Insurity

Policy, billing, claims and marine systems keep running with the models removed, since carriers bought them to run the business. Valen's custom predictive models on pooled consortium data predate the current wave, while the newest Borealis features are assistive, and the May 2026 positioning aims AI at configuration cost.

The boundary for straight through processing, what triggers referral and whether a declined submission is flagged for review are not published. The control Insurity does state is a division of labor: AI handles repeatable work while underwriters and service teams stay in control of judgment, exceptions and approvals, and simple risks are processed or declined automatically only under rules the carrier configures.

Calibrated consortium models go live in 12 to 18 weeks and production ready ones in as little as 4, a defined fitting cycle. Out of sample testing, drift monitoring and revalidation are not described. Ratio Policy AI flags missing dependencies and conflicts across rating, rules, forms and workflows before a release.

UBIC cut quote times from two weeks to 45 seconds with Insurity Predict, a named result on the analytics product. Across its analytics customers Insurity reports loss ratios 3 to 10 percent better than the industry average, 5 percent lower claims costs and growth three times the industry rate over five years, figures that come without a comparison set or method. More than 500 insurers run Insurity software, including 22 of the top 25 American property and casualty carriers and 7 of the top 10 managing general agents, across more than 400 cloud deployments.

Carriers that use the Valen Data Consortium contribute their premium, loss and policy data to it, and Insurity states that contributions are anonymized, cleansed and protected. Opt out terms, retention and the fate of contributed data when a carrier leaves are not published. The pool, $109 billion of premium across standard property and casualty lines, feeds models sold back into the same market by design.

Privacy policy content, data processing terms and retention schedules are absent, and nothing describes how policyholder data moves between a carrier's core instance and the analytics estate. Consortium contribution makes that flow matter more than it would for a pure core vendor.

A trust center, attestation report, security certification or security statement is not offered, so controls cannot be assessed before a nondisclosure agreement. Delivery is cloud based, and the analytics products are named as running on Amazon Web Services.

Built in regulatory intelligence updates rates, rules and forms automatically, placing the product inside state filing machinery where an error has regulatory consequence. How changes are maintained and propagated, and who answers for an out of date rate in production, are unstated, and no license is held.

A responsible AI statement, fairness testing or bias audit is not published. Pricing and risk selection models are the exposure, since pooled data across many carriers makes proxying for a prohibited rating factor easier to do inadvertently, and Ratio Policy AI now drafts rates and underwriting rules from filings.

An applicant declined automatically or priced up by a consortium model has no notice and no route to contest the inputs, and nothing allocates liability between vendor and carrier. A model fitted on pooled industry experience can err the same way across many carriers at once.

Predictive models are the vendor's own, fitted on a pool it owns, so that chain has no third party model dependency. The language model behind the premium audit assistant and conversational geospatial querying is unnamed, as is whether insured data reaches an outside service.

More than 400 cloud deployments cover policy administration, billing and claims, with analytics scores wired into decisioning where a risk is priced. Interoperability with a rival core is never addressed, so the evidence concerns integration inside the vendor's own estate.

Cloud only delivery is framed as a strength, and the analytics products are named as running on Amazon Web Services. Regions, residency options, consortium data location and any private path are unpublished.

Cost is the center of the public argument, from a specialty suite priced at a fraction of other cores to a May 2026 call for carriers to stop signing contract extensions, yet no price is published. Self service configuration is pitched as removing system integrator dependency, which goes unquantified like everything else.

Large and specialty carriers, brokers and managing general agents are all served, with a purpose built suite for the smaller specialty end. Lines run from homeowners and personal auto to workers compensation and commercial property, concentrated in American property and casualty.