Insurers

Judgment preservation, after-the-event coverage, and contingent risk transfer all price the same underlying uncertainty from the opposite side of the table from a funder: the probability that a judgment survives appeal, that a claim resolves in a particular range, or that a matter's timeline extends past a policy's assumptions. Underwriting that risk without a calibrated, case-level outcome layer means pricing largely on precedent and underwriter experience rather than a checkable distribution.

Criterica's production model fleet produces the same class of resolution and duration distributions used across the platform, applied here to the specific question a carrier is pricing: appellate reversal probability by jurisdiction and claim type, settlement-timing distributions relevant to a policy term, or claims-data patterns relevant to a book of contingent risk.

The relationship can run in either direction: Criterica supplies outcome intelligence that informs a carrier's own underwriting, or a carrier contributes claims data, appropriately governed, that sharpens the corpus for both sides.

Carriers are, in a real sense, on the other side of the same distributions a funder prices. A calibrated model does not take a side in that relationship; it describes the same uncertainty for whoever is pricing it.

This symmetry is deliberate. A model calibrated to describe reality rather than to favor one side of a transaction is the only kind of model that both a funder and a carrier can reasonably rely on.

The same intelligence also supports a carrier's own reserve-setting on contingent risk exposure, since a calibrated distribution for how a claim class typically resolves is directly relevant to how a carrier reserves against it, not only to how it prices new business.

Applies the model fleet to the specific risk a carrier is pricing, disclosed with confidence interval and support size rather than a single number presented as certainty.

Structures data-sharing arrangements where a carrier's claims history improves model coverage for the claim types and jurisdictions that matter to its book.

Supports ongoing monitoring rather than a one-time underwriting pass, since a judgment's or claim's status can change materially after a policy incepts.

Keeps the intelligence function separate from any capital desk, so a carrier evaluating Criterica's outputs is evaluating a probability estimate, not a position someone else has an incentive to see priced a certain way.

Confirms coverage segment by segment where a carrier's book spans several claim types or jurisdictions, so a carrier knows precisely which parts of its book the model fleet actually speaks to today.

Scopes engagements specifically around reserve-setting where that, rather than new-business pricing, is a carrier's interest, since the two questions draw on the same underlying distributions but are used differently inside a carrier's own process.

The specific product line, judgment preservation, ATE, or contingent risk, and the jurisdictions and claim types that make up the book.

A description of what claims data the carrier could make available, if a data-sharing arrangement is of interest.

Current underwriting methodology at a high level, so Criterica can identify where model output would add the most to an existing process rather than duplicate it.

Any regulatory or reporting constraints on how claims data can be shared externally, so a data-sharing arrangement, if pursued, is structured within them from the outset.

The first call scopes the product line and claim types, confirms model coverage for the relevant jurisdictions, and identifies whether the relationship is intelligence-only or includes a data-sharing component.

Where coverage is confirmed, the next step is typically a defined pilot: a bounded set of claims scored against the model fleet, compared to the carrier's existing pricing, before any broader integration is discussed.

A pilot typically runs over a defined claim set and time window, with results compared directly against the carrier's own historical pricing before either side commits to anything broader.

Whether the pilot leads to an ongoing intelligence relationship, a data-sharing arrangement, or both is a decision made jointly at the end of the pilot, based on what the results actually show.