Onyx Blog
What Black Book’s Payer Interoperability KPIs Tell Health Plans About Analytics, AI Governance and Value Realization
Black Book’s 2026 payer interoperability research evaluates vendors across 18 key performance indicators. In this series, I’m looking at those KPIs in five practical areas through the lens of what it actually takes to build, operate and scale payer interoperability infrastructure.
In Part 1, Building for Evolving Mandates, I looked at the regulatory, API and standards foundation. In Part 2, Making Clinical Data Work in Production, I focused on clinical data acquisition, normalization, workflow integration and provider activation. In Part 3, Reliability, Governance and Production Operations, I looked at what it takes to keep that infrastructure secure, resilient and supportable as it scales.
The next three KPIs shift the focus from moving and managing data to what health plans can actually do with it:
Together, they raise an important question:
Once the data is connected, governed and available, what value are you actually getting from it?
Interoperability creates a significant amount of operational information. API transactions succeed or fail. Data arrives from different sources with varying levels of completeness and quality. Providers and applications connect at different rates. Prior authorization transactions move through multiple steps. Exceptions occur and processing times vary.
All of that produces signals about how the interoperability environment is performing. The challenge is making those signals visible and useful.
A health plan should be able to understand how much data is moving, which connections are succeeding, where transactions are failing, how quickly information becomes available, which providers or applications are actively using APIs and where exceptions are occurring — without manually assembling that picture from multiple systems.
But operational visibility is only part of the requirement. Payers also need evidence. CMS requirements introduce reporting, audit and compliance obligations. When a health plan needs to demonstrate what happened, it needs more than a dashboard. It needs traceable records showing transactions, decisions, timestamps, status and exceptions.
That is why analytics and audit-ready reporting belong together. The architecture should capture operational evidence as part of the transaction itself rather than requiring teams to reconstruct it later.
Healthcare organizations already have enormous amounts of clinically useful information trapped in notes, summaries, reports and other unstructured records. The challenge isn’t simply extracting or summarizing that information. It is determining what is clinically relevant, connecting it to the right member and context, preserving provenance, and making the result usable by a specific downstream workflow.
A risk adjustment workflow, for example, may need to identify clinical evidence relevant to a suspected condition and help a reviewer locate the supporting documentation. Quality may need evidence related to a measure gap. Care management may need information buried in the longitudinal record that changes what a care team needs to know.
These are very different problems from asking a general-purpose model to read a document. They also depend on the data foundation underneath the AI. Identity has to be resolved, terminology normalized, source information preserved and results tied back to the underlying clinical evidence. The more consequential the workflow, the more important those controls become.
That is why Black Book pairs AI-enabled clinical document intelligence with automation governance. Health plans need to understand not only whether an AI system can produce a useful result, but what supported that result, where human review is required, how exceptions are handled and how the process can be audited.
The objective should not be automation for its own sake. It should be using AI to reduce repetitive work, surface relevant information faster and focus human attention where judgment is actually required. That means extending the same principles we apply to healthcare data — provenance, traceability, access controls, monitoring and auditability — to the intelligence created from it.
The final KPI in this group provides a useful test of the other two: Did any of this actually make the work better?
API availability, transaction volume, connected endpoints and data ingested are important operational measures, but they are not the same as value. Value becomes clearer when better data and intelligence reduce the effort required to complete real payer workflows.
That might mean a reviewer spends less time searching clinical documents, better clinical data reduces manual reconciliation, information reaches a workflow sooner, or teams investigate fewer exceptions. It can also mean that a health plan can launch a new use case using data it already acquires rather than building another ingestion and normalization pipeline.
This changes the economics of interoperability. If clinical data acquired for one requirement can also support quality, risk adjustment, care management or other workflows, the value of the underlying investment expands. The same is true of identity, normalization, provenance, analytics and governance capabilities.
The question therefore isn’t simply how much data a health plan has connected. It’s how much additional work that foundation can support without rebuilding the infrastructure underneath it.
This is where the progression across the Black Book KPIs comes together.
APIs create access to data, but access alone is not enough. The data has to be made trustworthy and reusable, the production environment has to keep it governed and observable, and analytics and intelligence have to make that information useful in the workflows that matter.
Within OnyxOS, that means acquiring healthcare data once, normalizing and organizing it into a reusable foundation, and making that foundation available across compliance and business use cases. It also means applying analytics and intelligence on top of the same governed data rather than requiring every new use case to rebuild identity, acquisition, normalization and provenance.
Those are platform problems. Solve them once, and teams can spend more time solving the business problem.
As health plans move beyond basic connectivity, I would ask:
These questions help distinguish having connected data from creating value with connected data.
AI isn’t the destination of a modern healthcare data strategy. Better work is.
That means faster access to useful information, fewer manual steps, better visibility, stronger evidence and measurable improvement in the workflows that matter. AI can help get us there, but its usefulness depends on the quality of the data underneath it, the governance around it and our ability to measure the result.
That is why these three Black Book KPIs belong together: measure what is happening, apply intelligence with evidence and governance, and determine whether it actually changed the work.
Get a clear view of where your organization stands today and what to prioritize as you prepare for the next wave of CMS mandates.
Download the 2026 Black Book Payer Interoperability Report to see the full 18-KPI framework and comparative vendor findings.
View the Black Book Report
