AI-native EHRs: A new era or just a new label?

What does it mean to be AI-native?

It's too early in the game for a standard definition of 'AI-native' to exist. Because of this, the term's meaning can vary significantly.

Crystal Broj, enterprise chief transformation officer at MUSC, weighed in on the phrase.

"If it simply means an EHR has generative text or summarization tools added on top, that's more of an enhancement than a reinvention," Broj said.

But when AI capabilities are deeply embedded and aligned with clinical workflows, rather than layered on top of a clunky legacy system, it can make a meaningful difference.

"If the AI is natively driving documentation, decision support and personalization in real time -- not bolted on, but built in -- then it can absolutely change usability," said Broj.

Another expert emphasized how much the origin of the platform matters.

"It's a buzzword for sure," said Nigam Shah, MBBS, Ph.D., chief data scientist at Stanford Healthcare. "People need buzzwords. But if an EHR company started after this AI revolution began, it can be AI native -- legitimately."

"You can only be a native if you began when the thing was around," he added. "If a company has already existed for 20 years and is now adding AI -- that's a bolt-on."

Inside the AI-Native playbooks at Oracle and athenahealth

Despite the conflicting views on what constitutes an 'AI-native' system, EHR vendors are going all-in on the term.

Oracle and athenahealth have both described their EHR platforms as 'AI-native,' emphasizing that the systems have been rebuilt around AI.

"Oracle's EHR is not an evolution of what came before -- it's a reimagined foundation for how work gets done in healthcare," said Seema Verma, executive vice president and general manager at Oracle Health and Life Sciences.

According to Verma, Oracle's EHR was redesigned from the ground up to make AI an active participant in the care process, generating draft content for workflows in the ambulatory setting. Some use cases include documentation, ordering, coding and prior authorization.

"Instead of sidecar bots, our AI agents are embedded directly into workflows. They anticipate tasks, surface relevant insights and support clinicians while keeping them in control," Verma says.

Athenahealth also focuses on embedding AI throughout the EHR. Chad Dodd, vice president of product management, explained that the platform incorporates both deterministic and generative AI to support functions such as inbox workflows, patient summaries and lab result interpretation.

"Our approach is fundamentally different. We don't bolt on AI -- we build it into a shared platform in the cloud," Dodd said.

According to Dodd, this cloud-based foundation is what enables athenahealth to embed AI capabilities natively across the entire clinical workflow for customers who opt in.

Evaluating the impact of 'AI-native' solutions

Both companies claim the goal of an improved user experience, though measurable outcomes are still emerging. According to Dodd, about 30-40% of athenahealth clients have fully opted into its suite of AI-native features. Many others have adopted specific tools individually, with high demand for ambient documentation and chart summaries.

Verma said that Oracle's system mirrors the ease of consumer apps. Clinicians can navigate with voice, access draft notes immediately and interact with agents to complete tasks.

But not everyone agrees that AI-native design dramatically transforms day-to-day clinician experience.

"From a cognitive burden standpoint, whether it's AI-native or not, I don't think that's going to be a big difference," Shah said. "The primary difference will not be on the end user, but on the IT teams -- maintaining, educating, upgrading, bug fixing."

Shah argued that instead of focusing on labels, organizations should ask whether AI -- native or not -- reduces the total cost of ownership.

"The metric we should be using is: does this reduce the total maintenance cost of the platform?" he said. "Because if making it AI-native means they charge me twice as much, what's the good in that?"

AI-native EHR systems may have potential, but according to Broj, it's too early to say whether these tools will deliver meaningful change in practice.

"If it reduces clicks, cognitive load and friction, then it's meaningful," she said. "But we need to move past the marketing and measure the impact."

Dr. Shah echoed the need for clear evaluation metrics.

"Just being AI-native isn't good enough," he said. "As someone buying the software, I want to know: What does it actually reduce?"

Elizabeth Stricker, BSN, RN, comes from a nursing and healthcare leadership background, and covers health technology and leadership trends for B2B audiences.