BY J.D. Brewer, Principal, Advisory, C&O Health and Government, and Saurabh Goyal, Principal, Advisory, Health and Government, KPMG LLC
AI as the new architecture of payer value
Healthcare payers are operating in one of the most difficult financial environments the industry has faced in years.
Medical costs continue to climb, regulatory constraints are intensifying, and administrative margins remain under constant pressure. To keep pace, most organizations have followed a familiar playbook: digitizing processes, automating transactions, and outsourcing some operations in a tireless pursuit of efficiency.
Yet the cost-to-serve remains stubbornly high. And, for many organizations, the ability to actually deliver transformative change is trapped in a Groundhog Day cycle of big-promise technology investments and underwhelming results.
Artificial intelligence (AI) offers healthcare payers a chance to break that cycle. Unlike earlier waves of automation, AI has the potential to reshape how they make decisions, improve operations, and elevate their interactions with members and providers.
But a familiar concern lingers. Because while investment in AI is widespread, the results have been uneven. Healthcare payers are testing and deploying AI capabilities across the business—from predictive models to worker productivity—yet too often these are targeted pilots that are focused on improving legacy processes.
To move from Groundhog Day to groundbreaking change, payers must fundamentally rethink how work happens across the enterprise. Translating AI’s potential into measurable value requires a commitment to structural change that makes AI integral to the overall operating model: embedding intelligence directly into decision cycles, integrating fragmented data environments, and governing AI securely so it can operate with trust at scale.
Setting up AI to deliver value
The default approach for many organizations—in healthcare and beyond—is to layer AI capabilities onto existing systems and workflows. While adding AI to legacy steps may increase efficiency at the margins, true transformational value demands a broader redesign of how work happens across the enterprise. But, across industries, many companies are struggling to make this leap, with just 10 percent reporting that their technology implementations are “fully scaled”—a decrease from previous years.¹
If scaling is difficult universally, healthcare payers have an even narrower needle to thread. Many of their operational workflows—from claims processing to utilization management to member outreach—were designed for a different technological era. They rely heavily on manual review, fragmented data sources, and sequential decision processes. Payers cannot enhance these outdated methods through automation alone; they must reimagine end-to-end processes while navigating industry-specific constraints.
Unlocking measurable value requires addressing the critical pressures facing the payer business:
Updating the operating model: AI must operate directly within operational workflows rather than acting as a separate analytical layer. Intelligent systems can assist decisions in real time by summarizing clinical documentation, identifying patterns in claims, flagging anomalies, or recommending the next best action for member engagement. The goal is to let AI actively orchestrate work.
Modernizing data: Many AI initiatives stall because of concerns over systemic data issues and the weight of legacy tech debt, with 56 percent of companies citing this as a primary barrier to AI-driven transformation.² But moving to “business-ready” data does not require a multi-year rebuild to fix every data lake. It requires modernized, interoperable architectures that allow AI to reach across fragmented systems and extract insights on the fly.
Keeping pace with the industry: Providers spend 30 percent of their time on prior authorization tasks.³ To combat this, they are using AI to help optimize documentation and billing. While this dynamic has been described as a “battle of the bots,” the true opportunity for payers is not about outmaneuvering providers. It is about looking inward. By adopting equally sophisticated intelligence to streamline their own operations, payers can achieve significant administrative savings. Even a conservative 10 percent improvement in processing efficiency translates into massive value.
Managing new risks: Unlike other sectors, a payer’s “cost of being wrong” extends to clinical outcomes and legal compliance. With only 31 percent of payers reporting a fully defined AI governance model,⁴ the path to scale requires a Trusted AI framework that prioritizes explainability, bias monitoring, and cyber resilience.
A common goal across all of these areas is to shift from “AI Roulette”—scattered bets on point solutions—to a disciplined architecture of value. By solving for these foundational hurdles, leaders can move to an operating model that orchestrates work intelligently across the front, middle, and back office.
Moving to “business-ready” data does not require a multi-year rebuild to fix every data lake. It requires modernized, interoperable architectures that allow AI to reach across fragmented systems and extract insights on the fly.
Intelligent transformation: Three actionable arenas for AI
Transitioning to an intelligent operating model requires shifting focus from isolated task automation to building adaptive systems. True transformational value is unlocked when AI integrates directly into decision cycles—allowing the organization to sense, decide, and adapt in real time. This transformation is currently unfolding across three high-impact arenas.
Opportunity #1
Experience Intelligence (front office)
Focus: Member and Provider Engagement
Member engagement remains one of the hardest frontiers in healthcare. Health plans consistently struggle to reach members at the right moment and motivate them to participate in preventive care or affordability programs.
Intelligence in Action: Predictive AI and conversational platforms shift outreach from generic campaigns to highly personalized interactions. By analyzing behavioral patterns and historical data, AI can anticipate churn, tailor messaging, and enhance affordability navigation. For providers, AI rebuilds trust by simplifying documentation exchange and easing friction in authorization cycles.
The Value: These experience gains translate directly into brand loyalty, higher retention, and a measurable “brand halo” that strengthens the payer-provider relationship.
Numbers to Know
30 percent
Resolution of routine member inquiries without human intervention when conversational AI and intelligent routing are deployed effectively in healthcare member services.⁵
Opportunity #2
Operational Intelligence (middle office)
Focus: Utilization Management, Claims, and Appeals
Core functions like utilization management (UM) and claims remain heavily cost-intensive and are ripe for intelligent redesign. Organizations are actively using AI to reframe denial management as proactive value management—directing members toward appropriate, evidence-based care more efficiently.
Intelligence in Action: Generative and agentic AI models are currently being deployed to expedite prior-authorization decisions, summarize complex clinical records, and predict denials before they occur. For example, recent AI pilots at regional health plans have successfully reduced manual handling of transactions, drastically improving turnaround times for providers and members alike.
The Value: These middle-office improvements represent one of the most immediate sources of measurable return on investment (ROI). By drastically reducing manual rework and expediting complex clinical reviews, payers can achieve massive cost savings while accelerating the care delivery workflow.
Numbers to Know
500 million
UM reviews are conducted annually across the U.S.—even just a 10 percent improvement in processing efficiency via AI would deliver enormous administrative savings.⁶
Opportunity #3
Decision Intelligence (back office)
Focus: Data Integration and Secure AI Governance
AI’s value depends entirely on data accessibility and reliability. Payers possess abundant information, but it is often fragmented across legacy mainframes, clouds, and line-of-business systems.
Intelligence in Action: Modern architectures—including interoperable clouds, federated analytics, and real-time data exchange—allow AI to act on complex data securely and at scale. This intelligence is operationalized through common industry platforms (such as Oracle, Salesforce, ServiceNow, and Azure), providing the scaffolding to extend and help optimize existing tech investments safely.
The Value: Integrating data must happen alongside strict governance disciplines—explainability, bias monitoring, audit trails, and cyber resilience.
Numbers to Know
$7.42 million
The average cost of a healthcare data breach, making healthcare the costliest industry for data breaches for the 14th consecutive year.⁷
Capturing the ROI: Making AI value measurable
While the opportunities are compelling, organizations without a strong strategy risk losing nearly 40 percent of their expected value to a gap between planned and realized benefits.⁸ The missing discipline in many failing AI programs is value management.
To transform AI from a standard expense line into a continuous performance engine, leaders must apply a rigorous, three-phase value management lifecycle:
1. Articulation
Define clear, quantifiable value objectives for every AI initiative up front, ensuring they align directly with overall organizational goals.
2. Investigation
Design and implement systems to continuously investigate value drivers through live data and feedback loops. Data-driven analysis provides the insights necessary for real-time course correction.
3. Realization
Embed measurement and accountability directly into the governance model. Continuously monitor key performance indicators to help ensure the transformation delivers the expected financial and operational outcomes.
For healthcare payer leadership, disciplined value tracking provides the concrete confidence required to scale AI securely and strategically. And this value analysis becomes especially critical as AI is interlaced into the structural fabric of payer operations. Organizations that design secure, intelligent systems around measurable value will ultimately lead the market on cost, compliance, and experience.
The next era of healthcare competitiveness will belong to the intelligent payers—enterprises that treat AI not just as a tool for automation, but as the architecture of trust and growth. t
References:
1 KPMG. 2026 KPMG US Technology Survey report.
2 KPMG. 2026 KPMG US Technology Survey report.
3 Salzbrenner, S. G., Lydiatt, M., Helding, B., Scheier, L. M., et al. (2023, July). Influence of Prior Authorization Requirements on Provider Clinical Decision-Making. The American Journal of Managed Care, 29(7), 331–337.
4 HealthEdge. (2026, February 27). 2026 Payer Survey: Are Health Plans and Members Aligned on the Future?
5 Gartner. (2024). Predicts 2024: U.S. Healthcare Payers.
6 KPMG. (2024). The Next Frontier in Utilization Management.
7 IBM. (2024). Cost of a Data Breach Report 2025.
8 KPMG. (2024). The Value Management Paradox in Healthcare.