Rewiring Cost Containment in Employer Healthcare

Employer healthcare spend is at an all-time high—up ~50% since 2017—despite tens of billions invested in digital health. Over the past decade, large venture-backed companies have been built in navigation, MSK, metabolic, and behavioral health. More than $60 billion has flowed into the category since 2015. Hinge Health’s IPO this year, alongside scaled players like Maven Clinic, Included Health, and Omada, shows how much capital and talent have poured into attacking high-cost categories. Yet, even with public market liquidity for some of these companies, costs for employers continue to rise, and member engagement remains low.

We believe AI represents the biggest opportunity we’ve seen in over a decade to break this cycle. The rise of AI-native infrastructure changes the cost equation, the engagement equation, and the speed at which new models can scale. For the first time, it’s possible to unify fragmented point solutions, personalize the member experience at scale, and deliver true navigation and cost containment tools at a price point that works for the mid-market—not just the Fortune 100. GenAI and LLMs reduce service costs, enable real-time data orchestration, and power incentives tailored to the individual. The result: a new system architecture that engages people earlier and shifts behavior in ways that were cost-prohibitive just a few years ago.

Our belief is that a large part of the issue has been structural—the explosion of point solutions has created siloed member experiences. Companies focus on engaging with patients in their specific swim lanes and do little to think about a patient’s holistic journey or shift their overall health journey. Meanwhile, “navigation” layers often lack the data integration and consumer orientation to deliver durable results.

What we’re hearing in the market

Over the past few months we’ve spoken to over 100 benefits leaders, brokers, operators, and actuaries, and a few consistent themes keep coming up:

The status quo isn’t working

The average employer has 6-10 point solutions with no unified layer to orchestrate them. Employees are confused. Adoption is low. True ROI is rarely there.

Costs are rising fastest, especially in the mid-market

Premiums continue to rise, with many employers reporting 10–20% YoY increases. In some regions, the cost of family coverage is projected to reach ~$50,000/year by 2030—exceeding average wages in many industries. At the same time, spend is highly concentrated: A relatively small set of high-need members can drive tens of millions in annual costs, leaving employers exposed to volatility they can’t predict or plan for. This financial pressure is increasingly pulling CFOs into benefits decisions and amplifying the demand for ROI that is clear, defensible, and immediate.

Broker buy-in is the key to employer adoption

Brokers control access to most employer accounts, and they won’t champion solutions unless the ROI for their clients is clear, defensible, and easy to explain. The offerings that break through will be the ones that deliver measurable savings and engagement without adding operational drag.

How AI changes this

From studying the largest first-generation employer-market companies, three constraints stand out: engagement is the holy grail, human service teams drive high costs, and data is siloed. GenAI tackles all three:

Engagement

AI-native navigation personalizes outreach across chat, text, and voice, sustaining member interaction over time

Human cost

Conversational AI resolves most routine questions and routes members to the right care, freeing staff for complex cases

Data silos

Real-time data layering unifies claims, labs, and SDoH, enabling early steerage and targeted incentives

These shifts make it possible for a $2-3 PEPM product to match the impact of $15+ PEPM service-heavy platforms, unlocking mid-market and SMB segments that were previously out of reach.

Rewriting the Cost Containment Playbook

If V1 was defined by fragmented point solutions, V2 will be defined by solving navigation. Navigation is the key unlock for cost containment—and for the first time, it’s possible to do so at scale and at a cost point that works beyond the F100.

The goal isn’t to invent another point solution, but rather to create a system architecture that connects the existing ones and intervenes earlier in the patient journey. The essential elements are:

• A centralized AI interface as the member’s front door to benefits and care decisions
• Steerage to high-value providers and contracted bundles
• Configurable plan design levers like waived co-pays or financial nudges to guide behavior in real-time

The individual tools aren’t new, but the infrastructure and cost structure that supports them is. AI should reduce the marginal cost of navigation to close to zero, allowing for scalable personalization and earlier intervention where traditional models struggled to show savings.

Open Questions We’re Exploring

• What does it take for AI to be trusted as a front door for high-stakes navigation?
• For a mid-sized employer, what is the minimum operational, technical, and organizational setup required to actually get a positive ROI from AI-driven navigation / cost-containment?
• Can this replace today’s bloated benefits stack or does it end up becoming just another layer?

Why We’re Spending Time Here

This feels less like a new trend and more like a second chance to get cost containment right – especially for the part of the market that’s been neglected and underserved. Lower delivery costs, flexible plan design, and smarter back- and front-end infrastructure offer a path beyond fragmented point solutions toward something integrated and effective.

We’re continuing to explore and would welcome conversations with others thinking about this space!

Reach out to sam@primary.vc or hannah@primary.vc