Abridge tackles a daily task for healthcare professionals: documenting consultations. This American artificial intelligence company develops an assistant capable of turning a medical conversation into a draft clinical note. Presented on abridge.com, its technology aims to reduce the time spent entering data and preserve the attention given to the patient.
A company born out of medical practice
Founded in 2018, Abridge has combined medical and technical expertise from the outset. Its co-founders are Shiv Rao, a cardiologist, Zack Lipton, a machine learning researcher, and Sandeep Konam, an engineer specializing in artificial intelligence. The company initially focused on the difficulty patients face in remembering and understanding the information exchanged during a medical appointment.
This approach subsequently led to an offering for healthcare professionals and healthcare organizations. The problem still revolves around the same raw material, conversation, but the output changes: the aim is now to produce usable clinical documentation. Abridge has notably established collaborations with American healthcare organizations such as UPMC and Yale New Haven Health.
From conversation to medical record
The solution belongs to the category of so-called “ambient” clinical documentation assistants. It processes spoken exchanges during a consultation to propose a note organized around the professional’s needs. It therefore goes beyond word-for-word transcription: it selects and rephrases the elements relevant to medical documentation. The healthcare professional must then review, correct if necessary, and validate the result.
Abridge also highlights the ability to link elements of the note to the passages of the conversation from which they originate. This traceability makes it easier to verify the generated text, without in itself guaranteeing accuracy. Integration with electronic health record software, particularly the Epic environment, is another important aspect of its offering: the operational value depends largely on the ability to avoid duplicate data entry and switching between interfaces.
The company thus positions itself in documentation rather than autonomous diagnosis. For its customers, evaluation focuses as much on the quality of the notes as on the time actually saved, adoption by teams, and compatibility with their internal procedures.
What next?
Abridge’s growth will depend on its ability to deploy its tool at scale without compromising the reliability of clinical notes. Accents, consultations involving multiple speakers, medical specialties, and ambiguous wording all present situations that must be handled effectively. Competition among medical assistants also requires demonstrating lasting benefits beyond initial trials.
The confidentiality of conversations, keeping patients informed, and data governance will remain crucial. For healthcare organizations, the challenge will be to measure gains without shifting the workload toward systematic correction of notes. The future of this category of tools will therefore depend as much on integration into healthcare practices as on model performance.