Insurance · Deep dive

Reserv

An AI-native third-party administrator for property-and-casualty insurance — it takes over carriers' and MGAs' claims files, runs them on its own Glance platform, and uses generative AI to let a smaller bench of human adjusters close claims faster and cheaper than legacy TPAs like Sedgwick and Crawford.

emerging

The question that decides it: Reserv's bet is that AI-native adjusting cuts loss-adjustment expense and cycle time enough that Global 2000 carriers and MGAs migrate claims off Sedgwick, Crawford and Gallagher Bassett — and keep paying a per-claim TPA fee for the privilege. But a TPA is fundamentally a labor-arbitrage services business, and the same large-language-model tooling that lets Reserv run claims with a thinner adjuster bench is available to the incumbents it is attacking (Sedgwick is even owned by KKR, Reserv's own lead investor). Does AI let Reserv structurally win share and expand margin on the strength of a purpose-built platform and clean data — or does it merely reset the cost curve for the whole industry, commoditizing adjusting and leaving Reserv a fast-growing but low-margin services roll-up rather than a software-margin moat?

My take ★★ Very interesting ★ Interesting − Maybe ✗ Not interesting

Snapshot

Reserv is an AI-native third-party administrator (TPA) for property-and-casualty insurance: carriers, managing general agents (MGAs), corporate captives and brokers hand it their claims files, and Reserv adjudicates them end-to-end on its own software while using generative AI to make each human adjuster more productive. Founded in 2022 by Snapsheet veteran CJ Przybyl and operator Martha Dreiling, it says it reached roughly $100M of annual recurring revenue in about four years, serves nearly 200 clients, and employs 500-plus claims adjusters across the US and UK. In May 2026 it raised a $125M Series C led by KKR — bringing total funding to nearly $200M — with the stated ambition of scaling from ~500,000 complex claims a year today toward 30 million within four years. The tension worth holding onto from the start: this is a scaled, fast-growing services business dressed as a software company, and its own lead investor already owns the incumbent it is trying to beat.

Founding story

Reserv is Przybyl’s second act in claims. At Snapsheet he spent more than a decade building what the company billed as the first virtual claims process for auto losses — a business that grew past 400 people, worked with seven of the top 10 US P&C carriers, and by Reserv’s account processed 2.5 million-plus claims and over $10B in indemnity. That vantage point produced a specific frustration: claims technology had gotten good at capturing photos and payments, but the actual adjudication — the reading, the coverage analysis, the negotiation, the correspondence — was still gated by human adjusters drowning in administrative work and legacy systems that took years to change.

Reserv’s founding thesis, in Przybyl’s framing, was to prove how seamless claims could be “if technology weren’t the bottleneck” — and, crucially, to automate the whole organization rather than ship a claims tool and hope carriers adopt it. So instead of selling software to insurers, Reserv became the TPA: it owns the claims operation, the adjusters and the platform together, which lets it deploy AI internally without waiting for a carrier’s IT roadmap. Martha Dreiling joined as co-founder and COO in August 2022, bringing operating experience from OnDeck, Attune and Rhino and a BCG analytics background, and built the adjuster-plus-AI operating model. The company came out of stealth in July 2022 with an $8M seed round.

How it works

Mechanically, a claim enters Reserv at first notice of loss (FNOL) and is ingested into Reserv Glance, the company’s claims platform. Glance lets a client consolidate historical and live claims into a single database — Reserv says customers can phase out legacy claims systems “within weeks” — and then applies what it describes as fully explainable AI to read, prioritize and act on claims. The automation is dialable per client: straightforward, low-complexity claims can be handled in a largely automated flow, while complex claims run in an “assisted” mode where AI does the drudge work (summarizing files, drafting correspondence, surfacing next-best actions, flagging missing documentation) and a human adjuster makes the judgment calls and owns the relationship with the carrier.

That is the core mechanic: not a robot adjuster, but an adjuster-led model with an AI copilot doing operational execution behind the scenes, so one adjuster can carry more files and spend time on the hard, high-severity ones. Reserv reports handling nearly all non-workers-comp P&C lines — commercial auto, business-owners policies, general liability, professional lines, medical malpractice, property — and is licensed and live across the US and UK. The company frames itself as operating “post-AI,” meaning new model capabilities get folded straight into the platform rather than shipped as bolt-on features, and it pitches the resulting proprietary claims data as a compounding “data advantage” for its MGA and carrier clients.

Product and business overview

Reserv sells two things that are hard to separate: a claims service (the TPA — it employs the adjusters and closes the files) and a claims platform (Glance plus the AI tooling). Around that core sit adjacent modules the company has been building — claims automation models, faster integrations, and “net-new non-claim” modules it flagged at the Series B. The strategic logic is that owning both the software and the operation is the moat: because Reserv runs the claims itself, it can push automation at its own pace and harvest structured data from every file, rather than being one vendor in a carrier’s stack. The risk baked into the same fact is that the revenue is still, at bottom, per-claim services revenue.

Business model and pricing

Reserv is paid the way TPAs are generally paid: a fee to administer claims, typically structured as a per-claim or per-file administration fee under a program agreement with a carrier, MGA, captive or self-insured, with adjuster labor bundled into that fee. Reserv does not publish a rate card, and precise per-claim economics are not disclosed. It describes ~$100M of ARR (May 2026) built on this file-flow revenue, triple-digit year-over-year revenue growth for two consecutive years, and a business that is near cash-flow positive. The economic pitch to carriers is loss-adjustment-expense (LAE) reduction and speed: Reserv has cited sustained cycle-time reductions on property and auto claims on the order of 1.6x to 2.6x. The strategic question the pricing raises is the one every services business faces — a per-claim fee scales with headcount and volume, not with software gross margins, unless AI genuinely decouples the two.

Traction over time

Read the growth with the base in mind: Przybyl himself has framed the arc as ~$20,000 to ~$100M of revenue in four years, which is genuinely fast, but the “50x ARR growth” cited around the 2023 Series A is off a near-zero starting point. The client count moving from ~100 (mostly MGAs) in mid-2025 to ~200 by mid-2026, alongside 500-plus adjusters, is the more telling signal that this is scaling as an operations business, not just a software seat count.

Market analysis

The insurance TPA market is large but slippery to size, because headline figures often fold in health-benefit administration and the premium/claim dollars flowing through TPAs, not the administration fees themselves. The Business Research Company put the insurance TPA market at roughly $372B in 2025 rising to ~$404B in 2026 (~8.7% CAGR); other houses cite $432B-$513B (2024) with forecasts toward $795-886B by the early 2030s. Reserv’s own serviceable target is narrower and more credible as a wedge: it frames the opportunity as the “non-field-based commercial P&C claims market,” and ties its 30-million-claim ambition to automating a substantial share of that. Two structural forces favor it: carriers face a persistent, aging adjuster shortage and rising LAE, and generative AI has, for the first time, made automating unstructured claims work plausible. The same second force is the threat — it lowers the barrier for everyone, incumbents and insourcers included.

Competitive intel

Reserv sits between two kinds of rival. On one side are the legacy TPAs — Sedgwick, Crawford and Gallagher Bassett — enormous, entrenched, and the direct targets of the “we’re faster and cheaper because we’re AI-native” pitch. Sedgwick is the giant (33,000+ people, 80+ countries) and the awkward one: it is controlled by KKR, so Reserv’s own lead investor already owns the market leader it aims to displace — a fact Przybyl spins as strategic access to insurance balance sheets and know-how, but which also means KKR is hedged on both sides of the disruption. Crawford (~$1.35B 2025 revenue, ~$500M market cap) is the cautionary comparable: a mature claims-services business the public market values at well under 1x revenue, which is precisely the multiple trap Reserv must escape. On the other side are the AI/software players — CCC Intelligent Solutions (which bought EvolutionIQ in December 2024) and Five Sigma — that attack the technology layer without running claims, and could arm carriers to insource. Reserv’s differentiated claim is that owning the service and the software together lets it move faster than the incumbents and capture data no pure-software vendor can. Whether that integration is a moat or just more headcount to feed is the crux.

History and evolution

• 2022 — Founded; begins operating around May; exits stealth in July with an $8M seed co-led by Altai Ventures and Bain Capital Ventures. Martha Dreiling joins as co-founder/COO in August.
• Oct 2023 — $20M Series A (Altai and BCV again co-lead) to expand generative-AI claims-analysis tools; ~70+ staff across US and UK.
• Jun 2025 — $25M Series B led by Flourish Ventures, with Accenture Ventures joining; ~350+ staff, ~80 MGA clients and ~20 carriers.
• Sep 2025 — Series B extended by $16M to $41M, led by QBE Ventures.
• May 2026 — $125M Series C led by KKR (via Next Generation Technology Growth), pushing total funding to nearly $200M; ~$100M ARR, 500+ adjusters, ~200 clients; sets a target of 30M claims/year within four years.

What people say

The case for. The commercial signal is strong: growing from near-zero to ~$100M ARR in four years, near cash-flow-positive, with a marquee investor that approached unsolicited, is not the profile of a hype cycle. KKR’s Patrick Devine publicly framed what Reserv has built as “truly differentiated,” Clients cite meaningful cycle-time reductions (roughly 1.6x-2.6x on property and auto), and the roster — nearly 200 insurers, MGAs, captives and brokers, reportedly including Global 2000 carriers and logistics firms — suggests the service actually clears enterprise procurement.

The complaints. Two categories. First, employees: Reserv’s Glassdoor sits around 3.5/5 with roughly 55% recommending it, and recurring themes point to a hard-charging metrics culture, thin training for new adjusters, high turnover and layoffs, and modest pay (a claims-adjuster median around $53K). That is the operational reality of running a fast-scaling adjuster bench, and it matters because service quality in a TPA is the product. Second, the structural skeptic case: a TPA is a low-margin, labor-driven services business — the public-market comps (Crawford at sub-1x revenue) show how such businesses get valued — and the very AI that powers Reserv is available to Sedgwick, Crawford and to carriers who might simply insource high-volume claims. Analysts explicitly warn that commoditized claims are under “AI and insourcing pressure.” If AI resets the cost curve for everyone, Reserv grows into a competitive, margin-pressured services market rather than a defensible software one.

Outlook: the open question

What would have to be true for Reserv to be more than a fast-growing services roll-up: AI has to structurally decouple its revenue from its headcount — driving loss-adjustment expense and cycle time down far enough, on a platform and data asset competitors can’t easily replicate, that Global 2000 carriers and MGAs keep migrating files off Sedgwick, Crawford and Gallagher Bassett and Reserv keeps a widening slice of the fee as software-like margin rather than passing all the savings through. The bull case is concrete and partly evidenced: real cycle-time gains, ~$100M ARR near breakeven, ~200 clients, and an integrated own-the-service-and-the-software model that can push automation at its own pace and compound proprietary claims data. If the 30-million-claim ambition is met with a thinner adjuster ratio per claim each year, the margin story becomes believable and the platform becomes a genuine moat.

What would sink the thesis: the same generative-AI tooling proves to be a rising tide, not a Reserv-specific edge. Sedgwick — owned by Reserv’s own lead investor — deploys comparable AI at vastly greater scale; carriers license Five Sigma or build on CCC’s data and insource the high-volume claims that are easiest to automate; and Reserv is left running the hard, high-severity, human-heavy files at TPA economics, where the public market pays Crawford less than 1x revenue. The tells to watch are unglamorous and specific: claims handled per adjuster over time, gross margin trajectory as volume scales, client and premium retention, and whether Reserv can win workers’-comp and field-heavy lines it currently sits out. Reserv has to prove that being AI-native is a durable advantage and not merely the new table stakes that its own investor’s other portfolio company adopts just as fast.

How a challenger would attack it

The wedge is Reserv’s own bench. Reserv’s product is ultimately its adjusters, and the Glassdoor signal — 3.5/5, ~55% recommend, thin training, high turnover, a $53K median adjuster wage inside a hard-charging metrics culture — says the bench is stretched. A challenger would go agent-first rather than adjuster-first: instead of using AI to make 500 modestly paid humans carry more files, it would build for the 2027-era model generation where low-complexity claims close with no human in the loop, hire a small bench of senior, well-paid adjusters for the judgment calls, and market the service-quality gap Reserv’s turnover creates. Second vector: pricing. Reserv still charges a conventional per-claim TPA fee, so a challenger can sell outcome-based pricing — a share of verified LAE savings, or software-style per-claim pricing at a fraction of TPA rates — which Reserv can’t match without repricing the $100M ARR base its near-breakeven economics depend on. Third: attack the lines Reserv sits out. It handles no workers’ comp and avoids field-heavy claims; owning workers’ comp first means meeting Gallagher Bassett, not Reserv, and arriving at Reserv’s core lines later with a fuller book. The window matters — Reserv’s own thesis concedes the AI is a rising tide, not a proprietary edge.

Same playbook, new buyer

Take the AI-native TPA model where Reserv isn’t. The most direct shift is workers’ compensation — the largest TPA line, explicitly outside Reserv’s non-workers-comp footprint, and the heart of Gallagher Bassett’s and Sedgwick’s book. Reserv won’t follow quickly: comp is a state-by-state regulatory and medical-management build its platform wasn’t shaped for, and its 30-million-claim target keeps it focused on high-frequency commercial P&C. Second shift: self-insured corporates and captives as the primary buyer rather than MGAs and carriers. Reserv’s client base skews MGA (~80 of ~100 clients at the Series B); a Fortune 1000 self-insured buys on different criteria — cost transparency and data ownership rather than program flexibility — and resents the opacity of per-claim fees. Third: geography. Reserv is US/UK only; Continental Europe, Canada and Australia have the same aging-adjuster problem, legacy TPA incumbents, and no AI-native option. A regional clone with local licensing and language models fitted to local claims documents gets years of head start before a US-scaled Reserv, still absorbing $200M of venture expectations at home, can justify the expansion.

Sources and further reading

• Reserv Announces $125 Million Series C Financing Led by KKR (BusinessWire, May 2026)
• Reserv raises $125m Series C led by KKR (FinTech Global, May 2026)
• Reserv’s Przybyl on the unsolicited $125M KKR raise and growing to $100M in four years (The Insurer, May 2026)
• Reserv Raises $25M Series B to Support Rapid Scaling (PR Newswire, Jun 2025)
• Claims tech firm Reserv extends Series B to $41m (FinTech Global, Sep 2025)
• Reserv Raises $20M Series A Co-Led By Altai Ventures and Bain Capital Ventures (Coverager, Oct 2023)
• Reserv Comes Out of Stealth Mode and Announces $8M Seed Round (Coverager, Jul 2022)
• Why We Invested in Reserv (Flourish Ventures, 2025)
• Working at Reserv — employee reviews (Glassdoor, accessed Jul 2026)
• Insurance Third Party Administrators Global Market Report 2026 (The Business Research Company via GlobeNewswire, Apr 2026)
• Crawford & Company revenue and market cap (StockAnalysis, accessed Jul 2026)
• CCC Intelligent Solutions Announces the Acquisition of EvolutionIQ (CCC Intelligent Solutions, Dec 2024)