For enterprise IT readers, the interest lies in that combination: an AI interface embedded in an existing information system, backed by regional cloud infrastructure and operating within established access permissions. But there is an important wrinkle in the launch narrative: Malaffi’s own documentation described an AI assistant before this announcement.
What the Malaffi assistant does
WAM reports that the assistant retrieves relevant patient-record information, produces concise health-history summaries and answers physicians’ queries. Malaffi holds information for more than 14 million patients and over four billion clinical records, including clinical documents, allergies, laboratory results, vital signs, radiology images and diagnostic results.
That describes a substantial information-retrieval task. However, the breadth of Malaffi’s database should not be mistaken for proof that the assistant interprets every data type. The announcement lists radiology images among the exchange’s holdings; it does not establish that this integration performs AI image analysis. Nor does it specify how records are ranked or how summary completeness is evaluated.
The distinction matters: finding useful information and proving that nothing important was missed are different challenges.
An earlier assistant complicates the “new” label
Malaffi’s official physician-assistant FAQ is dated November 2025. It already describes an assistant supporting custom questions and predefined queries about information such as medications, diagnoses and vital signs.
The FAQ documents this access sequence:
- Open the Delta View tab.
- Select Physician LLM in the left-hand panel.
- Review the disclaimer.
- Select Agree to proceed.
Those are documented portal instructions, not confirmation that the newly announced Ossi integration has an identical interface.
The evidence therefore supports a narrower conclusion than “Malaffi gets AI for the first time.” The 10 November 2026 announcement identifies Ossi as the underlying platform and states emirate-wide physician availability. It does not explain whether Ossi replaces, expands or rebrands the previously documented assistant.
Where Microsoft Azure fits
M42 announced Ossi on 22 October 2026, describing a multi-agent clinical platform built on Microsoft Azure cloud and AI infrastructure. Its launch materials identify four agents covering ambient documentation, health-record retrieval, clinical questions and medical coding. M42 also says the platform is multilingual, model-agnostic and supports integration through HL7, FHIR or custom APIs.
These are characteristics of the broader Ossi platform. They should not be read as a checklist of features enabled inside Malaffi: the integration announcement specifically describes retrieval, summaries and query responses.
For IT teams assessing comparable deployments, the architectural lesson is to distinguish the platform’s possible capabilities from the functions actually deployed. A product brochure is not a production configuration.
Access controls remain part of the story
Malaffi’s provider-portal documentation adds useful operational context. It describes access for authorised users at connected healthcare facilities, with the portal integrated into providers’ electronic medical record systems.
Its Break the Privacy Seal mechanism restricts specified sensitive information to the highest level of authorised users, where clinicians judge access necessary for treatment. Malaffi says these activities are monitored and audited. These are documented portal safeguards, rather than a newly disclosed technical specification for Ossi.
The practical implication is straightforward: an AI query box should not be treated as permission to see everything behind it.
Availability is not yet evidence of effectiveness
WAM says clinical responsibility remains with physicians and that usage results and physician feedback will inform further development. The announcement supplies no measured time savings, accuracy results or evaluation timetable.
That leaves a clear dividing line between the deployment and its promised benefits. Abu Dhabi has announced an Azure-hosted assistant inside a large connected health-data environment. Whether it reliably reduces searching without losing important context is the next question the evaluation needs to answer—not a result that availability alone can establish.