Generative AI startup Anthropic has been commanding attention from health system executives in the days since it launched Claude for Healthcare.
The company, with a reported valuation of $350 billion and the backing of a who’s who of Big Tech firms, made its entry into healthcare on Jan. 11, just a week after a major competitor, ChatGPT developer OpenAI, did the same. Anthropic’s offering promises faster prior authorizations, better care coordination, and quicker regulatory filings for drug development.
Healthcare executives told Becker’s they think Claude for Healthcare has an opportunity to be the preferred organizationwide AI platform for health systems, while OpenAI will likely be the go-to for consumers.
“Anthropic appears to be taking an enterprise-first approach and connecting to health systems, apps and their data and with a strong focus on HIPAA-compliance,” said James Whitfill, MD, senior vice president of strategic partnerships and chief transformation officer of Scottsdale, Ariz.-based HonorHealth.
“ChatGPT has 200 million people per week asking health questions and adding their data to it today. That is effectively a crowdsourced health record vault that even Google and Microsoft never could pull off. Anthropic does not have that user base, and so they are looking to be the trusted partner for others who already have relationships with patients.”
Phoenix-based Banner Health has been an early adopter of Claude at the enterprise level. In late 2025, the 33-hospital system offered BannerWise, a private chatbot built off Anthropic’s technology, to all of its 55,000-plus employees. It fits into the organization’s goal of cutting administrative tasks for clinicians in half by 2030.
The $15.6 billion system went with Claude because of its focus on privacy and accuracy, said Michael Reagin, executive vice president and chief technology officer of Banner Health. The chatbot tells you when it doesn’t know an answer rather than making one up.
An early use case has been chart summarization for oncology patients, whose medical records can run hundreds of pages and contain PDFs, images and even paper faxes. Manual previsit summarization takes about eight hours per patient. Claude has processed over 1,400 Banner oncology clinical notes since June.
Banner Health plans to next deploy the tool to streamline call center, revenue cycle and supply chain operations.
“We see a large ecosystem of AI relationships really contributing to the overall betterment of Banner patients,” Mr. Reagin said. “We view the relationship with Anthropic as being an anchor point that helps us orchestrate a lot of other AI.”
If anything, Claude for Healthcare gives hospital leaders another option in the burgeoning healthcare AI market, which includes everyone from the largest EHR vendors to a plethora of tiny startups.
“It’s hard to compare the other healthcare gen AI offerings as they change minute to minute, but it sure is fun fantasizing about instantaneous prior authorizations,” said Richard Zane, MD, chief medical and innovation officer of Aurora, Colo.-based UCHealth. “At the end of the day, Claude will likely be one of the models in an enterprise AI layer, with strengths in safety, connectors, and long-context reasoning rather than from exclusive capabilities no one else can match.”
At the same time, several health systems are building their own solutions. Take Palo Alto, Calif.-based Stanford Health Care, which created an EHR-integrated chatbot for clinicians called, naturally, ChatEHR. Both Anthropic and OpenAI’s new platforms include added healthcare AI capabilities for developers.
“I am quite excited about Claude for Healthcare (as well as the OpenAI for Healthcare offering),” said Nigam Shah, MD, PhD, chief data scientist of Stanford Health Care. “From our standpoint, it will accelerate the development of solutions such as ChatEHR, reducing our total cost of developing workflow automations that have enterprise value for us. I look forward to comparing the out-of-box functionalities of the different enterprise frameworks for generative AI in healthcare.”
But privacy remains paramount, healthcare leaders say. Muhammad Siddiqui, CIO of Richmond, Ind.-based Reid Health, said HIPAA readiness must go beyond marketing and be proven in practice. That means clear business associate agreement terms, data retention rules, audit logs, and access controls. Both Anthropic and OpenAI say patient data will not be utilized to train their models.
“If a vendor cannot answer who sees the data, how long it stays, and what it is used for, we do not feed it patient records,” Mr. Siddiqui said.
Reid Health filed a lawsuit Jan. 13, with Epic and some health system peers, against Health Gorilla and other health data companies over alleged misuse of patient information.
The accuracy of the AI’s medical advice is another concern. “Perhaps adding additional data will help, but I worry that these tools will be viewed by laypeople as accurate without truly understanding the risks,” said Patrick Woodard, MD, CIO of Rapid City, S.D.-based Monument Health. “In medicine, the first rule is primum non nocere. Last I checked, these chatbots haven’t sworn their Hippocratic oath. For now, I’ll stick with the humans who have.”
However, Sunil Dadlani, executive vice president and chief information and digital transformation officer of Morristown, N.J.-based Atlantic Health, called Claude for Healthcare a “game-changer” because it fills the “trust gap” that has long existed in health AI.
“Unlike its competitors, Claude is built on a ‘Constitutional AI’ framework, which essentially gives it a predefined ethical compass,” he said. “For a patient or a healthcare consumer, this is a massive win — it means the AI is hardwired to be cautious, honest, and far less likely to ‘hallucinate’ a medical answer just to be helpful.”
He said Claude also has the advantage of being able to handle large amounts of data (known in large-language model vernacular as “tokens”, or units of text). “You can drop a 300-page patient record into it, and because of its 200,000-plus token window, it won’t ‘forget’ the allergies listed on Page 1 by the time it reaches the surgical notes on Page 300,” he said.
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.