Dr. Tom Milam, chief medical officer at Iris Telehealth and a practicing psychiatrist
Photo: Dr. Tom Milam
Andy Flanagan, CEO of behavioral health virtual care provider Iris Telehealth, and Dr. Tom Milam, chief medical officer at Iris Telehealth and a practicing psychiatrist, have been closely examining how health systems are approaching artificial intelligence implementation in behavioral health. They see a critical shift on the horizon.
Currently, many healthcare provider organizations are experimenting with AI tools in isolated pilot programs. 2026 will mark the year when successful health systems move these systems from the pilot phase into core operational workflows, Milam said.
An essential operational tool
"AI will transition from experimental technology to an essential operational tool that helps behavioral health programs identify which patients need urgent attention and allocate limited clinical resources more effectively," Milam continued.
"The key distinction here is that AI in behavioral health should focus primarily on operational efficiency rather than clinical decision-making – helping organizations optimize scheduling, resource allocation and access to care rather than attempting to replace clinical judgment," he added.
Some top academic medical centers are moving in this direction. Duke University School of Medicine recently received a $15 million grant from the National Institute of Mental Health to expand its AI model.
Its model predicts worsening mental health up to a year in advance with 84% accuracy, moving from pilot studies into real-world clinics across rural North Carolina, Minnesota and North Dakota.
Patient intake and scheduling
"The transformation we anticipate centers on how health systems handle patient intake and scheduling for behavioral health services," Milam said. "Traditionally, these processes have been reactive: A patient calls for an appointment, gets scheduled based on availability, and then often has to wait weeks or even months for care.
"In 2026, we anticipate health systems will implement AI-driven models that can analyze patterns in patient data – such as appointment history, no-show rates and utilization patterns – to identify individuals who may be at a higher risk of crisis or escalation," he continued.
For example, if a patient has missed multiple appointments, has had recent emergency department visits or shows patterns suggesting increasing acuity, an AI system could flag that individual as needing more immediate attention.
"The critical operational distinction is these systems help organizations answer the question, 'Given our capacity to see 100 patients this week, which 100 patients have the most urgent need?'" Milam explained. "This represents a shift from first-come-first-served scheduling to intelligent prioritization based on operational data.
"However, you must understand what this technology is and isn't," he added. "As Andy frequently emphasizes in his discussions about AI in healthcare, current AI tools are fundamentally probability engines that excel at pattern recognition and administrative tasks, not clinical diagnosis or treatment planning."
Behavioral health risk stratification
The AI models being deployed for behavioral health risk stratification are machine learning systems trained on operational and utilization data – they work best when a human is involved in the workflow to review recommendations and make final decisions, Milam said.
Iris Telehealth recently surveyed 1,000 U.S. consumers and found that 73% want human providers to make final decisions in AI-flagged emergencies. The company encourages healthcare organizations to view AI as a support tool, not a decision-maker – ensuring safety and accountability in crisis care.
"Health systems that successfully scale these AI systems in 2026 will be those that maintain appropriate clinical oversight, clearly define the operational problems they are solving and resist the temptation to overreach into clinical territory where the technology is not yet ready," he concluded.
Follow Bill's health IT coverage on LinkedIn: Bill Siwicki
Email him: bsiwicki@himss.org
Healthcare IT News is a HIMSS Media publication.
