Becker’s Health IT + Digital Health + RCM Conference reinforced an important shift in the revenue cycle conversation. AI adoption is accelerating, but technology shaping processes itself is no longer the most interesting question.
The bigger issue is what happens around it: how performance is measured, how workflows are rethought and redesigned, how payer-provider dynamics evolve, how suppliers are held accountable and where human expertise creates the greatest value.
Five signals stood out.
1. Measure AI by its impact on the entire revenue cycle, not isolated KPIs
Individual KPIs can tell a misleading story. A lower denial rate means little if recovery costs rise, payments slow or additional manual work appears elsewhere. The same applies to automation: processing something faster (automation) does not necessarily improve the economics of the overall process.
For RCM leaders, this raises the bar for AI investment. Establish the baseline first, then measure performance across financial and operational outcomes, including manual review, exception handling and correction. The question is no longer simply whether AI works. It is whether it improves the economics of RCM as a whole.
2. Redesign the workflow around AI instead of automating the status quo
AI is already moving well beyond pilots. Larger health systems are deploying it across workflows, with encouraging results in areas such as eligibility and coverage checks, prior authorization documentation and transmission, and ambient documentation.
The harder problems are revealing something important. Complex handoffs, unworked EHR queues and gaps between clinical documentation and billing do not disappear when technology is added. Automating a fragmented process can simply make fragmentation move faster.
The opportunity is to redesign work around what AI now makes possible. That means eliminating unnecessary handoffs, clarifying ownership and deciding deliberately where automation ends and human judgment begins. As AI scales, understanding its true unit economics, including technology consumption and the cost of human intervention, will become increasingly important.
3. Get ahead of the payer-provider AI contest by strengthening both prevention and resolution
Both sides are becoming more sophisticated. AI can accelerate prevention, detection and response, but complex denials remain difficult precisely because they involve exceptions, differing payer requirements and judgment.
This creates a dual imperative for providers: prevent more issues before they occur while becoming much better at resolving those that remain. Prior authorization is especially important because friction here can affect not only revenue cycle performance but also how quickly patients move toward treatment, now also subject to CMS regulations.
The goal should not be to automate every dispute. It should be to reduce recurring disputes, strengthen evidence, create more predictable payment and reserve experienced people for cases for exception-handling where their judgment materially changes the outcome.
4. Hold RCM partners accountable for outcomes, not activity or labor
Traditional labor-arbitrage models are losing relevance, particularly among larger health systems. At the same time, leaders are asking harder questions about what suppliers can deliver today, whether critical human expertise will remain available and how technology roadmaps translate into measurable results.
AI also challenges another assumption: that long-term contracts built around relatively static delivery economics will remain fit for purpose.
RCM leaders should increasingly structure supplier relationships around outcomes, including revenue improvement, cost containment, expertise and performance, rather than activity alone. Contracts also need enough flexibility to evolve as AI changes delivery economics. The winning partner will not simply provide capacity. It will be accountable for ever-improving value.
5. Redesign the RCM workforce so automation amplifies human expertise
Health systems have more work than available people, making automation essential. But experience also shows that technology introduced without user involvement, clear ownership and effective controls can create new problems rather than remove old ones.
The workforce question therefore should not be framed simply around how many tasks AI can replace. It is about where people create the most re-thinking and value.
Routine work can increasingly move toward automation, while experienced teams focus on complex exceptions, judgment and intervention. That requires role redesign, trusted data, security involvement, clear accountability and defined thresholds for when people take control. Every automated process also needs a practical way to stop, override or recover when it does not perform as intended.
Don’t just automate today’s revenue cycle. Redesign it for what comes next.
Taken together, these signals point to something bigger than an AI transformation. The operating model of RCM itself is changing.
Technology, workflows, workforce, supplier relationships and commercial models are being reshaped at the same time. For healthcare leaders, the opportunity is not simply to automate today's revenue cycle. It is to design a more measurable, adaptable and accountable model for what comes next. Contact us to learn more.
About the author

James (Jamie) Burke
Partner, Healthcare Lead
Jamie is a Partner and the ISG Healthcare Lead, developing new business and supporting clients’ initiatives and ISG project teams around the world. Jamie has extensive experience in the design, development, implementation and support of strategic business initiatives and information systems in a variety of industries and environments. His consulting experience has ranged from Fortune 100 companies to start-up companies. For the past ten years, Jamie has been on the ground helping healthcare organizations be successful in their digital transformation and cost optimization strategies and execution. LinkedIn Profile