During the 2026 AHCA/ADVION Rehabilitation Symposium, AHCA’s Dan Ciolek moderated a panel on using AI to improve function and outcomes in nursing facilities and assisted living. Panelists came from operations, therapy, and the tech space. Their message: the greatest barrier to AI isn’t the technology. It’s governance, workflow, and trust.
Key takeaways:
- Start with the problem, not the product. Identify the resident, clinician, or operational need first, then match the right type of AI to it. Not every “AI” tool carries the same capabilities or the same clinical risk. Panelists added that the frontline staff who live the problem every day should help define it.
- You own the note. AI should support clinicians, not displace responsibility. The clinician remains the author, and using AI doesn’t dilute legal or clinical liability. In discussion, the panel’s bigger worry wasn’t AI errors but over-trust without ownership, like clicking “accept” at 5 p.m. to get out the door. That potential habit is also how notes start to look identical, a red flag for Medicare auditors who expect individualized plans of care.
- Governance and integration come first. The experts highly advise an AI governance committee and a signed BAA before any model touches PHI. Fragmented EHRs remain one of the biggest barriers. Panelists stressed that therapy platforms need to connect with the operator’s EMR, because data without full-chart context runs the risk of being “just a number.”
- Vet, then pilot, then scale. Pilots with measurable outcomes let you evaluate a tool before broad deployment. Good measures include documentation time, clinician satisfaction, resident outcomes, and ROI. Before piloting, have your prospective AI vendor meet your CIO and call their references. Then include skeptics in the pilot, not just champions, and watch for new burdens on staff.
- Measure enterprise value, not just productivity. According to McKinsey, 80% of organizations say AI improved their own productivity, but only 37% can tie it to the bottom line. The difference is workflow redesign. Track clinical outcomes, resident experience, efficiency, and total cost of care. As the panel put it, a few strong metrics beat a crowded dashboard.
If you have questions about where AI fits in your therapy program, please let us know.
The Point Is: Start small, prove value, build trust, and scale responsibly. Instead of “Human-in-the-loop” we should be saying “Clinicians in the driver’s seat.”
