FAQ: How AI is changing managed care in 2026 and what leaders need to watch

Artificial intelligence (AI) is quickly becoming part of daily operations in managed care. For example, health systems and plans are using AI tools to reduce documentation burden, support decision-making and improve workflow efficiency. However, as adoption grows, recent research has shown that leaders are sharing their concerns about value, oversight and accountability.

Here’s what executives should know, based on industry research and insight from Ben Scharfe, CPA, executive vice president of artificial intelligence at Altera Digital Health.

1. Where is AI delivering the most measurable impact today?

AI is being used across healthcare operations, including documentation support, workflow automation and clinical summarization tools. According to research published in the Journal of the American Medical Informatics Association, ambient AI tools that generate clinical notes can significantly reduce documentation time and ease administrative workload for clinicians.

Scharfe said the biggest impact is happening around the patient encounter itself. AI tools are helping reduce pre-visit chart review time and easing post-visit documentation through ambient listening and record summarization.

By reducing documentation tasks before and after visits, organizations can help clinicians focus more directly on patient care.

2. Can AI truly reduce administrative burden? How should leaders measure ROI?

According to an AMA survey of nearly 1,200 physicians, many see AI as most valuable for reducing administrative burdens that add hours to the workday. About 57% of physicians said automation could help with documentation, prior authorizations and other non-patient-facing tasks that contribute to stress and burnout.

The survey also highlighted that AI could improve work efficiency and job satisfaction, showing the strong potential for technology to support physicians’ daily workflow.

Measuring return on investment (ROI) for AI can also be challenging, according to Scharfe. Savings are not always immediate or easy to calculate using traditional cost models.

Scharfe advised executives to look beyond direct cost savings and suggested tracking physician and patient satisfaction as leading indicators of turnover, which carries real financial impact. He also recommends monitoring metrics referred as “pajama time,” or after-hours charting, and the human-AI agreement rate, which is how often clinicians agree with AI recommendations.

“If a tool lacks transparency or requires frequent manual overrides, it is adding cognitive complexity and frustration rather than true value,” Scharfe said.

In other words, ROI should include trust, how easy these tools are to use, and whether they’re actually cutting the workload, not just whether it automates tasks.

3. What governance safeguards should plans and systems have in place?

As AI tools expand, governance becomes more important. Organizations need clear processes for evaluating, monitoring and verifying AI systems.

According to the National Institute of Standards and Technology AI Risk Management Framework, organizations should establish oversight, testing and monitoring processes to manage AI-related risks.

Scharfe said health systems should develop internal expertise to evaluate tools rather than relying solely on external assurances. He stressed transparency, including the ability to trace AI outputs back to original patient data so clinicians can verify the information themselves.

“A critical safeguard is ensuring transparency through in-line citations to original patient data,” he said. “This ensures that the human expert can easily verify the AI's ‘work,’ keeping the ultimate accountability with the clinician rather than the algorithm.”

4. What is one common mistake organizations make when adopting AI?

One major mistake is treating AI like traditional software, according to Scharfe. Unlike deterministic systems, modern AI produces probabilistic outputs that require review and professional judgment.

He warned this mindset can lead to “automation complacency,” where staff accept AI outputs without proper review. He also noted that many organizations make the mistake of thinking a vendor’s contract transfers liability, when in fact the health system still holds the legal responsibility for patient care.

5. What should managed care leaders prioritize over the next 12 months?

Federal health IT policy is evolving. According to the Office of the National Coordinator for Health IT, the proposed Health Data, Technology and Interoperability (HTI-5) rule could ease some requirements and simplify certification.

As regulatory policies shift, organizations could carry more responsibility for evaluating AI tools internally.

Scharfe added that leaders should prioritize building strong internal governance structures and selecting vendors who act as responsible partners. Organizations must also develop internal expertise to test and verify AI performance rather than relying solely on federal standards.

As AI adoption grows, careful oversight and accountability will be just as important as innovation.

This FAQ was approved for accuracy and for publishing by Scharfe.