Microsoft and Epic put generative AI into hospital workflows

If you have visited a clinic recently, you have probably noticed your doctor spending more time typing on a keyboard than actually looking at you. Administrative burden has pushed 40% to 60% of clinicians into burnout, and HHS predicts a shortage of 90,000 physicians in the US by 2025.

Microsoft and electronic health record giant Epic announced a massive expansion of their strategic AI partnership to tackle this paperwork crisis. Unveiled at Epic’s annual Users Group Meeting, the initiative embeds dozens of generative AI copilot tools across Epic’s healthcare ecosystem, powered by Azure OpenAI Service and Nuance.

AI listens in the exam room

The centerpiece of the rollout is bringing Nuance Dragon Ambient eXperience Express directly into Epic’s native software. DAX Express is now integrated into the Epic Hyperdrive platform and the Haiku mobile app, allowing physicians to securely generate clinical notes automatically from ambient exam room conversations.

Nuance was also named one of the first vendors in Epic’s new Partners and Pals program, cementing the tight integration between both platforms.

Beyond direct patient notes, Microsoft and Epic are targeting the broader administrative back office. New note summarization tools are expanding inside the Epic In Basket, offering suggested text and context summaries to speed up reviews for doctors and nurses.

On the business side, generative AI is being deployed into revenue cycle management. The system analyzes clinical notes to offer medical coding suggestions, aiming to streamline billing accuracy in an industry where McKinsey reports administrative costs swallow nearly a quarter of national healthcare spending.

Epic is also hooking Azure OpenAI into SlicerDicer, its self-service reporting platform, letting early testers query clinical data with natural language to study rare diseases and fill evidence gaps.

Why it matters

For overworked clinicians, this could be the single most practical application of generative AI today. Reducing hours spent charting after hours is an immediate quality-of-life win, and if medical coding becomes automated, health systems can claw back massive operational expenses.

We think Microsoft's strategy here is textbook execution. Instead of trying to convince hospitals to adopt a standalone AI tool, Microsoft is injecting its tech straight into Epic, the dominant operating system of modern healthcare. That eliminates friction for thousands of physicians who already live inside Epic all day.

However, clinical workflows leave zero margin for error. While ambient listening sounds magical, AI hallucinations in dosage instructions or allergy histories carry life-or-death consequences. Clinicians will still need to manually verify every generated summary, meaning time savings will depend entirely on how accurate DAX Express proves in noisy, real-world clinics.

Still, recent data from UPMC and KLAS Research shows healthcare executives are prioritizing operational optimization above all else for AI spending over the next two years. Microsoft and Epic just positioned themselves as the default standard for that shift.