# Finic

> Finic builds AI for both kinds of fraud and risk decisions: Grayson, a frontier decision model fast enough for the real-time path, and agents that do the careful work of investigation.

Grayson is Finic's decision model API: send it a case and typed questions, and get a calibrated probability for every answer. Its developer docs have their own llms.txt.

## Grayson

- [Grayson docs](https://docs.finic.ai/llms.txt): the API, ready-to-run recipes for fraud, AML and risk decisions, and a CLI to evaluate Grayson on your own data
- [Finic portal](https://portal.finic.ai): API keys, the playground, usage and billing

## Blog

- [Question-Bottleneck Scoring](https://finic.ai/blog/question-bottleneck-scoring): A whitepaper on predicting customer outcomes by asking a decision model plain-language questions and weighting its calibrated answers: human-quality judgment calls at the speed and cost of code.
- [Can a decision model make loan default prediction more accurate?](https://finic.ai/blog/decision-models-xgboost-loan-default): We added a decision model's judgment to a traditional credit model. With Grayson, it predicted defaults better on later loans, worth up to $155,000 per 10,000 applications in real loan outcomes. Jev's answers added less.
- [Can a decision model make insurance fraud detection more accurate?](https://finic.ai/blog/decision-models-xgboost-insurance-fraud): We added a decision model's judgment to a traditional fraud model. With Grayson, it caught more auto insurance fraud in 10 of 12 test runs, worth about $410,000 a year for an insurer with 100,000 claims.
- [Introducing Finic Grayson](https://finic.ai/blog/introducing-grayson): Grayson is a universal decision model built for real-time decisions across fraud, risk and financial crimes. Send it any context and the questions you need answered, and get a calibrated probability for every answer in under 150 ms of model time for typical requests.
- [Black Hat 2026 Buzzword Index](https://finic.ai/blog/black-hat-2026-buzzword-index): BH 2026 felt more buzzwordy than usual. We measured every sponsor's Black Hat profile and landing page to see which buzzwords dominate, which are rising, and which are dying.
- [Open-source fraud investigation skills for AI agents](https://finic.ai/blog/fraud-investigation-skills-for-ai-agents): First Agent Skills purpose-built for fraud analysts. Drop them into Claude Code, Cowork, or Codex so your agent investigates this week's patterns.
- [The most expensive fraud event is the one you haven't named yet](https://finic.ai/blog/most-expensive-fraud-event): Rules and models are useful once a pattern has a name. Organized fraud rings make money before that.

## Policies

- [Acceptable use policy](https://finic.ai/policies/acceptable-use): What you may and may not do with Grayson and the Finic portal.
- [Data processing addendum](https://finic.ai/policies/data-processing): How Finic processes personal data on your behalf, including how long we keep requests and Zero Data Retention.
- [Master services agreement](https://finic.ai/policies/msa): The terms that govern use of Finic's services.
- [Privacy policy](https://finic.ai/policies/privacy): How Finic collects, uses and shares personal information.