Skip to main content

Corti, a Danish AI startup, has secured $60m in a Series B funding round, thereby creating one of Denmark’s most substantial financial rounds this year. The raise, spearheaded by Atomico and Prosus Ventures, and buoyed by existing investors such as French VC Eurazeo, Danish export fund EIFO, and Danish VC Chr. Augustinus Fabrikker, is set to catalyse Corti’s global expansion, particularly the US.

The essence of Corti’s innovation is encapsulated in its Generative AI tool, designed to assist healthcare professionals in their decision-making processes. Amidst the palpable tension of time constraints and staff shortages, GenAI is a tool that augments human conversations in healthcare, analysing patterns within a vast corpus of existing data to generate new outputs, specifically, diagnoses. Initially, Corti trained its large language model (LLM) using recordings of emergency medical phone calls related to cardiac arrest, gradually expanding its technological reach to be utilised more broadly by hospitals and other healthcare institutions across Scandinavia, the UK, and the US, impacting about 100m patient encounters per year, or a staggering 150k a day.

Imagine a scenario where a healthcare professional is engaged in a patient interaction, either through a call or in person. GenAI, acting as an AI “co-pilot”, comprehends the dialogue and information shared in real-time via the user’s computer, without necessitating additional hardware since the AI and the data collected are securely stored in a dedicated cloud online. As the interaction unfolds, GenAI gently nudges the professional with suggestions on follow-up questions and advice, akin to a second opinion, while also autonomously managing journalling and quality assurance. Once a clinical note is complete, the AI scrutinises the text to determine the most apt diagnosis and procedure codes, thereby streamlining the process and enhancing accuracy.

Lars Maaløe, an associate professor in machine learning and a co-founder of Corti, elucidates that their tech stack transcends just an LLM. It encompasses understanding the conceptual level of what’s said within the audio, language comprehension, natural language processing, and clinical predictors, with the LLM being merely one facet of it. However, the journey of integrating GenAI into healthcare is not without its challenges. One such challenge is the occurrence of hallucinations, instances where the AI generates outputs that do not accurately reflect data inputs, such as providing a factually inaccurate response. To circumvent this, Corti has developed software that ensures predictions adhere to protocols, guidelines, and textbooks, with healthcare providers maintaining control over the guidelines that the AI must adhere to.

Corti’s co-pilot is not confined to a single geography but is utilised in a myriad of hospitals across various locations. Since the AI is housed in a dedicated cloud, it can assimilate learnings from patients encountering similar issues across different geographies. Maaløe shares an insightful observation that there isn’t a substantial difference between patients in Denmark, Sweden, the UK, or the US. This enables Corti to optimise customer systems by suggesting a treatment that has proven effective on a patient in, say, London, to a healthcare professional in Denmark, when they encounter a patient with similar conditions. From a big data perspective, Corti thus has a profound impact, prompting a reflection on whether it’s necessary to have such varied procedures for the same conditions and instead, have systems that learn from one another.

According to Corti, its technology has been substantiated to enable healthcare workers to achieve up to 40% more accuracy in outcome predictions and be 90% faster in managing administrative tasks. In a world where healthcare is perpetually evolving, and the demand for precision and efficiency is paramount, Corti’s GenAI stands as a testament to the transformative power of AI, not just as a tool but as a co-pilot, navigating the intricate pathways of healthcare conversations and diagnoses, ensuring that every word spoken is not just heard, but understood, analysed, and acted upon with unparalleled precision and care.

Kevin McDonnell

Author Kevin McDonnell

Helping ambitious HealthTech, MedTech, Health and Technology leaders shape the future of healthcare.

More posts by Kevin McDonnell