If you ask Microsoft general manager Kenn Harper whether Microsoft wants to be the iPhone of health care scribes, he'll politely decline to repeat the analogy — but he doesn't think it's inappropriate. Why it matters: As Microsoft's AI medical scribe capabilities face increased competition from startups, Harper says its strength is in openness — not unlike the way the iPhone enables a sea of apps to function on top of it.
Context: Microsoft in 2022 acquired voice recognition and AI software company Nuance for $19.7 billion. Dragon Copilot is the name of the combined companies' AI clinical documentation tool.
- More than 600 health systems are using its ambient tool, Harper says.
This Axios Expert Voices interview has been edited for length and clarity.
Who are Microsoft's biggest competitors in the ambient clinical space?
- There are some 30 competitors [including] Abridge and Suki.
Given the pace at which these startups are innovating, would Microsoft see value in acquiring a company like an Abridge?
- I don't think acquisition is as key to [our] strategy relative to what it might have been in previous technology eras, when everyone has access to technology and is innovating.
How is Microsoft positioning itself to stay competitive in the landscape, particularly with agile newcomers that have integrated into major EHR systems?
- It goes back to this ecosystem that opens up scale — not by acquiring companies per se — but letting them plug in.
- Microsoft doesn't have to build everything directly or think of every idea directly. It can let others come in and innovate on top.
- This is a huge opportunity. It's going to drive innovation at scale that I don't think we've ever seen before in health care [with] any third party that's building a unique capability —could be clinical decision support, tooling, coding, neurological [disease] detection or vocal biomarker screenings. What those companies are missing is the data of what's actually unfolding inside of an encounter between physician and patient.
- That's what's going to separate Microsoft from competitors.
What care environment uses Microsoft's ambient medical scribe tools the most?
- The lion's share today is primary care. Specialists are quickly catching up.
- The emergency department (ED) is also growing, and in-patient is a new area we just introduced.
What challenges does the technology face moving from primary care to specialty care, and how do you adapt it to suit?
- We have a very large clinical team at Microsoft. This is their job. They look at every specialty, the unique aspects of what needs to be documented, templates that need to be pulled in, unique workflows specialists have, and how the ambient tool can be used depending on how you practice and run from room to room, patient to patient.
Can you give me an example?
- One of the things that's unique about the ED is the physician might see the patient multiple times — what's called a course — so we've modified our workflow and documentation where every time you check on the patient, you want to log a timestamp and an additional course of assessment and plan.
When it comes to enhancing Microsoft's AI scribe capabilities, how does it think about its relationships with health systems, as customers or partners in innovation?
- Most customers fall under the partnership bucket. It's not just "We're selling you software." It's a constant, "Hey, how is it being used? Is it driving value?"
- There isn't a single meeting where a customer doesn't have a very strong opinion about what we should be doing next. This is a lively conversation about "when are you going to have diagnosis management? When are you going to have document specificity that will help us with some of our downstream coding initiatives?"
What's on your product road map?
- There's a lot we're going to be doing in revenue cycle management [RCM] to make it more proactive and automated to help optimize for downstream billing.
- We look at what's in the EMR. We pull that data out, we look at what the current visit is, and we can actually make some suggestions automatically based on context, where all you have to do is press the button and something changes in your documentation to make sure we're optimizing for reimbursement. That will be a big focus of ours.
Since you were at Nuance, I have a question for you. Why was it called Nuance?
- There's so much nuance in conversation. It was a reference to speech and the uniqueness of understanding speech.
How does Microsoft envision the role of AI in transforming clinical documentation and overall health care delivery going forward?
- It's not just about documentation anymore. This is about expanding into as much of the administrative work that we can help semi-automate for a clinician and save time.
