Business Choices for TPAs at the Inflection Point

This document does not make recommendations. It frames choices. The people reading it know their own capital structure, their own team, their own broker relationships, and their own risk tolerance. What follows is a set of structural questions and genuine strategic alternatives, informed by the market data in FWD.01 through FWD.04, designed to make the leadership conversation that follows more honest than it would otherwise be.

The Market You Are In Today

#

The level funded market is substantial and growing. The KFF 2025 Employer Health Benefits Survey found that 37 percent of covered workers at firms with 10 to 199 employees are in level funded plans, a share that has held steady over the past two years. Sixty-seven percent of all covered workers are in self-funded plans of some kind, including 27 percent at small firms and 80 percent at large firms (KFF, “2025 Employer Health Benefits Survey”). Enrollment in the fully insured medical market has dropped nearly 17 percent since 2019 as employers migrate to self-funding (Oliver Wyman, “Stop-Loss Market Update 2025”). The direction of the market is toward employer-borne risk with stop loss protection, not away from it.

The stop loss market that supports level funded is a $39 billion premium market as of 2024, up from $31.6 billion in 2022, reflecting a compound annual growth rate of roughly 12 percent (Oliver Wyman). Those numbers mask a deterioration: loss ratios worsened from 81.6 percent in 2019 to 86.0 percent in 2024. Million-dollar-plus claims are rising steeply, driven by cancer treatments, premature births, specialty pharmacy, and cell and gene therapies. The 2025 Aegis Risk Medical Stop-Loss Premium Survey reported that 49 percent of respondents had experienced a claimant exceeding $1 million in paid claims in the last two policy years (Aegis Risk; International Foundation of Employee Benefit Plans). Several of the largest stop loss writers, including Cigna, Voya, and Sun Life, reported difficult claims experience in the fourth quarter of 2024, which will translate into higher renewal rates for 2026 (IFEBP). The TPA’s operating environment is a growing market with worsening loss economics. Both facts matter.

The structural advantages of the current model, stated honestly: transparency over fully insured means the employer can see their own claims data, which is real and valued but overrated as a differentiator because most employers who say they want transparency do not use it effectively without help. The cost control narrative rests on the surplus return mechanism, but how often this actually happens, and at what magnitude as a percentage of total premium, is data most TPAs do not publish. The honest assessment is that surplus returns are meaningful in good years and nonexistent in bad ones, and the average employer’s experience over a three-year period is closer to breakeven than the marketing suggests. Flexibility, meaning plan design customization through RBP, DPC carve-outs, and custom formulary, is genuine for employers with specific needs and irrelevant for those who accept a standard design.

The structural vulnerabilities, stated without the varnish: stop loss dependency means the TPA’s value proposition depends partly on stop loss carrier relationships that the TPA holds but the employer does not. When the stop loss carrier declines a renewal, raises rates beyond viability, or exits the market, the employer is exposed and the TPA’s ability to retain the account is compromised. Carrier consolidation is accelerating: in January 2025, Nationwide acquired Allstate’s employer stop loss segment for $1.25 billion, and Prudential entered the market in September 2024 with a new stop loss product (Oliver Wyman; market reports). Network leasing means most TPAs do not own their provider networks, leasing access from carriers or aggregators (MultiPlan, PHCS, First Health), which is a structural disadvantage because the same network is available to any competitor that leases it and is also a source of margin compression as aggregators raise access fees. Claims adjudication is commoditizing: AI is making automated adjudication cheaper and more accurate, and the per-claim cost of adjudication processing is falling. The TPA whose value proposition rests primarily on claims processing accuracy is building on eroding ground. Data ownership ambiguity, where who owns the employer’s claims data varies by contract and state law, means a TPA whose retention strategy depends partly on making it hard for employers to leave with their data has a defensible position in the short term and a trust problem in the long term.

The competitive pressure that is not coming from other TPAs: the large carriers have entered the small group self-funded space. Nationwide’s $1.25 billion stop loss acquisition signals intent. Prudential’s market entry signals the same. UnitedHealthcare, Aetna, and Cigna all have distribution, capital, and network ownership advantages that no TPA can match. The question is not whether they will compete more aggressively. It is on what timeline. The insurtech entrants are not TPAs. They are vertically integrated benefits platforms. Angle Health raised $134 million in a Series B round in December 2025, bringing total funding to nearly $200 million. The company reports 26-fold revenue growth since its 2022 funding round, now serving over 3,000 employers across 44 states with an 80 percent renewal rate and a 36 percent reduction in median rate increases compared to the broader small group market (Fortune; Pulse2). Sana Benefits raised $60 million in 2022 and reports that 40 percent of its new customers were small businesses that previously offered no health coverage at all (Business Wire). HR platforms (Rippling, Gusto, Justworks) already have the micro-employer relationship. Benefits is a natural product extension. Their technology infrastructure handles small employer administration at lower marginal cost. They lack benefits depth but are building or acquiring it (FWD.03, Section 4).

The Market That Is Forming

#

Three structural shifts, each documented in the preceding articles, with their implications for TPA strategy stated directly.

Shift 1: The employment unit is shrinking. The 10 to 25 life sweet spot for level funded is being eroded from below by micro-employer formation (FWD.01, FWD.03). A TPA already serving 1 to 10 lives is experiencing this as a growing share of micro-groups in its book. Either the product economics improve at micro-group sizes, through technology investment and pooling with pool-level reinsurance (FWD.03), or the addressable market stagnates in relative terms while the total employer population grows.

Shift 2: The reimbursement model is competing with the group model. ICHRA adoption grew over 1,000 percent from 2020 to 2025 (HRA Council). It is not destroying level funded. It is taking the easiest sales: employers who would have moved to level funded but find ICHRA simpler (FWD.02). The employers remaining in group products are more complex, higher-value, higher-expectation accounts. The bar for TPA capability is rising. The PTC expiration complicates ICHRA’s value proposition for above-subsidy earners but is accelerating ICHRA demand among employers who previously offered no coverage (83 percent of ICHRA adopters in 2025 had not previously offered benefits).

Shift 3: AI is changing the cost structure of administration. Claims adjudication cost per claim is falling. The value of claims data is rising. Human labor is the component most exposed to cost pressure. Member experience (navigation, advocacy, care coordination) is the least automated and most differentiated. FWD.07 provides the Tier 1/2/3 readiness assessment for each business process. The TPA’s value is migrating from processing to intelligence and from administration to navigation.

The Business Model Choices

#

Five genuine alternatives. Each with what it requires, what it assumes, and where it breaks down. No ranking. The reader’s situation determines which is right.

Choice A: Deepen the Core. Level funded administration for 10 to 499 employees. Better technology, stronger broker relationships, geographic expansion, deeper stop loss carrier partnerships. This choice requires capital for the FWD.07 Tier 1 AI deployments (quoting automation, claims anomaly detection, employer dashboards), patience for organic growth, and a broker network that can be expanded. It assumes level funded continues growing at current rates, large carriers do not enter small group self-funded aggressively in the next five years, and the TPA’s existing differentiation (service quality, transparency, data quality) is sustainable against both carrier and insurtech competition. It breaks down when an insurtech with 26-fold revenue growth (Angle Health’s trajectory) builds a comparable product with better technology, or when the broker channel consolidates enough that relationship-based distribution ceases to be an advantage. The operational test: can you name the three capabilities your best brokers say you do better than anyone else? Are those capabilities defensible against a competitor with $200 million in investor capital and a purpose-built technology stack?

Choice B: Extend Into ICHRA. Add ICHRA administration alongside level funded. Help employers choose between models. This choice requires new compliance infrastructure, training for broker partners on model selection (FWD.02 Section 3: how brokers actually choose and why the TPA must provide guidance), and a clear theory of which employer profiles belong in which model. It assumes employers want a single administrator for both models, ICHRA margin is acceptable as a complement to level funded margin (it is lower), and the broker channel can present both options without defaulting to whichever is easier. It breaks down when ICHRA administration commoditizes and the margin disappears, which Take Command, Remodel Health, and comparable platforms are already driving with over $100 million in combined investor funding in 2025 alone (Healthcare Dive), or when running both models simultaneously creates service quality problems in the level funded core. The operational test: is your ICHRA strategy driven by employer need or by broker path-of-least-resistance? What is the ICHRA-to-level-funded revenue ratio in your pipeline, and is it moving in the direction you chose or the direction the brokers defaulted to?

Choice C: Double Down on Micro-Employers. Not “enter the micro-employer market.” TPAs already serving 1 to 10 lives are in it. The choice is whether to invest in the infrastructure to serve it profitably at scale, or to continue absorbing it as a relationship cost. This choice requires the technology investment that changes the unit economics (FWD.07 Tier 1: automated quoting, AI-assisted eligibility, automated stop loss reporting), access to a viable pooling mechanism with pool-level reinsurance (FWD.03 Section 2), and regulatory navigation for association or captive structures. It assumes the micro-employer market is large enough and growing fast enough to justify the investment (FWD.03 sizes this: 4.9 million employer firms with fewer than 10 employees, new business applications at record pace, the 55 to 64 cohort forming businesses at historically high rates), and that the administrative cost floor with technology is low enough that the segment is profitable (FWD.03 Section 3 models the delta from $3,000 to $6,000 per group to $800 to $1,500). It breaks down when the adverse selection problem makes the product unprofitable even with pooling and technology, or when Rippling or Gusto enters the segment with lower administrative costs because their infrastructure was built for this scale from day one. The operational test: what is your current administrative cost per group for your 1 to 10 life book? What would that cost be with FWD.07 Tier 1 capabilities deployed? Model the delta against your own numbers. Do not guess.

Choice D: Become the Fractional Workforce Benefits Coordinator. Build the product for workers with multiple income sources and no single employer offering group benefits. In the near term, this requires an association-based product for incorporated fractional professionals using existing legal mechanisms (FWD.04 Section 4), where the 120,000 fractional leaders identified in industry data represent the target population. In the longer term, it requires regulatory change to enable multi-employer contribution at scale. This choice assumes the near-term product can achieve enough pool scale (150 to 300 groups) to make reinsurance viable, and that the regulatory environment will eventually move toward portable benefits or multi-employer contribution with first-mover advantage accruing to operating knowledge rather than product exclusivity. It breaks down when the regulatory change does not materialize on a commercially viable timeline, when association formation proves harder than expected because the bona fide requirement limits participation after the DOL’s 2024 rescission of the AHP expansion rule, or when a well-capitalized competitor builds the same product faster once the regulatory path clears. The operational test: do you have a relationship with a professional association that could form the nucleus of the pool? Can you name the association, its membership count, and the share of members currently without group coverage? If you cannot, this choice is aspirational, not operational.

Choice E: Become a Platform Rather Than an Administrator. Build the operating system for the market. License technology and methodology to other TPAs. This choice requires significant technology investment across both FWD.06 (architecture) and FWD.07 (AI capabilities at Tier 2 depth across most processes), a genuinely differentiated platform that other TPAs would pay to use, a business model transition that disrupts existing revenue, and probably external capital, which means PE involvement with the timeline pressure and return expectations that implies. It assumes platform economics in benefits administration are better than operator economics (higher margin at scale, recurring revenue, network effects), that the TPA has something other operators would actually pay for, and that the existing customer base survives the transition. It breaks down when the platform cannot achieve scale before capital runs out, or when the existing TPA operation deteriorates during the transition because management attention is divided. Platform transitions in services businesses have a high failure rate. This is the highest-upside, highest-risk choice. The operational test: have you asked three other TPA CEOs whether they would pay for your technology? Not whether they admire it. Whether they would write a check. If the answer is not an enthusiastic yes from at least two, the platform thesis is untested.

The Questions Worth Sitting With

#

These are not rhetorical. The leadership team’s honest answers determine which choice is realistic and which is aspirational.

What is the actual source of your competitive advantage today? Name it precisely. Then test it: ask your three best brokers. If their answer does not match yours, one of you is wrong.

What does your stop loss carrier think your business looks like in five years? If you have not had that conversation, you are missing intelligence about your own risk exposure. In a market where loss ratios have deteriorated from 82 percent to 86 percent in five years and million-dollar claims are accelerating, the carrier’s view of your book matters.

What is the micro-group ratio in your book? Groups under 10 lives as a percentage of total groups and as a percentage of total premium revenue. How has that ratio changed over three years? Where is it heading? If you do not track this, you are managing a cost center without measuring it.

What would it cost, in revenue, relationships, and market position, to lose your five largest employer clients simultaneously? What would they need to find elsewhere, and could they find it? If the answer is “easily,” your retention is based on inertia, not value.

Which of the FWD.07 Tier 1 capabilities (quoting automation, eligibility parsing, claims anomaly detection, automated employer dashboards) could you deploy in the next six months with your current team? Which would require new hires or an outside partner? The gap between those two answers is the gap between your current technology capability and the market’s minimum expectation in three years.

If AI reduces the cost of claims adjudication by 60 percent over the next three years, what is the TPA for? The answer is not “nothing.” The answer is “whatever cannot be automated.” Name those things. That is your future value proposition.

What This Document Does Not Do

#

This document does not tell you which choice to make. It is not a strategic plan. It is a frame for a conversation that benefits from having been thought through before it happens.

The choices are not mutually exclusive in theory. In practice, most organizations cannot execute more than one significant strategic shift simultaneously. The honest question is not which choice is best in the abstract. It is which choice your organization can actually execute given your current capabilities, capital, and culture.

Capital structure deserves more honesty than strategy documents typically give it. Choices C, D, and E require capital that a privately held TPA may not have. The PE path is available: SIIA’s Corporate Growth Forum attracts more than a dozen private equity firms with specific expertise in the self-insurance marketplace. But PE capital comes with return expectations and timeline pressure that changes the strategic calculus. A strategy that requires external capital is not just a business plan. It is a partnership negotiation with investors whose interests are aligned with yours in the short term and may not be in the long term.

The market will reward deliberate choice and punish drift. The difference between the two is often less visible from inside the organization than from outside it. FWD.06 describes the architecture. FWD.07 describes what AI can do today. FWD.08 assesses who is positioned to build the benefits infrastructure the emerging workforce needs. The strategic choices framed here are the decisions that determine whether the TPA is one of the actors that builds it.