Actuarial AI Integration
Where actuarial rigour meets modern AI
I help insurance firms integrate AI into actuarial and pricing workflows — building tools grounded in regulatory reality and designed for people who understand the business.
Regulatory depth
Solvency II
Multi-model AI
Claude · GPT · Gemini
Domain coverage
Pricing · Reserving · ALM
Tools
Purpose-built for insurance
Each tool targets a specific actuarial workflow. Some use AI where it adds genuine value. Others are pure computation — because not everything needs a language model.
Policy Document Intelligence
Extracts structured data from unstructured policy documents — conditions, exclusions, and coverage terms surfaced in seconds.
Matching Adjustment Calculator
Full Solvency II MA calculation engine. Fundamental spread decomposition, asset eligibility, and benefit projection — no AI involved.
Assumption Setting Assistant
Drafts assumption papers from market data, experience analysis, and regulatory guidance — structured for peer review.
Mortality Calibration Tool
Fits Makeham and Lee–Carter models to portfolio experience data. Produces calibrated tables with confidence intervals.
Pricing Intelligence
Suggests GLM rating factor structures from historical loss data — identifies interaction effects and territorial patterns.
Annuity Data Converter
Paste unstructured deal text and extract structured annuity pricing fields — purchase price, income, mortality basis, and more — in seconds.
Actuarial Document Analyser
Upload any actuarial document and receive a structured TAS compliance review — severity-rated findings linked to specific clause references.
ORSA Narrative Generator
Drafts Solvency II Own Risk and Solvency Assessment disclosure narratives from quantitative inputs and board risk appetite.
Tetris
A classic Tetris game built with React. Take a break from actuarial work and clear some lines.
Approach
How I think about AI in insurance
Domain first, technology second
Every tool begins with the actuarial problem, not the model. The regulatory and commercial context shapes the architecture before any code is written.
Auditability is non-negotiable
Every AI-generated output links back to its source data and prompt chain. If an examiner asks "where did this number come from," there is always an answer.
Model selection is a deliberate choice
Claude for nuanced document work. GPT for structured reasoning. Gemini for data-heavy pattern recognition. The right model depends on the task, not the trend.
Augment the actuary, not replace them
These tools draft, suggest, and compute. The actuary reviews, judges, and signs off. Professional judgement remains where it belongs — with the professional.
About
I'm an actuary who builds software. My background is in Solvency II regulatory work — matching adjustment, SCR modelling, and annuity pricing — for UK life insurers. I now focus on where large language models and actuarial practice intersect: building tools that handle the repetitive, document-heavy parts of the job while keeping the actuary in control of every decision that matters.
The tools on this site are working demonstrations of that approach. Some use AI. Some don't. All of them reflect the principle that technology should serve the domain, not the other way around. If you're exploring how AI fits into your actuarial or pricing function, I'd welcome the conversation.
Domain
Technology
Contact
Let's talk about what AI can do for your actuarial function
Whether you're exploring AI integration, need a second opinion on an approach, or want to discuss a specific workflow — I'm always happy to connect.