Services

Start with the decision. Build only what changes it.

Three ways to work together, from a fixed-scope diagnostic to a system in production to ongoing advice. Each one links to something you can open and try, so you can judge the work before we talk.

01

Decision diagnostic

Two to three weeks · fixed scope

Find the decision worth improving, what it is costing today and what to build first.

  • The decision framed: owner, cadence, success measure, guardrails and baseline
  • A data audit against that decision, not against a wish list
  • The prize sized at realistic adoption, with a go or no-go
  • A build plan your team can run with or without me

For Leaders with a budget for data or AI and no shared view of where it should go first.

See it working
02

Build a decision product

Eight to sixteen weeks · one system in production

One decision system built, shipped into the workflow, instrumented and handed over with your team alongside.

  • Forecasting, pricing, prioritisation, recommendation or an AI assistant
  • Data contracts and automated pipelines that fail safely
  • Recommendations with reasons, inside agreed guardrails
  • Monitoring, retrain policy and a model card from day one

For Teams who know the decision and need it working in production, not in a notebook.

See it working
03

Measurement & AI advisory

Monthly · judgement on call

Proof that the work pays, and a team that knows when to trust the model and when not to.

  • Experiment design and read-outs the business believes
  • Value tracking at measured adoption, reported quarterly
  • Human–AI workflow design, overrides and reliance
  • Mentoring for analysts and data scientists

For Teams who own delivery and want a senior second opinion on what is working.

See it working
How the work runs

The same seven stages, every time.

Whether the engagement is a diagnostic or a full build, the work moves through the stages shown in the flagship product. Click any stage to see it done.

  1. 01 Frame One decision, an owner, a success measure, guardrails and a baseline to beat. Decision diagnostic
  2. 02 Data A data contract with checks that block a bad refresh, and why raw price data misleads. Build a decision product
  3. 03 Model A backtested demand forecast and causal price elasticities by segment. Build a decision product
  4. 04 Decide A recommended rate inside guardrails, with reasons a revenue manager can check. Build a decision product
  5. 05 Prove A switchback test, powered properly and read out against a pre-registered rule. Measurement & AI advisory
  6. 06 Run Accuracy and drift monitoring, retrain triggers and a model card. Build a decision product
  7. 07 Value Value at measured adoption, payback and a delivery plan. Measurement & AI advisory
Before we start

Questions people ask.

Which industries do you work in?

Most of my production work has been in financial services and lending, and the methods carry across: forecasting, pricing, prioritisation, recommendation and experimentation work the same way in retail, hospitality, insurance and B2B sales. The flagship product is set in a hotel on purpose.

How is the work priced?

The diagnostic is a fixed price agreed up front. Builds are scoped in phases, with a go or no-go after each one. Advisory is a monthly retainer. You always know the cost before the work starts.

What do you need from us?

A decision owner who can give an hour a week, access to the data behind the decision, and someone on your side who will run the system after handover. I work alongside your team rather than around it.

What about our data?

Your data stays in your environment, under your controls. Nothing from an engagement is reused elsewhere, and every demo on this site is built on synthetic or public data.

Do you only build models?

No. Often the highest-value work is upstream (framing the decision, fixing the data contract) or downstream (the workflow, the test, the adoption). The model is one of seven stages.

Start a conversation

Most engagements start with a two-to-three-week diagnostic.

Bring the decision you want to improve. You will leave the first call with a clear view of whether it is worth doing and what it would take.

Discuss a diagnostic
Decision Lab · live demo
Full screen ↗
Starting the demo…