Inpatient Care Drives AI Scribes
Brendan Keeler, interoperability lead at HTD Health, on GTM for AI medical scribes
The biggest upside in AI medical scribes now sits inside the hospital, not the clinic, because that is where documentation is most fragmented and where shaving even a few clicks or minutes off the workflow changes care delivery at scale. Outpatient visits are repetitive and easier to model. Inpatient care spans the ER, surgery, nursing rounds, transfers, orders, and specialty specific note types, so the product has to handle more chaotic audio, more handoffs, and much deeper EHR actions to matter.
-
The technical problem is not just better transcription. In a clinic visit, the workflow is usually check in, conversation, note, bill. In the hospital, the system must understand many entry points, multiple providers, department transfers, and very different documentation patterns across settings like the ER, ICU, surgery, and inpatient rounds.
-
The near term beachhead is controlled inpatient sub workflows, not the whole hospital at once. Nursing rounds on admitted patients are relatively structured, which is why Abridge has started there. That fits the broader pattern that leading vendors first win ambulatory care, then expand into adjacent environments that look similar enough to reuse the product stack.
-
This is also why Epic aligned players have an edge. Hospital deployments usually require deep integration, enterprise procurement, and security review, which favors vendors already tied into major EHR workflows. Abridge built around Epic and reached an estimated $100M ARR by May 2025, while PLG players like Freed grew fast in outpatient settings but from a smaller base and with lighter integrations.
The next phase of the market is a move from ambient note capture to workflow automation in high acuity settings. The winners will be the companies that can turn messy hospital conversations into finished notes, orders, and structured data inside the EHR, one unit and one role at a time, starting with the most repeatable inpatient workflows and then expanding outward.