Health plans and provider organizations have adopted AI across value-based care, but the operational discipline behind it, governance, data quality, and audit readiness, has not caught up. When a CMS RADV audit lands, the question is simple and unforgiving: can you defend the risk scores you submitted? In this panel, Reveleer and Mathematica leaders draw on year two of their national research to separate the organizations that can answer that question from the ones that cannot.

‍

What the session covers:

• Why the real gap in value-based care is not the technology, but the workflows and governance that turn tools into defensible outcomes
• What a practical AI governance model looks like across procurement, vendor measurement, and internal readiness
• How data traceability and chain of custody determine whether risk scores hold up under a multi-year audit
• What accountability payers are starting to demand from AI vendors, including performance-based contract terms
• Why funding the data foundation first creates the most leverage as CMS moves toward mandatory models
• How to keep a human in the loop so AI accelerates high-stakes work without creating liability

‍

The conversation gives payer and provider leaders a clear read on where to focus next, from governance ownership to data investment to vendor accountability. Watch the full panel to see how Reveleer approaches the defensibility standard in value-based care.

‍

Defensibility in value-based care FAQs

What is the defensibility standard in value-based care?

The defensibility standard is the ability to substantiate every submitted risk score with traceable evidence, from the source clinical document through the final diagnosis. CMS is moving toward more frequent RADV audits across a wider set of Medicare Advantage contracts, so health plans can no longer treat documentation as something to assemble after an audit notice arrives. Meeting the standard means maintaining a continuous chain of custody for clinical data, so that when an auditor questions a condition, the plan produces the supporting record and the path it traveled rather than reconstructing it under deadline pressure. Reveleer builds this traceability into risk adjustment workflows, so audit readiness becomes a byproduct of daily operations instead of a separate scramble.

‍

Why does AI governance matter for audit readiness?

AI now touches the workflows that determine payment accuracy, including risk adjustment, coding, and quality measurement, so an ungoverned model can introduce errors that surface later as audit findings and repayment risk. AI governance is the set of documented controls that keep those outputs accountable: defined criteria for evaluating vendors, a consistent method for measuring model accuracy, a process for catching errors before submission, and a human reviewer responsible for the final decision. Without governance, a plan cannot explain how a diagnosis was identified or why it was submitted, which is the exact question a RADV auditor asks. Reveleer applies structured oversight across its risk adjustment workflows, so AI assists clinical review while preserving the evidence trail behind every suggestion.

‍

What should payers require from AI vendors in value-based care?

Payers should require three things before signing. First, methodology transparency: the vendor should explain how its model identifies and validates a diagnosis, not simply report that accuracy is high. Second, evidence from real data: a working sample run against the payer's own records demonstrates performance more reliably than a benchmark drawn from a curated dataset. Third, accountability in the contract: performance terms tied to audit and coding outcomes hold the vendor responsible for results, rather than terms tied only to uptime or feature delivery. Reveleer supports this scrutiny by exposing the evidence behind each diagnosis suggestion and maintaining the audit trail payers need to verify accuracy for themselves.

‍

What is the single highest-leverage investment right now?

The data foundation delivers the most leverage. A longitudinal, patient-level record is the input every downstream capability depends on, so investing there first raises the accuracy of risk adjustment, quality measurement, and any AI applied on top of it, while producing the traceable evidence RADV audits require. It also buys optionality: as CMS expands mandatory value-based payment models, plans with a clean, connected data layer adopt new programs without rebuilding their infrastructure each time. Point solutions bolted onto fragmented data produce results that are harder to defend and more expensive to reconcile. Reveleer consolidates clinical data acquisition, risk adjustment, quality improvement, and member management on one data layer, so the foundation supports every workflow instead of splintering across systems.