Agentic enterprise: an antidote to the health insurer business model

AI can help payers move from administering benefits to orchestrating better care.

In brief

• Health insurers face a structural squeeze that efficiency alone won’t fix; AI can help redesign how care, costs and decisions are managed.
• The winners will use AI to build a better payer model—one that improves outcomes, lowers medical costs and strengthens retention.
• EY-Parthenon teams work with clients to turn AI ambition into value.

Health insurers are trying to solve a structural problem with incremental patches. Medical costs are rising, regulatory and pricing pressures are tightening, and employers and members want more value for every dollar spent. Growth, meanwhile, is harder to find and defend as plans compete to retain employer relationships and attract profitable members.

The EY-Parthenon CEO Survey 2026 shows that 78% of healthcare CEOs plan to increase AI investment this year, yet only 23% intend to use it to redefine the business model. That gap is significant. Efficiency gains alone will not be enough. The larger opportunity is to use AI to redesign how value is created across the care journey — shifting payers from benefit administrators to orchestrators of better decisions, lower costs, stronger retention and more personalized care.

The case for change for health insurers

The payer model was built for scale and efficiency, but it no longer fits today’s economics. Medical costs are rising faster than administrative savings can offset, exposing the limits of a broad, population-based model in a market that increasingly rewards more personalized, higher-value care.

These pressures point to a deeper issue: This isn’t a temporary squeeze but a challenge to the model itself. If the problem is structural, the response must be as well. AI creates the most value when it changes decisions, not just tasks. Embedded in workflows, it enables earlier risk detection, more proactive care management, smarter prior authorization and real-time guidance.

But even if the direction is clear, the path is constrained by capital, regulation, limited execution capacity and fragmented data. Point solutions will not change the economics. The better path is to sequence a few high-impact journeys, realize value quickly and reinvest it to fund broader transformation.

Many organizations still focus on pushing more volume through legacy structures because it is measurable and defensible. But what looks like a design choice today is quickly becoming a source of market separation. Reinvention requires redesigning workflows, value streams and decision ownership across the enterprise.

The market is splitting

That distinction is already starting to shape competitive outcomes. As early movers reinvent the model, they accumulate proprietary data, embed learning loops and build cost structures that improve with scale. As they reset cost and experience benchmarks, late movers may face higher unit costs, weaker experiences and less pricing flexibility.

The next-generation payer 

What, then, does a structurally advantaged payer look like in practice? It is not simply a benefit administrator with better tools. It is a decisioning and orchestration platform that operates continuously across the member lifecycle and care journey. That shift requires redesigning how insight is generated, decisions are made and actions are triggered across fragmented processes.

Care management journey

In practice, the distinction comes down to how AI is applied: as an overlay on existing work or as part of how the work itself is designed. Most payer AI investments are still bolt-ons that automate steps within existing functions, improving efficiency without changing the operating model. Built-in AI goes further by redesigning the work itself and closing the gap between insight and action.

An agentic health plan model

An agentic health plan is an operating model in which built-in AI agents continuously sense member needs, recommend next best actions and coordinate work across channels so humans can focus on clinical judgment, relationship building and handling exceptions.

At the center of this model is a dynamic view of each member that helps AI drive action, while humans provide judgment and oversight. Unlike traditional care models, it enables continuous risk monitoring, proactive engagement and personalized guidance to reduce friction and help high-cost members navigate care with providers.

AI use cases across the payer value chain

Seen this way, the model is not abstract. Many drivers of medical expense — avoidable admissions, unmanaged chronic disease, medication nonadherence, out-of-network leakage and member friction — are operational problems disguised as actuarial ones. An agentic health plan addresses them earlier and more consistently at scale.

Agentic AI does more than cut costs; it can strengthen the competitive position. We estimate that agentic AI can help generate $85 million to $225 million of value per $1 billion of revenue, along with better experiences, lower medical trends and stronger retention.

Execution depends on data, talent and scale

The opportunity is significant, but few organizations can pursue it all at once. Capital, regulation and limited leadership capacity constrain progress, and smaller plans may fall behind as data, talent and scale advantages compound.

AI only scales on high-quality, interoperable data. Fragmented records, inconsistent coding and disconnected vendors limit impact, regardless of model sophistication. Data investments may not pay off immediately, but they determine whether future AI efforts scale or stall.

With core data integrated, payers can add real-time signals — such as wearables and engagement data — to improve prediction and enable earlier intervention.

Funding the shift 

Transformation needs a practical funding model. The most effective path is self-funding: Early gains create capacity and cash to reinvest in redesign, platforms and global delivery models that add scale, flexibility and efficiency.

The path forward for health insurers

The window is narrowing. Payers likely have two to three years to reposition before structural disadvantage sets in. Leaders are already building governance, modernizing data and redesigning journeys end-to-end to compete on faster, better decisions at lower cost. The market will not wait for consensus. By the time the path feels fully proven, the economics will already have shifted.

Leaders can take a phased approach:

Organizations don’t need another AI strategy deck. They need to decide where to play, what to redesign and how fast to move. The real risk is not moving too early. It’s moving too late.

We would also like to thank our colleagues Katherine Kohatsu, Robert Sim and Joseph Baez for their contributions to this article.