Criterica Intelligence — production models trained on real court records, not synthetic data
Company
█

Structuredpredictionasinfrastructure.

Criterica Intelligence is the data science and model infrastructure division of Criterica Group. We build the prediction and decision systems that underwrite complex litigation, price risk, and power capital deployment. Production predictive models — one per jurisdiction, case type, and prediction target, each individually cleared through temporal-holdout gates — trained exclusively on real court records. Defensible probability is the product.

We answer four questions from the record.Criterica Intelligence is the risk intelligence brand of Criterica. How likely is each outcome. How long will it take. What will it settle or be awarded for. What is it worth today. Every client problem we take on reduces to one of those, and every answer we give traces to real cases and a published method. We validate on cases that resolved after the ones a model learned from, and we review calibration before a number is used, because a client prices capital and reserves on it.

Criterica Intelligence is a Criterica brand, founded by Jack Cosentino and built by research, data, engineering and client teams in Palm Beach, Florida.

Mission

Precision is not a feature. It is the foundation. Every model we build, every dataset we ingest, every prediction we deliver is held to the standard of institutional-grade evidence. We do not approximate. We tune until a 70 percent prediction is right about 70 percent of the time.

Approach

We acquire, structure, and model outcomes data from regulated industries. Our infrastructure is built on real judicial records, financial filings, and regulatory enforcement data. Not synthetic. Not approximated. The moat is the data. The edge is how reliably it predicts.

The Tangent Point
The Mark

The point where the rate of change is steepest.

CURVE

A sigmoid — the natural shape of outcome probability as evidence accumulates. It starts uncertain, accelerates through a critical zone, then resolves.

LINE

A tangent at the inflection. It represents the derivative: not the probability itself, but the rate at which certainty is emerging. That is the actionable signal.

RING

The exact point of contact between curve and tangent. Where rate of change is maximum. Where the system has found the answer. The ring says: we found the exact point.

Nothing else in this space locates that point. That is the product.

Global Presence
United States
Active
Real court records across federal circuits and state courts
Federal circuits and state courts nationwide. Licensed financial intelligence and regulatory enforcement datasets integrated.
United Kingdom
Active
England & Wales, Northern Ireland · production models deployed
England & Wales, Scotland, Northern Ireland. ABS-eligible — non-lawyer ownership unlocks law firm equity structures.
Australia
Active
NSW, Federal Court · active models deployed
All states. ABS jurisdiction with mature non-lawyer ownership frameworks. PI and commercial litigation coverage.
Canada
Active
BC, AB, ON, QC · production roles deployed
British Columbia, Alberta, Ontario, Quebec. Common law provinces with predictive feature parity to US courts.
South America
Coverage Established
Data ingested — model training underway.
Brazil, Chile, Colombia, Argentina. Commercial litigation and international arbitration expansion.
Asia Pacific
Coverage Established
Data ingested — model training underway.
Singapore (SIAC), Hong Kong (HKIAC), Japan. Major arbitration seats and growing institutional litigation finance.
Middle East
Coverage Established
Data ingested — model training underway.
UAE (DIFC/ADGM), Saudi Arabia, Qatar. ICC seat expansion and sovereign capital moving into legal finance.
India
Coverage Established
Data ingested — model training underway.
Supreme Court 2018 third-party funding clearance. $1.3T+ legal market. Expansion in formation.
European Union
Coverage
CJEU and regulatory enforcement data integrated. ICC and ICSID international arbitration coverage.

Statistics shown reflect historical or illustrative model outputs derived from real case data. They are not predictions or guarantees of any individual outcome. Litigation results depend on facts, jurisdiction, judge, and counsel, and vary case by case. Model accuracy is subject to selection effects and changing legal dynamics.