AI Dominates HIMSS26, but the Real Story Is How Hospitals Are Managing the Risks.

AI took center stage at HIMSS26, but the deeper conversation was about balance—how health systems can adopt agentic AI tools at speed while keeping patient safety, privacy, and cybersecurity front and center. Across sessions, speakers painted a picture of a healthcare sector eager to innovate, but mindful of the operational risks that come with large‑scale automation. For health IT investors, the event revealed where opportunity is meeting caution in 2026.

AI Agents Step into Real Operations

This year’s conference made clear that AI agents—autonomous systems capable of executing multi‑step workflows—are moving from pilot projects into production. Health systems and vendors showcased early successes in areas like staff onboarding, claims management, reimbursement, and patient scheduling, where administrative complexity drives costs and clinician burnout.

Leaders from Google Cloud (Alphabet Inc., NASDAQ: GOOGL) and Optum (UnitedHealth Group, NYSE: UNH) said these tools can now manage documentation and billing end‑to‑end, compressing reimbursement timelines with minimal human oversight. The message was consistent: agentic AI has evolved from experimental hype into practical technology, especially for revenue cycle automation and back‑office workflows.

Interoperability Emerges as a Critical Enabler

Panelists agreed that interoperability will determine whether AI can truly scale in healthcare. Speakers from Mayo Clinic and Wolters Kluwer N.V. (Euronext Amsterdam: WKL) warned that fragmented data remains the biggest barrier to effective AI deployment. Even the best models can’t deliver meaningful insights if critical patient information remains siloed across different EHRs, portals, or even fax systems.

Federal officials reinforced that improving data liquidity is now a key policy focus, citing risks like medication errors and care gaps when providers lack complete information. With AI increasing demand for robust, real‑time data access, interoperability is no longer just a compliance topic—it’s becoming a strategic necessity.

EHR Vendors Expand Embedded AI Suites

Major EHR vendors seized the HIMSS26 spotlight to demonstrate how deeply embedded AI has become in day‑to‑day clinical operations. Epic Systems reported that more than 85% of its clients now use built‑in AI to generate discharge summaries, automate documentation, and streamline revenue cycle tasks. Meditech (Medical Information Technology, Inc.) showcased its latest AI‑driven recording, claims‑denial, and data‑query tools.

CIOs emphasized that EHR‑native AI holds a unique advantage because it operates directly within clinical workflows, drawing on cleaner and more contextual data. Still, both Epic and Meditech reaffirmed their support for third‑party integrations, signaling that the EHR ecosystem is evolving into an open platform model rather than closed architecture.

AI Governance Becomes a Continuous Discipline

Governance was another recurring theme. Health system leaders, including executives from Mass General Brigham, noted that static governance committees no longer suffice in an era when staff can easily access models from OpenAI, Perplexity (Perplexity AI), and Anthropic’s Claude (Anthropic PBC). Organizations are adopting tiered, use‑case‑specific frameworks that tailor oversight depending on whether the AI is administrative or clinical in nature.

Speakers stressed the importance of continuous education, policy updates, and documentation—turning AI governance into a living system that evolves alongside the technology. Mature programs, they said, increasingly focus on practical controls rather than policies that exist only “on paper.”

Non‑Human Identity Management Becomes a Security Priority

As hospitals scale up their use of AI agents, non‑human identity management has emerged as a new cybersecurity frontier. Security experts warned that every autonomous agent interacting with patient or financial systems represents a new digital identity that must be tightly managed. Without granular access controls, organizations risk allowing poorly governed software to operate unchecked.

Vendors such as Imprivata Inc. introduced “agentic identity” tools designed to manage, monitor, and audit AI agents separately from human users. Unlike traditional accounts, these identities can proliferate rapidly, making them harder to track and secure. The takeaway was clear: the next wave of cybersecurity in healthcare will focus on protecting digital actors, not just people.