AI in Healthcare & Life Sciences

Where AI in healthcare and life sciences actually stands

Healthcare AI has two visible commercial surfaces right now, and neither of them is the diagnosis. The first is the back office — the billing, coding and paperwork layer that sits between a clinician and getting paid. Commure, a $7 billion startup, is built around automating exactly that, and STAT's reporting from inside the company describes the messy realities of the effort. The second is the patient-facing portal: Oracle Health's AI-backed patient portal is generally available in the US, built on OpenAI foundation models, integrated with Oracle's EHR, and offering record review, provider engagement, appointment scheduling and AI-generated explanations of medical jargon.

The shift worth tracking is not the capability but the delivery shape. Oracle's portal is not a pilot a health system opts into; it is a generally available feature of the EHR that system already runs. For a hospital on that stack, the live question is no longer whether to buy patient-facing generative AI — it is who reviews what the model tells patients, and whether the existing contract already includes it. That turns a procurement decision into a clinical-governance one, and hands it to organisations that were not necessarily staffing for it. Inference

The two stories read against each other usefully. Administrative automation is where the money is concentrating, and it is also where failure is least visible: a mis-coded claim or a dropped scheduling request does not surface as a clinical error, it surfaces as friction in a revenue-cycle report months later. Inference The STAT piece is the closest thing available to a look under the hood of that business, and it is paywalled — a fair indicator of how much of this sector's operational detail is publicly checkable at all. Inference

It was a thin week for the sector, and it is worth saying so rather than padding it. Nothing in the week's coverage touched diagnosis, clinical decision support, drug discovery or trials; no regulator acted, no outcomes data landed, and no vendor made an accuracy claim that needed checking. What did land was two developments about the administrative envelope around care rather than care itself — which, on this week's evidence, is where the sector's AI story actually is.