Google Cloud has launched Gemini for healthcare, an AI agent built for hospitals, medical groups and health plans. The company announced it on 10 November 2026, and it is available now in preview in the United States. It is the healthcare version of the Gemini agent Google introduced in October 2026, extended with healthcare skills, connectors to clinical and billing systems, and a governed platform.
What Gemini for Healthcare Does
Gemini for healthcare is aimed at the assembly work that sits around care. Patient records, coverage policies and clinical evidence are spread across disconnected EHRs, claims systems and administrative tools, and Google says clinicians spend nearly 28 hours a week manually reconciling that information. The agent is meant to pull the pieces together and carry multi-step tasks through to completion, with a person still overseeing the work.
For providers, it builds a patient's history across care settings, prepares the chart before a scheduled visit or procedure, and assembles authorisation packets against payer policy before submission, so fewer claims come back denied. For health plans, it checks incoming requests against medical policy, organises the findings for the clinical reviewer, and explains benefits and costs to members from their own plan documents. For Medicare Advantage members, it applies Medicare's coverage criteria where a plan's own rule is stricter.
Four Parts of the Product
Google describes four components. Skills are reusable instructions written by domain experts that teach the agent to run a specific task, such as medical record synthesis, pre-appointment preparation, coverage review or benefit determination, using the organisation's own policies and plan rules. Connectors link the agent directly to the EHR, claims, care management and administrative systems a team already runs, without migrating data or changing permissions.
The third component is an open partner ecosystem, with systems integrators and digital health partners extending the agent into specialised tools. The fourth is a governed platform with open model choice, where clinical, security and IT leaders set policies for every agent interaction, with audit logging and citations so reviewers can trace where an answer came from. Google says the product is built to support HIPAA compliance, and that clinical data, prompts and outputs stay inside the customer's governed environment and are never used to train its foundation models.
EHR and Coverage Connections
The EHR connections are read only. The Epic connector gives authorised users access to conditions, medicines, labs, clinical notes and specialist documentation, with no write-back to the chart. The Oracle Health connector surfaces patient histories, diagnostics and lab results in the same way. Google also lists an InterSystems FHIR server for standardised record exchange, Verato for patient identity matching, and Keragon for connections to more than 400 healthcare applications.
On the coverage side, the agent connects to CMS coverage rules, including national and local coverage determinations, so its reasoning rests on published policy rather than model memory. Optum eligibility and coordination of benefits tools are in limited release, and TriZetto verifies eligibility, coverage, co-pays and deductibles across Medicare, Medicaid and major commercial plans. GoodRx surfaces prescription pricing and savings options, and Infinitus Relay automates payer phone calls for prior authorisation and benefit verification. Google says a NextGen Healthcare connection is in development and due in the coming months. In total, Google named 27 connectors at launch, along with 16 systems integrators.
Who Is Using It First
Health systems, health plans, academic medical centres and clinical leaders are already working with Google Cloud on the product. The named organisations include HCA Healthcare, Northwell Health, Highmark Health, Optum, Blue Cross Blue Shield of Massachusetts, Hartford HealthCare and Stanford School of Medicine, with the American Society of Clinical Oncology contributing an oncology guideline assistant. HCA Healthcare's chief clinical officer, Michael J. Schlosser, said: "Making it easier for physicians to access and understand a patient's medical history helps them spend less time gathering information and more time focused on patient care."
The preview is open to Gemini Enterprise Standard and Plus customers, offered as per-user subscriptions or on a pay-as-you-go basis. Google has not announced prices or a date for general availability. The launch follows earlier industry versions of the Gemini agent for government services, financial services and legal work.
Why This Matters for Healthcare Organizations
Prior authorisation and chart preparation are two of the most time-consuming jobs in a clinic or hospital, and they are also where delays and denials start. An agent that can gather the record, check it against the payer's policy and draft the packet before submission speaks directly to that problem. The read-only EHR design matters here too. It lets the agent work across systems without touching the chart itself, which should make clinical and compliance leaders more comfortable than a tool that writes back.
The same idea is already reaching smaller practices through patient-facing tools. A patient portal that handles intake, scheduling and messages takes the same kind of administrative load off the front desk, and our look at how an AI medical receptionist handles after-hours calls shows what that looks like day to day. Gemini for healthcare aims at the heavier enterprise version of the same work, inside the EHR and the payer's systems rather than at the front desk.
Analysis: A Preview, Not a Proven Result
This is a preview launch, and it should be read as one. Google has published no benchmarks, error rates or pilot results for the healthcare workflows, and the launch organisations have not published measured outcomes either. The announcement describes what the agent is designed to do, not what it has been shown to do in live care settings.
Three things are worth watching. First, whether the read-only design and citation trail hold up under real clinical review, since an authorisation packet is only useful if a reviewer can trust every line in it. Second, pricing and general availability, neither of which Google has announced, since both will decide whether this reaches community hospitals and larger medical groups or stays with the biggest systems. Third, the NextGen Healthcare connector, which will show whether the platform genuinely spans the EHR market or stays centred on Epic and Oracle Health.
Sources and References
- Google Cloud, Introducing Gemini for healthcare, by Thomas Kurian, 10 November 2026: https://cloud.google.com/blog/products/ai-machine-learning/gemini-for-healthcare
- Futurum Group, Gemini for Healthcare Reads the EHR, 10 November 2026: https://futurumgroup.com/insights/gemini-for-healthcare-reads-the-ehr/
- Unite.AI, Google Cloud Opens Gemini for Healthcare Preview in the U.S., 10 November 2026: https://www.unite.ai/google-cloud-opens-gemini-for-healthcare-preview-in-the-u-s/
- Crypto Briefing, Google Cloud launches Gemini for healthcare to tackle clinical and admin workflows, 10 November 2026: https://cryptobriefing.com/google-cloud-launches-gemini-for-healthcare/
Reporting based on the sources above, accessed 11 November 2026. The Google Cloud blog post was read directly, and the Futurum Group, Unite.AI and Crypto Briefing reports were read directly and cross-checked against it. Med Clinic X did not conduct original reporting or interviews for this article. Corrections: contact us via https://medclinicx.com/contact