5 health IT leaders’ bold AI predictions for 2027

Advertisement

Artificial intelligence could reshape hospital IT jobs, introduce new financial pressures and change how patients interact with health systems in 2027 as the technology moves beyond generating information to taking action across healthcare workflows.

Five hospital AI and technology leaders shared their biggest predictions for 2027 with Becker’s, pointing to a shift toward agentic AI and a growing need for health systems to rethink how they manage costs, employees, cybersecurity and patient interactions.

“2025 was the year predictive AI went mainstream, 2026 was the year of generative AI, and 2027 will be the year of agentic AI,” Micky Tripathi, PhD, chief AI implementation officer at Rochester, Minn.-based Mayo Clinic, told Becker’s.

Dr. Tripathi expects agentic AI to begin seeing meaningful deployment in 2027, initially concentrated in administrative functions and revenue cycle operations. Federal requirements for payers to make claims data and electronic prior authorization capabilities available through application programming interfaces could help accelerate adoption, he said.

He also expects AI to become more deeply integrated with EHRs, continuously synthesizing medical records and helping clinicians navigate information and make decisions.

John Rares Almasan, PhD, senior vice president and chief artificial intelligence, data and infrastructure officer at Phoenix-based Banner Health, anticipates a similar transition.

“My biggest prediction for 2027 is that AI will move decisively from tools that help people complete individual tasks to agentic systems that orchestrate work across entire care journeys,” Dr. Almasan told Becker’s.

Banner is investing more than $1 billion in AI, robotics, telemedicine and cloud-based platforms, according to Dr. Almasan. He expects the next generation of AI to connect clinical, operational and administrative workflows rather than operate as isolated productivity tools.

But he cautioned that health systems may be underestimating the organizational changes required to support those capabilities.

“The constraint won’t be access to more powerful AI models. It’ll be whether the enterprise is ready for AI to act,” he said.

That readiness will require trusted data, interoperable platforms, identity and access controls, real-time governance and clear accountability, Dr. Almasan said.

At Richmond, Va.-based VCU Health, Alok Chaudhary, vice president and chief data and AI officer, expects AI to change not only how health systems use data but also how their IT teams are structured.

Mr. Chaudhary told Becker’s that VCU Health is increasingly moving toward conversational AI tools that allow employees to ask questions about organizational data rather than rely exclusively on traditional dashboards.

More than 200 users are already using conversational AI at the health system, he said.

The shift could reduce the need for analytics teams to continually build and modify dashboards or generate reports in response to employee requests. Mr. Chaudhary also expects traditional distinctions between data engineers, data architects, business analysts and business intelligence developers to become less rigid as AI enables employees to handle more tasks across those functions.

He said VCU Health is exploring how to organize analytics employees around solving specific clinical and operational problems rather than maintaining narrowly defined technical roles.

However, another development concerns him as AI adoption accelerates.

“One of the things that does keep me up at night is the cost,” Mr. Chaudhary said.

He pointed to the financial implications of consumption-based AI pricing as vendors introduce more AI capabilities into enterprise platforms.

VCU Health has established usage caps for developers and is paying closer attention to whether AI investments generate measurable returns.

Mr. Chaudhary raised the possibility that an AI solution could reduce staff hours but still cost more than the labor savings it produces. He expects health systems to pay closer attention to usage, expenses and return on investment in the coming year.

VCU Health is also working with Epic on an AI functionality roadmap and participated in an early buildathon involving the EHR vendor’s agent factory.

As health systems begin deploying more autonomous AI, other technology leaders expect new challenges in how organizations manage those systems.

Tom Barnett, chief information and digital officer at Memphis, Tenn.-based Baptist Memorial Health Care, predicts that oversight of AI agents will increasingly resemble processes already used to manage employees.

That could include establishing “agent IDs,” authentication requirements, role descriptions, onboarding and offboarding procedures, and policies governing which systems and data autonomous agents can access.

Mr. Barnett also anticipates organizations will need common frameworks to monitor agent performance, outcomes and costs, including token consumption.

As AI-related security threats grow, he expects those safeguards to become a “price of entry” for organizations deploying autonomous agents.

For Dr. Tripathi, one of the biggest changes could come from outside hospital walls.

He predicts patients will increasingly use their own AI agents to interpret medical records and interact with health systems on their behalf.

“As a result, the primary digital ‘front door’ to a health system may begin shifting from websites and patient portals designed for humans to navigate toward interactions with personal AI agents acting on patients’ behalf,” he said.

Those agents could seek to schedule appointments, obtain real-time pricing and benefits information, make payments and coordinate care. Insurers, referring physicians and other organizations could also deploy agents that interact with health systems.

That shift could require hospitals to rethink how they authenticate AI agents, determine what information they can access and establish which actions they are permitted to take.

Meanwhile, Simon Nazarian, chief digital and technology officer at Duarte, Calif.-based City of Hope, expects AI to expand where specialized healthcare expertise can be delivered.

“In 2027, I think we’ll see AI begin to fundamentally change how and where specialized healthcare expertise can be delivered,” Mr. Nazarian told Becker’s.

He anticipates health systems will increasingly use AI to extend the expertise of clinicians and scientists, support complex treatment decisions, identify clinical trial opportunities and help patients navigate care.

At City of Hope, he sees potential for AI to make specialized cancer expertise accessible to more patients, regardless of their location.

However, Mr. Nazarian said hospital executives may be underestimating the organizational transformation required to scale AI responsibly, including the need for stronger data infrastructure, cybersecurity, clinical validation and trust among employees and patients.

Together, the predictions suggest hospital leaders could face a different set of AI challenges in 2027 as the focus shifts from testing individual applications to managing technology that can take action across the enterprise.

“The organizations that lead this next chapter won’t necessarily be the ones with the most AI,” Dr. Almasan said. “They’ll be the ones that redesign care and operations around it, using AI to remove friction and complexity while strengthening control, trust, clinical judgment, and the human connection with patients.”

Advertisement

Next Up in Artificial Intelligence

Advertisement