Reserv

Automation does real work, handling simpler claims outright and supporting adjusters on complex ones, and the growth thesis depends on it entirely since scaling from 500,000 to 30 million claims cannot be done by hiring. What sits beneath is a substantial services and platform business: more than five hundred licensed adjusters, a migration tool that consolidates legacy claim files, and the administration function itself. Apply the removal test and a modern, well run third party administrator remains, which is exactly how the company differentiates itself from software only competitors.

Two mechanisms combine and both are unusual. The client chooses the automation level, from fully automated handling of simpler claims through to supported approaches where an adjuster leads on complex ones, so the boundary belongs to the carrier rather than the vendor.

And the humans in that loop are licensed adjusters carrying their own professional obligations, more than five hundred of them, which the company describes as an adjuster led experience supported by artificial intelligence rather than the reverse. Structural accountability of that kind is stronger than a configurable threshold. What is not published is where the automated tier stops by claim value or type.

The client selected automation level is a genuine control and functions as a staged adoption path, letting a carrier keep adjusters on anything it is not ready to automate. Beyond that the model layer is undocumented. No accuracy or error rates for automated claim handling, no evaluation methodology, no model documentation, no description of how automated determinations are sampled and reviewed, and the stated practice of putting the newest tools straight into production cuts against the validation discipline a carrier's own risk function would expect.

The evidence is quantified where it counts and includes a figure most private vendors withhold. Annual recurring revenue is stated at 100 million dollars, the client base at roughly 200 insurers, captives, managing general agents and brokers, and current volume at 500,000 complex claims a year with capacity more than doubling annually. A 125 million dollar round led by a major private equity firm arrived in 2026, with that investor's insurance sector experience behind the diligence. Operations span North America, the United Kingdom and Europe. What is missing is attribution: no individual carrier is named as a client and no per client outcome is published.

Two things deserve scrutiny. The company describes operating in a post artificial intelligence environment where the newest tools are immediately production ready and integrated, which is a velocity claim in a setting where a model decides whether an injured claimant is paid, and nothing describes what testing sits between a new tool and a live claim.

Second and larger: handling claims for roughly 200 carriers concentrates their loss experience in one platform, and loss data is the raw material of pricing, so what separates one carrier's claims history from another's is a first order competitive question that no public material addresses.

Claims files are among the most sensitive records in insurance, containing injury descriptions, medical evidence, employment details and personal circumstances, and in liability claims that material concerns a claimant who has no relationship with the carrier and never chose the administrator. Operating across the United Kingdom and Europe brings a separate and stricter regime to bear on the same data. No published privacy framework, retention schedule, special category data position or subprocessor list was located.

No trust centre, enumerated certification list, attestation scope or audit period was located in this pass. The platform holds claim files including medical evidence for roughly 200 carriers across three regions, and a major private equity investor conducted diligence before committing 125 million dollars, so independent assurance certainly exists and has been supplied privately. The grade records what an outside buyer can verify.

Licensure sits inside the operation rather than beside it. The claims administration arm is a licensed third party administrator, a regulated status in most states, and it employs more than five hundred licensed adjusters performing an activity that requires individual licensing in the majority of jurisdictions. That means the entity making claim determinations is directly answerable to insurance regulators rather than merely supplying tools to someone who is. Operations in the United Kingdom and Europe add further supervisory regimes, and the corporate separation between the administration and technology arms is disclosed openly.

Automated handling of simpler claims means a model is deciding whether a person is paid and how much, which is the most directly consequential automated decision in this index outside credit. Unfair claims settlement practices statutes exist in every state precisely to govern that determination, and bad faith exposure follows a wrongly denied or underpaid claim.

Nothing public reports accuracy for automated handling, analyses whether outcomes differ across claimant types or claim characteristics, or explains how a claimant disadvantaged by automation would be identified.

Accountability here is structural rather than promised, and it is the reason this grades above most of the index. As a licensed administrator Reserv assumes contractual and professional responsibility for adjusting claims on a carrier's behalf, its adjusters hold individual licences with duties attached, and errors in claim handling expose the administrator directly rather than only the carrier. A wrongly handled claim therefore has a named accountable party and a regulator to complain to.

What is absent is anything specific to the automation: no accuracy commitment, no remediation term, and no described route for a claimant to learn that their claim was handled automatically or to request human review.

The company states that the newest artificial intelligence tools are immediately production ready and integrated into its platform, which is an implicit acknowledgement that external model providers sit in the path, and it names none of them. No subprocessor list is published, no infrastructure providers are identified, and nothing describes where claim files containing medical evidence are processed. The composition of the pooled claims data across roughly 200 carriers is likewise undisclosed.

The migration capability addresses the specific obstacle that keeps carriers with incumbent administrators: moving historical and open claim files between systems conventionally takes up to nine months, and the company's tool is described as completing it in about two weeks, which changes the economics of switching.

Beyond migration the platform runs the whole claim lifecycle from first notice of loss to settlement rather than plugging into someone else's claims system, and it competes directly with the established claims platforms rather than integrating with them.

Delivery is cloud hosted with operations across North America, the United Kingdom and Europe, which means claim files containing medical and injury information cross regimes with materially different rules on special category data. Arrangements almost certainly exist to satisfy European clients. None are published: no hosting regions, residency options, transfer mechanisms, tenancy separation between carriers or subprocessor chain were located.

No rates, tiers or billing unit were located. Administration work in this market is conventionally priced per claim or on a fee schedule, and the dual structure raises a further question nothing answers: whether the software is licensed separately from the administration service, or whether a carrier must buy the handling to get the technology.

Five buyer types are served rather than one, spanning insurers, corporate captives, managing general agents, brokers and self insured businesses, which covers most of the ways property and casualty risk is carried and administered, and coverage extends across North America, the United Kingdom and Europe.

The boundary is functional and deliberate: this is claims, from first notice of loss to settlement, concentrated on non field based commercial work, with nothing addressing underwriting, distribution, policy administration or any line outside property and casualty.