR™ empowers developers to embed this clinical-grade intelligence directly into their healthcare applications, fostering safer, leaner, and more trusted AI experiences for clinicians.
The significance of this recursive approach lies in its foundational shift from passive summarization to active reasoning. Traditional ambient solutions typically process raw transcripts through generic models after a consultation has concluded, resulting in lengthy summaries that clinicians can spend considerable time—up to three hours a week—correcting. FactsR™, conversely, enables developers to build AI systems that deliver more concise and accurate results, which clinicians can interact with live, during and after consultations, ensuring relevance and precision.
The FactsR™ system operates through four key stages. First, it 'Listens and Extracts in Real Time,' continuously identifying and surfacing structured clinical "facts" such as symptoms, vital signs, medications, and social history as the conversation happens. Second, each extracted fact is 'Vetted and Refined with Specialized AI' through an AI-driven feedback loop, ensuring clarity, accuracy, and consistency. Third, the 'Clinician-in-the-Loop' design allows medical professionals to quickly review, accept, or adjust these facts, maintaining control and ensuring AI augments their judgment. Finally, once facts are finalized, the system 'Generates EHR-Ready Notes,' assembling a clean, concise summary free from irrelevant content.
The benefits for clinicians are substantial, including significantly less screen time spent on corrections, allowing for better patient focus as decisions are made within the consultation rather than in hindsight. Furthermore, FactsR™ provides audit-ready transparency, with every fact carrying a timestamp, a confidence score, and a direct link back to the relevant part of the conversation. Evaluation results on the public benchmark Primock57 dataset demonstrate that FactsR™ increases clinical completeness by 13 percent, capturing more relevant medical information compared to traditional ambient scribes, while drastically reducing note bloat by over 65 percent with clinician involvement.
Early adopters have praised the system's ability to accurately capture clinical details even in natural conversation. A beta user from an IT team in a regional Danish hospital highlighted the shift towards structured, recursive fact extraction, noting, "It goes beyond basic transcription to surface the right clinical facts in real time - exactly what busy clinicians need to stay focused and cut down on documentation overload." This sentiment underscores the practical advantages of FactsR™ in demanding clinical environments.
Lars Maaløe, CTO and co-founder of Corti.ai, emphasized the company's core philosophy: "healthcare AI must be purpose-built, real-time, and accountable." He added, "By breaking conversations into structured clinical facts and validating each one through recursive reasoning, FactsR™ elevates ambient documentation into a foundation for real-time clinical intelligence. When AI can listen, understand, and reason with medical context, it becomes more than a scribe - it becomes a trusted collaborator. With a developer-friendly API, we're enabling any healthtech company to embed this capability directly into their applications - safely, scalably, and in minutes."
FactsR™ is available today through a consumption-based API that is compliant with enterprise-grade HIPAA and GDPR standards. Developers can begin exploring its capabilities for free at docs.corti.ai and can access SDKs for safe and effective application building, often in as little as 30 minutes. A self-hosted option, suitable for on-premise or sovereign-cloud deployments, is slated to enter a limited preview phase this summer.
Corti.ai, the company behind FactsR™, is a research and development firm dedicated to building state-of-the-art AI foundation models specifically for the healthcare industry. Its mission is to eliminate administrative burdens and extend expert-level reasoning capabilities globally, thereby driving down healthcare costs and enhancing the quality of care. Corti's advanced models are designed to integrate seamlessly into any healthcare application via SDKs and APIs, enabling vendors, providers, and payers to leverage safe, cutting-edge AI across a wide array of clinical use cases.