Criterica Intelligence — 23,706 production models trained on 475M+ real court records
Platform

Five layers. One architecture.

The Criterica Intelligence platform is a five-level stack — Atlas backbone, infrastructure products, diagnostic audits, operating modules, and vertical applications. Every layer is Atlas-native. Every output is traceable to the model and data that produced it.

Platform Architecture
Level 05
Vertical Applications

12 market product families — one per client segment, each built on the same Atlas backbone with market-specific model configurations, audit types, and workflow integrations.

Litigation Finance
Law Firms
Insurance & Claims
Enterprise Legal
Legal Technology
Financial Institutions
Healthcare
Courts & Justice
Government
Real Estate
Employment
Consumer Legal
Level 04
Operating Modules

The intelligence applications institutional buyers operate day-to-day: underwriting, monitoring, resolution, capital allocation, claims triage, and litigation risk management.

Underwrite
Monitor
Resolve
Counsel
Capital IQ
Claims IQ
Litigation IQ
Court IQ
Risk IQ
Partner API
Level 03
Diagnostics

The first commercial engagement is always a diagnostic audit using the client's own data. Ten audit types spanning every major vertical — the wedge that converts institutional trust into production deployment.

Portfolio Intelligence
Claims Intelligence
Litigation Intelligence
Capital Structure
Data Readiness
Counsel Performance
Court System
Medical-Legal Risk
Employment Risk
Construction Dispute
Level 02
Infrastructure

The core product suite that sits between Atlas and the operating modules. Seven infrastructure products delivering storage, model access, signal delivery, financial data, graph queries, API access, and observability.

Vault
Models
Signal
Ledger
Graph
API
Control Room
Level 01
Atlas

The regulated outcomes intelligence backbone. The data, model, decision, workflow, and audit trail infrastructure that all other layers depend on.

Ontology
Data Layer
Model Layer
Decision Layer
Workflow Layer
Audit Trail
Atlas API
What you get back.

A decision, not a data dump.

Every case scored against jurisdiction-specific models. Win probability, settlement curve, duration, damages range, and a single recommendation — structured for the decision you actually have to make.

Criterica · Case Score Output2026-05-26 14:22 UTC
CASE IDCC-2847
CASE TYPESDNY · Medical Malpractice · NOS 362
FILED2024-03-11
PLAINTIFF WIN PROB
71%
SETTLEMENT PROB
84%
EXPECTED DURATION847 days
DAMAGES — EXPECTED$1,240,000
DAMAGES — RANGE$680K – $2.1M
RISK SCORE28 · LOW
PRIMARY MODELfed_nysd_medmal_outcome_v3
Recommendation
Case clears fund threshold. Plaintiff win probability and settlement curve consistent with portfolio targets. Proceed to term sheet.
FUND
Start with a diagnostic audit.
Every engagement begins at Level 03. Your data, your portfolio.