Industry: Healthcare and Behavioral Health
Make improvement stick in hospitals, clinics, and behavioral health
Most of the organizations we work with are not large health systems. They are critical access hospitals, FQHCs, community mental health centers (CMHCs), ambulatory clinics, post-acute providers, and community health nonprofits. Margins are thin, the quality director wears three hats, and every improvement project lands on staff who are already carrying the last one.
That is why we start with the people side of quality improvement, not with technology. A sound process fix still fades when the care team drifts back to the old way. Change management is how you keep that from happening.
Start with the people side of quality improvement
Our Quality Improvement modules layer change management onto the improvement method your teams already use, whether that is PDSA cycles, a Lean Six Sigma belt program, or both. Nothing in your method changes. Each module adds the sponsorship, communication, training, resistance, and reinforcement tools that belong in that phase of the work. Then we apply them with you on your live project.
Why AdoptionLab.AI in healthcare
Founder Matt Humer has coached QI teams at ten New Hampshire nonprofits and led change management at a Fortune 150 company.
The method comes from clinical quality improvement practice, not from AI.
When AI belongs in the picture
AI is optional here. Some organizations bring it in after the people side is working; some never do. When it fits, the same change management discipline is what makes it stick, and the highest-value uses are usually the least glamorous:
- Clinical documentation. Visit notes, progress notes, and treatment plan drafts, with HIPAA-eligible tooling and a clear human review standard.
- Prior authorization and appeals. Drafting appeals, summarizing denials, and navigating payor language.
- Quality reporting and grant work. HEDIS, UDS, state reporting, grant narratives, and board reporting.
- Internal operations. Policy drafts, training content, and board memos: a lower-risk place to build fluency first.
The compliance shape of healthcare AI
- HIPAA and BAAs. Any AI tool that touches PHI needs a Business Associate Agreement. The list of vendors offering BAAs has grown: Microsoft Copilot for Microsoft 365 (with the right tenant configuration), AWS Bedrock, Azure OpenAI, several specialized clinical AI vendors. The list of consumer tools that do not offer BAAs has also grown, and using those tools with PHI is a HIPAA violation.
- 42 CFR Part 2. Behavioral health and substance use treatment data carries additional federal protections beyond HIPAA. AI tools that are HIPAA-eligible are not automatically Part 2-eligible. The policy treatment of these data classes needs to be explicit.
- State law. Several states have adopted or are adopting AI-specific healthcare regulations (Colorado, California, and others). The policy needs to be regionalized for the states you operate in.
- Clinical decision support. AI used in clinical decision-making has additional FDA implications depending on how it is positioned. Most generative AI uses we recommend stay clearly outside CDS (the AI drafts, the clinician decides), which keeps the regulatory shape simpler.
How engagements look in healthcare
Most start with the Quality Improvement modules for a PDSA team or a belt cohort. Organizations that are ready for AI can use the same methods to run a small pilot, such as documentation or prior authorization. GenAI+ Belts™ fit clinical and operational leaders who want to build the skill set themselves.
Something to share with your CEO? Read the one-page overview →
Let's chat about your organization
Tell us about your care setting, your improvement program, and the workflows that matter. Thirty minutes, no slides, no obligation.
Let's chatOr copy the address: mhumer@adoptionlab.ai