Why Behavioral Health Can’t Wait for AI and Telehealth Policy Clarity

The following is a guest article by Andy Flanagan, CEO at Iris Telehealth

Behavioral health providers are expanding access to care while navigating three regulatory uncertainties: When will DEA prescribing rules become permanent instead of getting last-minute extensions? Will telehealth reimbursement parity return after enforcement stopped? And who’s legally responsible when AI flags a patient in crisis, but no one intervenes in time?

States passed more than 250 AI health bills in 2024 while federal guidance remains voluntary. Critical telehealth policy decisions in 2026 will determine whether virtual behavioral health stays financially viable. Organizations like Providence are already deploying virtual nursing and AI triage tools in emergency departments, but most providers still don’t know what rules they’re operating under — or what liability they’re assuming.

The answer isn’t waiting for policy clarity that may never arrive. It’s building internal frameworks that treat telehealth delivery, prescribing compliance, and AI governance as connected systems — with clear clinical accountability at every decision point.

The Legal Accountability Gap No One’s Discussing

Medical groups carry legal responsibility for every tool they deploy, including AI. But as healthcare AI adoption climbed from roughly 5% to 8.3% between 2023 and 2025, the conversations about implementation focused almost entirely on efficiency gains, not liability exposure. AMA President Bruce A. Scott, MD warned in last year’s AMA webinar that “voluntary standards are not going to be enough — we need to make sure that the principles of AI implementation are regulated.”

Here’s what most organizations miss: Medical groups are legally accountable for harm caused by any tool they deploy, including AI. Yet purchasing decisions get made based on competitor pressure and vendor promises about efficiency gains, not malpractice exposure or what happens when AI flags a high-risk patient who never gets reached.

The organizations seeing real returns understand where AI actually works. Sharp HealthCare reported work RVU increases of 3.5% to 6% per encounter using ambient AI for clinical documentation, capturing more complete records of care that physicians were already providing. This improved both billing accuracy and care continuity while supporting physician judgment.

AI delivers value for administrative tasks, like scheduling, resource allocation, and documentation. But when it moves into clinical decision-making or crisis response, risk increases sharply. Our survey found that 49% of Americans would use AI to monitor their mental health, but 73% want humans to make final care decisions. For now, AI’s value in behavioral health is operational, not clinical.

What Works: Connected Frameworks, Not Isolated Policies

The mistake most organizations make is addressing telehealth reimbursement, DEA compliance, and AI governance as if they’re unrelated problems. They’re not. When reimbursement gets squeezed, organizations turn to AI for efficiency. AI tools generate clinical insights that require documentation. That documentation becomes part of DEA compliance for controlled substance prescribing. Treating these as separate initiatives creates duplicated work and gaps in accountability.

Organizations building successful frameworks keep AI focused on administrative functions, where it delivers proven value without clinical risk. They build systems that explain how decisions get made — something 56% of Iris survey respondents called “extremely important.” They design escalation protocols that satisfy DEA documentation and AI liability standards simultaneously. And they assume policy uncertainty will continue rather than betting on extensions.

Regulatory clarity isn’t coming soon. The organizations scaling responsibly now are the ones building internal frameworks that can handle policy shifts without operational paralysis.

About Andy Flanagan

As CEO, Andy Flanagan is responsible for Iris Telehealth’s strategic direction, operational excellence, and the cultural success of the company. With significant experience in all aspects of our U.S. and global healthcare system, Andy is focused on the success of the patients and clinicians Iris Telehealth serves to improve people’s lives. Andy has worked in some of the largest global companies and led multiple high-growth businesses, providing a unique perspective on the behavioral health challenges in our world. Andy holds a Master of Science in Health Informatics from the Feinberg School of Medicine, Northwestern University, and a Bachelor of Science from the University of Nevada, Reno. His prior experiences include being a three-time CEO, including founding a SaaS company and holding senior-level positions at Siemens Healthcare, SAP, and Xerox.

The following is a guest article by Andy Flanagan, CEO at Iris Telehealth

Behavioral health providers are expanding access to care while navigating three regulatory uncertainties: When will DEA prescribing rules become permanent instead of getting last-minute extensions? Will telehealth reimbursement parity return after enforcement stopped? And who’s legally responsible when AI flags a patient in crisis, but no one intervenes in time?

States passed more than 250 AI health bills in 2024 while federal guidance remains voluntary. Critical telehealth policy decisions in 2026 will determine whether virtual behavioral health stays financially viable. Organizations like Providence are already deploying virtual nursing and AI triage tools in emergency departments, but most providers still don’t know what rules they’re operating under — or what liability they’re assuming.

The answer isn’t waiting for policy clarity that may never arrive. It’s building internal frameworks that treat telehealth delivery, prescribing compliance, and AI governance as connected systems — with clear clinical accountability at every decision point.

The Legal Accountability Gap No One’s Discussing

Medical groups carry legal responsibility for every tool they deploy, including AI. But as healthcare AI adoption climbed from roughly 5% to 8.3% between 2023 and 2025, the conversations about implementation focused almost entirely on efficiency gains, not liability exposure. AMA President Bruce A. Scott, MD warned in last year’s AMA webinar that “voluntary standards are not going to be enough — we need to make sure that the principles of AI implementation are regulated.”

Here’s what most organizations miss: Medical groups are legally accountable for harm caused by any tool they deploy, including AI. Yet purchasing decisions get made based on competitor pressure and vendor promises about efficiency gains, not malpractice exposure or what happens when AI flags a high-risk patient who never gets reached.

The organizations seeing real returns understand where AI actually works. Sharp HealthCare reported work RVU increases of 3.5% to 6% per encounter using ambient AI for clinical documentation, capturing more complete records of care that physicians were already providing. This improved both billing accuracy and care continuity while supporting physician judgment.

AI delivers value for administrative tasks, like scheduling, resource allocation, and documentation. But when it moves into clinical decision-making or crisis response, risk increases sharply. Our survey found that 49% of Americans would use AI to monitor their mental health, but 73% want humans to make final care decisions. For now, AI’s value in behavioral health is operational, not clinical.

What Works: Connected Frameworks, Not Isolated Policies

The mistake most organizations make is addressing telehealth reimbursement, DEA compliance, and AI governance as if they’re unrelated problems. They’re not. When reimbursement gets squeezed, organizations turn to AI for efficiency. AI tools generate clinical insights that require documentation. That documentation becomes part of DEA compliance for controlled substance prescribing. Treating these as separate initiatives creates duplicated work and gaps in accountability.

Organizations building successful frameworks keep AI focused on administrative functions, where it delivers proven value without clinical risk. They build systems that explain how decisions get made — something 56% of Iris survey respondents called “extremely important.” They design escalation protocols that satisfy DEA documentation and AI liability standards simultaneously. And they assume policy uncertainty will continue rather than betting on extensions.

Regulatory clarity isn’t coming soon. The organizations scaling responsibly now are the ones building internal frameworks that can handle policy shifts without operational paralysis.

About Andy Flanagan

As CEO, Andy Flanagan is responsible for Iris Telehealth’s strategic direction, operational excellence, and the cultural success of the company. With significant experience in all aspects of our U.S. and global healthcare system, Andy is focused on the success of the patients and clinicians Iris Telehealth serves to improve people’s lives. Andy has worked in some of the largest global companies and led multiple high-growth businesses, providing a unique perspective on the behavioral health challenges in our world. Andy holds a Master of Science in Health Informatics from the Feinberg School of Medicine, Northwestern University, and a Bachelor of Science from the University of Nevada, Reno. His prior experiences include being a three-time CEO, including founding a SaaS company and holding senior-level positions at Siemens Healthcare, SAP, and Xerox.