Hospitals and health systems are entering a period of rapid acceleration in AI. Clinical leaders are reporting tangible gains, from automated note-taking to clearer visibility into patient histories, while executives are increasingly focused on what it will take to scale these tools responsibly. Pressure is building from several directions: new federal guidance on interoperability, rising expectations around patient experience and persistent concerns about administrative complexity.
AI now touches each of these challenges, and some organizations — including Oracle Health and Life Sciences — see the EHR as a critical lever for translating promise into sustained improvement. Rather than viewing the EHR as a static system of record, the company frames it as the foundation for a more intelligent healthcare infrastructure, shaped by its experience across cloud technology, data management and healthcare operations.
Few leaders have had a more panoramic view of these shifts than Seema Verma, executive vice president and general manager of Oracle Health and Life Sciences, and former administrator of CMS. In a recent conversation with Becker’s Healthcare, she outlined how she expects AI’s role in healthcare to evolve in 2026 and why several foundational challenges still require industry-wide focus.
Three themes stood out in particular: the opportunity to broaden access to care, the urgency of improving interoperability and the transformative potential of AI agents for both patients and clinicians.
2025: a year of firsts + accelerated innovation
Ms. Verma reflected on the past year and the momentum Oracle is carrying into the next. The company’s most significant accomplishment of 2025, according to Ms. Verma, was its AI-native EHR securing certification from the HHS Assistant Secretary for Technology Policy and Office of the National Coordinator for Health Information Technology (ASTP/ONC). She described the shift as comparable to the moment mobile technology moved from early smartphones to the first iPhone.
“We’re talking about a next-generation product,” Ms. Verma said. “It’s going to be a game changer.”
The approval marked a turning point for Oracle, which has been working to develop tools that can relieve clinicians of time-consuming administrative tasks while offering more useful guidance at the point of care. For example, the organization saw rapid adoption of its Clinical AI Agent — recognized by KLAS for its EHR integration capabilities — with nearly 10,000 clinicians across dozens of specialties using the tool to streamline documentation and generate patient summaries.
Ms. Verma pointed to a measured outcome that aligns with the “patients over paperwork” ethos she championed at CMS: Physicians using Oracle’s Clinical AI agent are seeing more than 40% reduction in time spent on documentation. Many have regained more than an hour of active EHR time in their workdays.
Another area of momentum came from specialty care. Ms. Verma described how academic and community cancer centers expressed strong interest in partnering with Oracle after reporting that existing systems were not meeting the demands of oncology care. More than a hundred sites have engaged with Oracle to explore how its platform could support the complexity of cancer workflows, including the need for more precise data, integrated decision support and closer alignment with emerging therapies.
Underlying all of this, Verma emphasized, is Oracle’s broader AI strategy. She noted that healthcare organizations often speak about AI in sweeping terms, but the strength of any model or AI system depends on the quality and diversity of the data it can interpret.
“The quality of the AI that we’re bringing to the market is really important,” she said. “If your AI is working off of 13 years of historical data, it’s going to give you advice and practice patterns from 13 years ago — whereas ours is multimodal. It can consider data from multiple sources, whether it’s CMS rules, current literature or changes going on in the market.”
This approach is intended to mirror how clinicians make decisions in practice. Just as important, each output includes a citation path to its underlying data source, so users can inspect the reasoning rather than accept a black-box suggestion. For Ms. Verma, this transparency is central to earning the trust of clinicians and ensuring AI enhances — rather than overrides — human judgment.
Expanding access to care through the EHR
One of the most promising applications of AI, Ms. Verma said, lies in expanding equitable access to clinical trials. Participation in clinical trials has remained stubbornly low in the U.S., driven by factors like geography, limited awareness and uneven access to specialty care. Patients receiving treatment in rural or community settings may never learn they qualify for a trial, while those treated at large academic medical centers are more likely to encounter these opportunities.
The EHR can play a central role in closing that gap. Ms. Verma explained how today, clinical trial matching typically occurs outside the EHR; eligibility criteria are often managed through separate research systems, requiring clinicians and researchers to manually move data between platforms. Oracle’s unique position in both clinical and research workflows creates an opportunity to integrate trial matching directly into care delivery.
“We’re the only company in the world that actually has both an EHR and an electronic data capture system,” she said. “By bringing those pieces together, we can better understand clinical trial criteria and identify which patients may be eligible.”
AI can help identify eligibility in far more nuanced ways, particularly in oncology, where trials may hinge on specific genomic markers or tumor characteristics. It’s an area where Oracle is beginning to see “some early wins,” Ms. Verma said. She described a future in which both clinicians and patients are notified of relevant trials directly within the EHR or patient portal, making trial awareness part of routine care. What was once a flyer taped to a registration desk could instead become an automated, integrated element of care planning.
Ms. Verma also pointed to the broader economic stakes. Much of the cost of drug development stems from slow and uneven trial enrollment. If matching becomes more efficient, pharmaceutical firms could reduce operational complexity and shorten development timelines. For health systems, this could eventually influence drug pricing and expand treatment options for patients, Ms. Verma posited. She added that the work around trial matching reflects Oracle’s mission to address structural challenges that have resisted progress for decades.
Interoperability and the data foundation AI requires
Despite meaningful policy advances, interoperability remains difficult. Ms. Verma spoke candidly about the historical role EHR companies have played in blocking broader data exchange. The industry has made progress on common standards, yet real-time data flows between organizations remain inconsistent, limiting the usefulness of AI tools that depend on complete information.
“AI works well when it has good data underneath it,” Ms. Verma said. “If we don’t have complete, real-time data, we will not be able to leverage AI. For Oracle, interoperability and an AI strategy go hand-in-hand.”
Oracle has taken several steps to strengthen interoperability, including participating in federal frameworks and securing TEFCA Qualified Health Information Network designation. Its approach is rooted in an open invitation to exchange data with any organization, not only existing clients. Ms. Verma also highlighted Oracle Cloud Infrastructure’s “military-grade” security that is used to protect the most sensitive data at some of the largest and most sophisticated businesses, national defense agencies, and governments around the world, noting that it already hosts many large AI companies and enables closer collaboration across the ecosystem.
Interoperability also intersects with a challenge Ms. Verma has emphasized since her time at CMS: friction between payers and providers. Prior authorization remains one of healthcare’s most persistent administrative pain points. Ms. Verma sees potential for AI-supported EHR workflows to streamline record retrieval, reduce manual back-and-forth and move claims more efficiently toward adjudication. The goal, she said, is to return time and resources to both sides of the system and reduce the sizable share of healthcare spending devoted to administrative work.
Agentic AI and the evolution of the patient portal
For many patients, the portal has not kept pace with advances in consumer technology. Most portals allow scheduling, basic messaging and access to test results, but offer limited support in interpreting those results or understanding changes in health over time. Ms. Verma noted that patients increasingly turn to public AI tools to make sense of lab values — signaling a broader demand for more interactive, personalized guidance.
Oracle previewed its next version of the patient portal in 2025. It brings several of the same capabilities clinicians are beginning to use (ambient listening, summarization, conversational search) into a patient-facing environment. Through Oracle’s relationship with OpenAI, patients may soon have access to an assistant that can help them locate information in their records, explain unfamiliar terms or highlight patterns in recent tests.
The goal is to shift the portal from a transactional interface to a tool that helps patients make sense of their care. “Some of the things you’re used to doing with modern technology in your car, in your home — that ambient experience — we’re bringing it to the portal, which is very exciting,” she said.
Looking to 2026, Oracle plans to introduce a broader set of AI agents across revenue cycle, nursing and clinical operations, many of which are expected to be available within existing systems as well as the new EHR. Ms. Verma also previewed an AI data platform for life sciences that is being developed to combine de-identified data with analytical tools to help researchers explore potential treatments and accelerate discovery.
A turning point for 2026
As health systems navigate financial pressure, workforce shortages and rising patient expectations, the question is no longer whether AI will influence healthcare but how quickly organizations can adopt it in ways that deliver lasting value. Ms. Verma believes the coming year represents a pivotal moment: Progress in access, data exchange and AI-enabled workflows will depend on how effectively the industry addresses long-standing structural challenges.
Oracle’s work in 2025 demonstrates how quickly the landscape can change when foundational tools evolve. If the next stage of AI in healthcare delivers on its early promise, the EHR may become not only a system of record but a catalyst for broader transformation.
“We feel like this is a turning point in terms of what we’re bringing,” Ms. Verma said.
