Hear from leaders across Medicaid, technology, and policy as they build trust, transparency, and accountability into the future of AI-driven Prior Authorization (PA). This on-demand session brings together real-world experience, federal policy insight, and modern AI with Human-in-the-Lead (HITL) principles — all in one practical and thought-provoking discussion.
What you'll learn:
- Federal & State AI Landscape:
Understand how CMS and MACPAC are shaping AI adoption in Medicaid PA, including privacy, oversight, and policy guardrails. - Washington State Health Care Authority's AI Readiness:
Gain insight into the Washington State Health Care Authority's AI strategy, including how they identify low-risk use cases for implementation. - AI in Action:
See how Acentra Health uses generative AI methodologies and clinical automation tools to support HITL decisions to accelerate better outcomes. Understand the impact of policy changes, claims coordination, and the evolving Medicaid landscape under the new administration.
Participants
- Mike Barabe, Deputy Chief Information Officer, Washington State Health Authority
- Sean Harrison, Chief AI Officer, Acentra Health
- Verlon Johnson, Chief Government & Corporate Affairs Officer, Acentra Health and Chair, MACPAC (Medicaid and CHIP Payment and Access Commission)
- Katherine Rogers, MPH, PhD, Deputy Director, MACPAC
Read the transcript
Note: This is a polished transcript of the full session and is not intended to be a verbatim record.
Opening and policy context
[Approx. 00:00] Verlon Johnson opens the session by framing the central issue: how Medicaid programs can use AI in prior authorization in ways that build trust, improve transparency, and maintain accountability.
[Approx. 03:30] The discussion then turns to the federal and state policy environment. Dr. Katherine Rogers highlights the importance of guardrails, oversight, and clear governance as AI adoption moves forward in Medicaid.
State readiness and responsible adoption
[Approx. 10:00] Mike Barabe describes how Washington State is approaching AI readiness by starting with lower-risk use cases and focusing on operational discipline, governance, and measurable outcomes.
[Approx. 12:30] The panel emphasizes that readiness is not only about technology. It also depends on internal alignment, strong oversight, and confidence that AI-supported workflows can be introduced responsibly.
Human-in-the-loop AI in action
[Approx. 15:00] Sean Harrison explains how generative AI and clinical automation can help organize information, reduce repetitive manual work, and support better workflow management in prior authorization.
[Approx. 18:30] A key point throughout the discussion is that AI should support expert judgment, not replace it. Human-in-the-loop design remains essential, especially in complex or clinically sensitive cases.
Reducing provider burden while preserving rigor
[Approx. 22:00] The panel discusses provider burden and the challenges created by fragmented documentation, inconsistent requirements, and manual review processes. AI-supported workflows can help reduce unnecessary administrative effort and create a more predictable provider experience.
[Approx. 25:00] At the same time, the conversation makes clear that efficiency must be balanced with compliance, quality, and sound clinical review.
Transparency, data, and implementation
[Approx. 27:00] The panel reinforces that transparency and explainability are critical. Teams need to understand how AI-supported outputs are generated, reviewed, and monitored over time.
[Approx. 30:00] Sean Harrison and Mike Barabe also emphasize the importance of data quality and interoperability. AI works best when the underlying data is reliable and systems can exchange information effectively.
[Approx. 33:00] The implementation guidance is practical: begin with governance, define clear use cases, set measurable goals, and treat AI adoption as an ongoing operational capability rather than a one-time launch.
Closing takeaway
[Approx. 33:30] In closing, Verlon Johnson returns to the central message of the session: trust in AI-driven prior authorization is built through thoughtful implementation, explainable workflows, strong oversight, and consistent collaboration across policy, clinical, operational, and technology teams.