Actuarial AI Practice

AI forInsuranceand Actuaries

Reserving, pricing and claims models built by actuaries who write the code. Statistically defensible, regulator-ready, and fitted to the portfolio you actually underwrite.

AI Consultation

No two portfolios are alike, so we don't sell a product. We work with your team to find where AI genuinely earns its place in your reserving, pricing and claims workflows.

Looking for inspiration? Check out our Autonomous AI Hedge Fund to see a real-world example of what our models can achieve.

Our Services

Built for Insurance & Actuarial Professionals

Pick a service to see what it does. Purpose-built models, not off-the-shelf software repurposed for insurance.

CLAIMS TRIANGLEDevelopment year →202020212022202320242025ObservedProjectedRESERVE DISTRIBUTIONP50P75P99.5Best estimate, risk marginand capital from one model

Stochastic & deterministic

Illustrative data

Advanced Reserving

Beyond the chain ladder: a full range of outcomes, not a single number.

  • Full uncertainty

    P50 to P99.5 reserves from one consistent model.

  • Methods side by side

    Mack, bootstrap ODP and Bayesian against your chain ladder.

  • Regulator-ready

    Documented assumptions, diagnostics and reproducible code.

Chain ladderMackBootstrap ODPBayesianGLM

Engagement

How an Engagement Runs

Four stages, each ending in something you can put in front of a committee.

  1. 1

    Discovery

    We review your current methods with your team and agree the question the model must answer.

    You get: Scoping note
  2. 2

    Data & Modelling

    Data validated end to end, then candidate models fitted and benchmarked against your incumbent.

    You get: Model benchmark
  3. 3

    Validation

    Back-testing, sensitivity analysis and documented assumptions, ready for review.

    You get: Validation report
  4. 4

    Deployment

    Shipped into your stack with monitoring and a handover, so your team owns it.

    You get: Live model + handover

Have a reserving or pricing problem worth solving?

Tell us about the portfolio and the question you need answered. We'll come back with an approach, not a sales deck.