People
This article was originally published in Law360 on June 23, 2025.
Healthcare providers and revenue cycle management vendors face an increasingly challenging environment that is ripe for legal disputes.
For providers, margins are being compressed as labor, supplies and capital costs continue to rise faster than reimbursement rates. Meanwhile, payors are faced with their own economic pressures and have increased their scrutiny of every claim. This scrutiny translates into increasing procedural requirements, expanding the use of artificial intelligence in claims adjudication, and more frequent downcoding of procedures.
Providers often address these challenges by hiring revenue cycle management vendors. These vendors are expected to leverage their specialization, economies of scale, proprietary tools and global labor pool to help providers to receive timely reimbursement for their services.
At a macro level, revenue cycle management vendor performance seems easy to judge. A provider should have clarity around three basic elements:
- What did I do? (Translating clinical notes into diagnosis and procedure codes);
- What did I bill? (Tracking and reporting gross and net claims);
- What did I collect? (Applying cash to claims and calculating the net collections rate).
These figures should all connect — there should be codes that correspond to clinical activity, traceability from the codes to a claim or bill, and a clear accounting of cash receipts associated with individual claims.
When disconnects emerge between metrics that ought to move in tandem (e.g., claims are up and collections are down), or with historical performance (e.g., the net collection rate is lower now than it was last year), it's an indication that something has gone awry.
Unfortunately, when underperformance occurs, the fundamental drivers are not always immediately clear. They could include some combination of:
- Changes in provider behavior;
- Changes in payor behavior;
- Changes in internal vendor systems;
- Changes in how vendor systems integrate.
With multiple plausible-seeming explanations for why revenue has been lost, provider-vendor disagreements can easily spiral into serious legal disputes over millions of dollars in damages. So, in the age of artificial intelligence, how should leaders prepare to resolve minor disagreements quickly or prevail if things escalate? We suggest five approaches that should become standard practice for any provider or vendor leveraging AI in the revenue cycle management services.
- Create tailored, AI-specific service-level agreements — include specific performance standards for AI tools and model performance over time, not just overall process outputs.
- Leverage targeted AI models to flag anomalies and then leverage human expertise to conduct quality assurance and quality control on processes, connect users, and diagnose issues.
- Develop an independent view of revenue cycle management performance and drivers, rather than relying only on counterparty reporting.
- Routinely conduct back tests on models to confirm performance in identifying known historical issues that drove past performance dips — targeted AI models will require consistent maintenance and updating to remain reliable.
- Create issue resolution forums — start flagging concerns early and establishing a good faith effort to resolve them, before interpretations of events begin to diverge.
General counsel for both providers and vendors will have a critical role in the development and delivery of the legal frameworks that support these approaches. Early involvement will pay dividends by avoiding misunderstandings and unnecessary friction as issues emerge and require resolution.
1. Counsel should collaborate with business and technical leaders to develop standard, AI-specific language for contracts and policies.
Rather than assume existing terms cover data, workflows, reporting or service levels, it behooves executive teams to create tailored language that describes what an AI tool will do, how it will do it, what will be measured, and what will be paid. For example, if an AI screening tool ingests all claims, identifies a subset for remediation, and passes them along for action, should volumes be counted on the basis of inputs, outputs or actions?
2. Counsel should support the negotiation of agreements or amendments of existing contractual oversight mechanisms to include special considerations around AI.
AI tools are becoming part of standard workflows, and so will make themselves felt in existing provider-vendor relationships without any explicit discussion of the topic. There is no way to opt out of the age of AI. Thus, spending time with critical counterparties to agree upon and formalize the application of existing contract language to AI creates the space for amicable discussion and resolution before issues arise.
3. Counsel can provide structure and clarity to both internal and external AI governance.
Internally, counsel can collaborate with business leaders to align on how corporate oversight and governance will manage AI, which existing forums support those purposes, and what new bodies and charters might be necessary. Externally, counsel can support the establishment of joint committees and oversight bodies to surface and resolve issues early. This can prevent the occurrence of disputes, generate documentation of issues and demonstrate good faith efforts at resolution.
Regardless of the tools involved, providers want to be reimbursed, payors want to manage medical costs and vendors want to make a profit. Those basic incentives ensure that differences of opinion will continue to occur. Protecting your interests in the age of AI requires anticipating the impact of technology, taking steps today to get ahead of potential disputes, and setting conditions for a favorable resolution.
