I help firms with proprietary knowledge, data and evidence-heavy workflows build AI-enabled products and services that are useful, testable and stand up to client, legal and procurement scrutiny.
The gap
Many professional services firms are experimenting with AI, but few have a clear route from prototype to a defensible product.
Before taking them to market, they need to answer questions such as:
What I do
My work combines product thinking, AI workflow design and practical assurance, so professional services firms can build AI-enabled products that are useful, testable and ready to take to market.
What it answers
What should become AI-enabled, client-facing or accessible through AI assistants?
For example: internal tools, client portals, datastores, knowledge bases, documentation, training materials — shaped around your firm’s expertise, not what everyone else is doing.
Typical outputs
What it answers
How should the workflow, product, tool, agent or service actually work?
For example: custom GPTs, agent workflows, client-facing tools, MCP services, connectors or agents.
Typical outputs
What it answers
How is it tested, benchmarked, explained and made defensible?
Typical outputs
Evidence-led
Specialist firms cannot rely on vague claims such as “we use a good model”. Credible AI-enabled products need:
Defined tasks
A clear scope for what AI should and should not do
Clear success criteria
Evidence that outputs meet the job, not just look plausible
Source traceability
Answers users can verify back to underlying material
Version control
Known models, prompts and releases — not a moving target
Testing evidence
Golden cases, benchmarks and evaluation pipelines
Human review points
Where a person checks, approves or overrides
Understandable limits
Honest boundaries on what the system will not do
Client-ready explanation
Language that company leadership and sales teams can stand behind
This is where product enablement and assurance meet.
Proof
Retained as AI Assurance Lead to design the evaluation, testing and benchmarking approach for a legal analytics AI portal, including golden test sets, workflow controls and client-facing assurance materials.
Result
The product could be graded against defined cases and explained to clients, lawyers and procurement — tested and defensible rather than taken on trust.
Built a specialist planning appeal search product, turning unstructured public decision notices into a structured research tool for planning professionals.
Result
Tens of thousands of public decisions became a searchable, structured dataset, so professionals find and summarise relevant precedent in minutes instead of trawling raw notices.
Developed multi-model AI visibility benchmarking, using prompt sets, entity extraction and classification to measure how brands appear in AI-generated answers.
Result
Run at scale — 9,000+ AI responses across three leading models and 300+ organisations — producing benchmarks used as published research and white-label client deliverables.
Created the blueprint for a machine learning model trained and tested on thousands of human-generated metadata records, to automatically classify and categorise client-specific topics in news articles.
Result
An AI-assisted, human-in-the-loop process integrated into internal workflow tools – saving analyst time and improving accuracy.
Developed a machine learning model to predict corporate litigation before it happened – providing an early warning system for publicly-listed companies, and a prospecting pipeline for law firms.
Result
A risk assessment tool for publicly-listed companies and a prospecting pipeline for law firms.
AI/ML product strategy, analytics products, B2B technology positioning and go-to-market for specialist software and data businesses.
Result
Multiple products taken from concept to market across specialist software and data businesses.
Background
I’m an independent AI product and assurance consultant with experience across legal analytics, professional research tools, AI benchmarking and B2B technology strategy. I work with specialist firms that need AI-enabled products to be useful in practice and defensible in front of clients, lawyers, regulators and procurement teams.
Fit
I focus on firms with specialist knowledge, valuable data, evidence-heavy workflows or existing products that could become more capable through AI.
Next step
If you are building, evaluating or commercialising an AI-enabled product or workflow, I can help you decide what to build, how it should work, and how to make it defensible enough to take to market.