M42 launches Ossi AI platform for clinical workflows in Abu Dhabi

UAE-based healthcare company M42 introduced Ossi, a multi-agent artificial-intelligence platform for clinical workflows, at the Abu Dhabi Future Health Summit on 22 October. Information reported by Emirates News Agency from the company’s announcement said the platform is intended to reduce documentation work and give clinicians more time with patients. It is built on Microsoft Azure’s cloud and AI infrastructure.

Ossi is launching with four agents, including an Ambient Scribe Agent that transcribes clinical conversations and prepares draft notes for clinician review, and a Health Record Agent that retrieves and summarises relevant information from electronic medical records. M42 said the system is intended to bring together clinical context, prior patient information and evidence-based guidelines through secure ambient intelligence.

The platform has been activated at seven healthcare facilities in the UAE and Bahrain: Amana UAE, Amana Bahrain, Healthpoint, HealthPlus, Mubadala Health Dubai, Sheikh Sultan bin Zayed Hospital and Imperial College London Diabetes Centre in Abu Dhabi. The list defines the initial deployment, but the announcement did not say how many clinicians were using the tool, how frequently it was used or how much time it saved.

M42 said data security and patient privacy were considered in the platform’s design. The release describes a product intended to support clinician focus and evidence-based workflows; it does not present a study validating patient outcomes. Nor does it say that Ossi independently makes medical decisions: the first agents assist with documentation and information retrieval.

The four-agent launch illustrates that the platform is divided into assistive tasks rather than one general clinical decision-maker. The Ambient Scribe prepares documentation for a clinician to review; the Health Record Agent assembles relevant chart context; Clinical Q&A helps retrieve clinical information and prepare patient-facing explanations; and the Coding Agent suggests codes for review by appropriate teams. M42 says the platform is model-agnostic and multilingual, and can connect with existing records through HL7, FHIR or custom APIs. Those interoperability options describe supported integration routes, not proof that every facility’s record system is already connected in the same way.

The company says Ossi is designed to meet regional health-data governance requirements and emphasizes privacy and security. The launch announcement does not publish an independent security audit, named model providers, language coverage, error rates, clinician uptake or measured documentation time. It also does not report clinical outcomes. The seven facilities identify deployment footprint; they do not establish how many departments or users are active or whether the agents are used for every encounter. Those distinctions matter when evaluating a deployment that sits inside clinical workflows and handles sensitive health records.