Key facts
- Launch: Founder Ryan Steil launched Prediction Lab publicly in 2025.
- Recognition: Prediction Lab was one of the finalists in State Farm's 2026 startup pitch competition for insurance technology.
- Team: Prediction Lab is a small, remote team and was hiring actuarial, machine learning and platform engineers in 2026.
Focus areas
About
Prediction Lab sells a hosted predictive modelling platform to property and casualty insurers and their pricing and actuarial teams. Actuaries use it to build pricing, underwriting, claims, retention and lifetime value models, connecting directly to data in Google BigQuery, Snowflake, Amazon S3, PostgreSQL and Microsoft Azure. Teammates work on the same models in real time, with every change versioned for review and rate filing. Founder and chief executive Ryan Steil launched the product publicly in 2025.
How they use AI
Prediction Lab fits generalised linear models, the standard tool for insurance rating, and can build a starting feature set automatically with gradient boosting before an actuary refines it by hand. A context-aware assistant answers questions about the user's own models and data. Through a Model Context Protocol connection, AI agents can run modelling experiments on branches, which humans review and approve before they merge into production models.
Products and services
- Modeling engine — Fast GLM fitting and refitting on multi-million-record insurance datasets
- Model review — Branches, validation checks, comments and approvals with a full audit trail
- AI assistant and agents — Questions on models and data, plus MCP-connected agents that run experiments for human review
- Model deployment — Hosted API endpoints that serve approved models
Key people
- Ryan Steil — Founder and Chief Executive Officer