From Bot vs Bot to Collaborative AI: Why the Revenue Cycle Arms Race Is a Dead End

The healthcare revenue cycle AI landscape is shifting from adversarial "provider AI vs payer AI" to collaborative models where both sides use AI to reduce friction, accelerate payments, and improve patient experience. In 2026, this shift is being driven by CMS transparency mandates, HFMA advocacy, and platform companies building interoperable AI systems that benefit both payers and providers simultaneously.

Ten days ago, HFMA 2026 opened with a stark warning. BJC Healthcare CFO Scott Hawig described the industry's AI trajectory in three words: "Bot versus bot is the fundamental problem." This week, the narrative is already shifting. R1's CEO told MedCity News on June 18 that payers and providers are "starting to show more willingness to work together on improving the payments process" through AI. The adversarial era isn't just unsustainable — the industry is already moving past it.

The Bot vs Bot Dead End: Why Adversarial AI Fails Both Sides

The adversarial model works like this: payers deploy AI to accelerate claim reviews, flag potential overpayments, and issue automated denials. Providers respond by deploying AI to submit more aggressive appeals, detect denial patterns, and resubmit claims with optimized documentation. Payers counter with more sophisticated denial algorithms. Providers counter with more aggressive appeal engines. Both sides spend more. Neither side wins.

The numbers tell the story. Kodiak Solutions reports a 25% increase in net revenue leakage at hospitals driven by this dynamic — payer AI denying faster than provider AI can appeal. The cost isn't just financial. Every denial-and-appeal cycle consumes administrative resources on both sides: payer staff reviewing appeals, provider staff researching denials, both sides generating documentation that exists only to justify a position in a dispute that shouldn't have happened.

The fundamental problem with adversarial AI is that it treats revenue cycle as a zero-sum game. Every dollar a payer denies is a dollar the provider loses, and vice versa. But the revenue cycle isn't zero-sum — it's negative-sum when friction is high. Administrative costs eat into both payer margins and provider revenue. The American Medical Association estimates that prior authorization alone costs the U.S. healthcare system $34 billion annually in administrative overhead, with costs split roughly evenly between payers and providers.

When both sides deploy AI to fight harder over the same claims, total system costs increase while total system value stays flat. The AI doesn't create revenue — it just shifts which side absorbs the friction cost. That is why BJC Healthcare's CFO called it a dead end. And that is why the industry is pivoting.

The Shift: Why Collaboration Is Emerging Now

Three forces are converging to push the industry from adversarial to collaborative AI in 2026.

1. CMS Interoperability Mandates Create Common Ground

The CMS-0057 rule — effective January 2027 with adoption accelerating now — mandates electronic prior authorization, requires 72-hour urgent and 7-day standard PA response times, and forces payers to disclose specific reasons for every AI-assisted denial. These mandates create common data standards that both payer and provider AI systems must operate on.

When both sides share the same data format (FHIR R4), the same response timelines, and the same transparency requirements, the incentive structure flips. It becomes cheaper to validate a claim correctly the first time than to deny it and process the appeal. Payer AI that works with provider AI to ensure clean submissions saves both sides money. Adversarial AI that generates denials now creates a paper trail of specific disclosure requirements that makes automated denials more expensive to defend.

2. HFMA 2026 Signals Industry-Wide Readiness

Chief Healthcare Executive's June 16 summary of HFMA 2026 highlighted "affordability, AI, and surprising optimism" — with a key insight that progressive health systems are focused on finding recoverable revenue "before it ages, is reduced, or is written off." This language signals a shift from reactive (fight denials after they happen) to proactive (prevent revenue loss upstream).

The "surprising optimism" is notable. HFMA conferences have trended pessimistic for years as payer-provider tension escalated. The 2026 shift toward optimism reflects a growing recognition that collaborative AI creates a better outcome for both sides than continuing the arms race.

3. Platform Companies Are Building the Infrastructure

AGS Health launched InnovationWorks on June 18 with an explicit mission: "Turn the promise of Revenue Cycle AI and Automation into Outcomes That Matter to Providers." The emphasis on outcomes over automation reflects the market's maturation from "deploy AI to do things faster" to "deploy AI to achieve better results for all stakeholders."

Omega Healthcare's recognition as the only Star Performer in Everest Group's 2026 RCM Intelligent Operations PEAK Matrix validates that collaborative, outcome-oriented AI operations are becoming the benchmark. NTT DATA's research (June 17) puts concrete numbers on the opportunity: AI-native and agentic operations can achieve up to 35% cost savings through optimization — savings that come from eliminating friction, not winning disputes.

What Collaborative AI Actually Looks Like

The collaborative model replaces the adversarial cycle (submit → deny → appeal → counter-deny) with a friction-reduction approach where AI prevents disputes from occurring. Here's what that looks like across key revenue cycle workflows:

Real-Time Eligibility That Prevents Claim Rejections

AI eligibility verification in the collaborative model doesn't just check whether a patient has coverage — it confirms the specific benefit details, deductible status, and authorization requirements for the planned services before the encounter. When provider AI and payer systems exchange structured eligibility data through FHIR APIs, both sides benefit: the provider submits a clean claim, and the payer processes it without the cost of a denial-and-appeal cycle.

AI Prior Authorization That Submits Complete Documentation Upfront

The adversarial approach to prior authorization submits the minimum documentation and fights if it gets denied. The collaborative approach uses AI to analyze payer-specific medical necessity criteria, compile comprehensive clinical documentation, and submit a complete authorization request that meets the payer's decision requirements on the first attempt.

The result: fewer denials, fewer appeals, faster approvals. The provider gets the authorization in hours instead of weeks. The payer spends less on adjudication. The patient gets treated sooner. This is not theoretical — Healthcare IT Today reported on June 17 that payers are modernizing operations with AI and automation specifically to "reduce labor gaps through collaborative automation, not adversarial replacement."

Predictive Denial Prevention That Catches Issues Before Submission

AI denial management in the collaborative model shifts from post-denial response to pre-submission prevention. AI analyzes historical denial patterns, payer-specific rules, and claim characteristics to flag potential issues before the claim is submitted. The provider fixes the issue. The claim goes through clean. The payer doesn't have to deny it. The appeal never happens.

This is where the NTT DATA 35% cost savings figure comes from. When AI prevents the denial-appeal-resubmission cycle from starting, both sides save the administrative cost of processing the dispute — and the patient doesn't experience delayed care or surprise bills from unresolved claims.

Shared Data Standards Through CMS Interoperability

CMS interoperability rules require payers to expose Patient Access APIs, Provider Access APIs, and Payer-to-Payer data exchange. These APIs create a shared data layer that collaborative AI can operate on. Provider AI queries real-time benefit information through standardized APIs instead of scraping payer portals. Payer AI receives structured claims data instead of parsing faxed documents. Both sides operate on the same information, reducing the mismatches that drive most denials.

Adversarial vs Collaborative: The Numbers

Why This Matters for Individual Practices

The collaborative AI shift isn't just a health system play. Individual medical practices benefit disproportionately because they lack the negotiating leverage and administrative infrastructure to win the adversarial game.

A 5-physician ENT practice doesn't have the resources to build an AI denial-fighting machine that matches UnitedHealthcare's claim review algorithms. But it can deploy AI that submits clean claims with complete documentation, verifies eligibility in real time, and meets every payer's authorization requirements on the first submission. The collaborative approach doesn't require matching the payer's AI firepower — it requires working with payer systems rather than against them.

This is the strategic insight that separates collaborative AI from the arms race narrative. Adversarial AI favors the side with more resources — larger health systems that can out-invest payers in denial management technology. Collaborative AI favors the side that gets it right the first time — which any practice can do with the right tools, regardless of size.

BAM AI's Collaborative Approach

BAM AI's AI agents for healthcare revenue cycle are built for the collaborative model from the ground up:

• Pre-submission claim validation — AI checks every claim against payer-specific rules, coding guidelines, and authorization requirements before submission. Issues are caught and fixed in seconds, not discovered through denial notices weeks later.
• Complete documentation assembly — AI compiles clinical documentation that meets medical necessity criteria for each payer, reducing denials from insufficient documentation to near zero.
• Real-time eligibility and benefit verification — AI confirms coverage details, deductible status, and authorization requirements through electronic channels, eliminating the portal-navigation and phone-call friction that consumes billing staff hours.
• CMS compliance built in — All workflows align with CMS-0057 interoperability mandates, FHIR-based data exchange requirements, and electronic PA standards — positioning practices for January 2027 compliance while capturing immediate efficiency gains.
• Purpose-built for individual practices — Not enterprise-only technology scaled down. AI agents designed for the volume, workflow, and payer mix of individual specialty practices.

The Transition Is Happening Now

The shift from adversarial to collaborative AI is not a prediction — it's an observable trend with data points accumulating weekly. R1's CEO confirmed the willingness shift on June 18. HFMA 2026 registered surprising optimism about collaborative approaches. AGS Health launched a platform explicitly focused on collaborative outcomes. Omega Healthcare earned Star Performer status for intelligent operations. CMS mandates are creating the interoperability infrastructure that makes collaboration technically feasible.

Practices that continue investing in adversarial AI — better denial fighters, more aggressive appeal engines — are optimizing for a game that the industry is abandoning. Practices that shift to collaborative AI — clean first-pass claims, complete upfront documentation, real-time eligibility — are positioning for the model that R1, HFMA, and CMS are all converging toward.

The bot-vs-bot era had a good run. It's ending because it had to. The question for every practice is whether they'll make the shift before or after their competitors do.