Skip to content
    Back to writing
    August 8, 2024 · updated May 8, 2026 · 2 min read

    Ambient scribes were never going to be a product category.

    Ambient scribes were never going to be a product category — by Thomas Jankowski, aided by AI
    Always a feature, not a product— TJ x AI

    The JAMA study landed in early August 2024 with the numbers everyone in the ambient-scribe category had been hoping for. AI scribes reduce physician EHR time by 13.4 minutes per day. Burnout drops from 51.9% to 38.8% in 30 days. The trade press wrote it up as the legitimacy milestone the category had been waiting for. The reading was: ambient scribes are validated, and the standalone-scribe-vendor path is now the obvious operator path.

    The structural read is the inverse. Ambient scribes were never going to be a product category. They were always going to be a feature.

    The reasoning is operating-economics, not product-quality. The standalone ambient-scribe vendor sells a single workflow tool that integrates into the EHR and produces draft documentation. The EHR vendor sells the EHR, the workflow infrastructure, the billing-codes layer, the patient-portal, and a long tail of clinical-decision-support modules. By 2025-2026 the EHR vendor is also shipping a clinical-AI surface that includes ambient scribing as one feature among many. The standalone vendor's product is, at that point, a feature competing with a feature inside a platform the customer already runs.

    The JAMA study accelerated the absorption rather than extended the standalone runway. The trade press read the study as proof that ambient scribes are a real category. Epic and Cerner read the study as proof that ambient scribes are a feature worth building into the EHR-native clinical-AI product they were already shipping in 2024. The standalone vendors are now competing against the EHR-vendor build cycle, and the EHR vendors have a structural advantage the standalone vendors cannot offset: the EHR vendor's distribution is the customer relationship the standalone vendor is trying to enter.

    This is, of course, the feature-vs-product collapse pattern that has played out repeatedly in adjacent categories. Mailchimp built a category around email marketing; HubSpot absorbed the workflow into a marketing-automation platform; the standalone email-marketing-vendor category is, by 2024, much smaller than it was. The ambient-scribe category in healthcare is on the same curve, on a faster timeline, because the EHR vendors have a regulatory-and-distribution moat that mailchimp's competitors did not have.

    The structural read for the standalone-scribe vendors in 2024 is: the company you are building is, structurally, an acquisition target by Epic or Cerner or one of the cloud-AI-platforms. The exit window is the next 18-24 months, while the EHR vendors are still building their native version. After that window the acquirer's price is set against the build-vs-buy comparison, and the build option starts winning.

    The structural read for the health systems is: do not sign a multi-year ambient-scribe contract with a standalone vendor. The EHR-native version is shipping in 2025-2026, the migration path from standalone to EHR-native is going to be ugly, and the contract terms with the standalone vendor are not going to favor the buyer in that migration. Sign one-year terms with renewal options. Wait for the EHR-native version. Pay the productivity cost of the gap year if necessary.

    The structural lesson, of course, is the load-bearing one. A category that ships a single load-bearing workflow inside a larger platform is not a category. It is a feature. The platform absorbs it. The operators who recognize this in 2024 are the operators who time their bets correctly. The trade press will, on the same curve, write up the absorption as a series of "Epic announces ambient scribing" stories without recognizing that the announcement was the predictable endpoint of the category's structural shape from the beginning. The absorption is the story. The trade press will get there in 2027.

    —TJ