Last updated: 11 November 2026, 09:15 GMT.
In short
- Google Cloud announced Gemini for healthcare on 10 November, initially in preview.
- It is designed to coordinate clinical and administrative work across existing systems.
- Google says its Epic and Oracle Health connections read records without writing back to those EHR systems.
- The launch does not establish deployment in Brighton’s NHS services or independently proven improvements in care.
The Gemini for healthcare preview was announced by Google Cloud on 10 November 2026, bringing a new healthcare-specific AI workflow product to organisations using clinical, payment and administrative systems. The immediate development is a product preview, with named launch customers largely in the United States.
Gemini for healthcare preview: what is new
In its original announcement, Google describes a system that brings information from health records, billing platforms and clinical guidance into coordinated workflows. Tasks include preparing records before appointments and assembling information for coverage reviews.
Google names launch customers including HCA Healthcare, Northwell Health and Stanford School of Medicine. It says direct connections to Epic and Oracle Health use read access, without writing changes back to the underlying electronic health record. A NextGen connection is still in development.
The company’s technical launch explanation sets out access controls inherited from existing systems, separate agent identities, audit logs and citations that reviewers can inspect. These are Google’s descriptions of the product, rather than findings from an independent assessment.
Why the read-only distinction matters
Reading a record and altering the record of care are different powers. A read-only EHR connector limits what that particular connection can change. It does not, on its own, demonstrate that every answer is accurate, every source is complete or every action elsewhere in a wider workflow is harmless.
That distinction is useful for Sussex readers following healthcare AI. A product launch should prompt questions about exactly what has been connected, whose permission applies and which person checks the result. It should not be treated as evidence that a local service has bought or adopted the product.
What is confirmed and what is still uncertain
The preview and the stated connector design are confirmed by Google’s launch materials. Claims about easing pressure and freeing staff time remain claims by the supplier in this report. The announcement is not an independent clinical trial, and it does not provide evidence here of improved outcomes for Brighton patients.
For wider context, our report on the WHO–OECD TArGET Health tool explains a separate framework for assessing emerging health technology. That framework is not an endorsement of this Google product.
Brighton Locals’ view
Our assessment: connecting evidence and retaining a clear audit trail are sensible priorities. The meaningful test will be whether real deployments show reliable outputs and useful time savings after staff checking and corrections are included. Preview availability is the start of that assessment, not its conclusion.
What organisations should do next
Healthcare teams considering the preview should read the launch explanation and ask for the scope of available connectors, access restrictions, review responsibilities and evaluation evidence relevant to their intended use. Patients should continue using their existing care and booking channels; this announcement does not create a new local treatment or appointments service.
Watch for wider availability, the promised additional EHR connection and published evaluations that distinguish administrative efficiency from clinical benefit.
Google has announced a more connected healthcare workflow product. Whether it makes care better depends on evidence from its use.
Image: Google headquarters sign, 2 April 2014, archive photograph by Anthony Quintano, CC BY 2.0. This depicts the company, not a hospital or a patient system.
Leave a Reply