CADi is intelligence with integrity: a control layer that sits between a model's recommendation and the system's action and decides, per case, whether the machine acts alone, acts with a human, or hands the decision to a person. It records the reason, the evidence, the policy and a hash-chained audit trail for every decision.

Is CADi in production?

Yes. CADi is built and running in production, with nine UK patent applications pending. This website is a demonstrator: it runs the engine live in the browser on a simulated payments datastream, so nothing here routes on real customer data. CADi participates in cohort 2 of the FCA's Supercharged Sandbox, an AI-focused programme in which firms experiment in a secure, controlled environment; participation is not FCA approval, endorsement or authorisation.

Does CADi replace human accountability?

No. CADi supports, evidences and governs decisions; it does not replace the accountable human. It decides who decides, and vulnerable customers always reach a human.

THE TEAM

Phil Baker

CO-FOUNDER · MATHS ARCHITECT

Phil is a Co-Founder of CADi Limited, where he is pioneering his new approach to decision intelligence, based on causal inference and confidence measurement - an alternative to the predictively biased approaches that have caused disproportionate harm to the most vulnerable in society.

He brings more than 20 years’ experience leading transformation across financial services, insurance and customer operations, including senior roles at NatWest Group, Barclaycard, Travelport, Computershare and Endsleigh.

His work on Consumer Duty, vulnerable customers and Ai-enabled operations exposed a persistent gap: predictive systems could estimate what might happen, but could not explain what a decision caused or prove that it was justified.

At CADi, Phil applies causal inference to verifiable decision intelligence, helping regulated organisations improve outcomes, strengthen trust and show why a decision was appropriate when it was made.

Phil also looks after university and partner relations while CADibrings its Sales Director on board.

Ste Spence

CO-FOUNDER · BUSINESS ARCHITECT

Ste is a hyper-curious, evidence-driven technologist who has spent almost four decades helping organisations solve complex problems through better systems, better measurement and better decisions.

Never confined to a single industry, his career has taken him from recording studios to global contact centres and everything in between, designing auditable, statistically controlled decision systems for environments where precision, governance and disciplined execution are essential.

Today, Ste is immersed in the world of causal inference, building the decision intelligence that moves organisations beyond correlation to understanding cause and effect. He believes the best ideas should always be challenged, holding strong opinions lightly and welcoming new evidence that overturns established thinking.

At CADi, Ste is focused on creating practical decision ecosystems that help organisations take the best next steps, learn continuously and build trust through evidence rather than assumption.

ABOUT

CADi keeps a human accountable for the decisions that need one.

As Ai moves from advising to deciding, CADi decides, case by case, who decides: the machine alone, the machine with a human, or a human entirely. And it can show why.

01WHY WE EXIST

Ai is moving from advising to deciding. Agents now pay, refuse, approve and escalate. In high-consequence work, that can quietly remove the accountable person.

Most systems rely on opaque confidence thresholds. Our mission is to keep a human accountable where needed, making every choice between human and machine principled, legible and auditable: case by case, on the evidence, and shown.

02WHAT WE MAKE

Everything runs on the CADiTruth Engine (“Powered by CADi”): nine layers, a do-calculus core, and a hash-chained audit trail. It estimates what each action changed, what a better choice could have delivered, and who decided under what authority. That is decision integrity, made operational.

Each product applies that engine to one environment:

CADi PayTHE PAYMENT EVENT
Authorisation, fraud, routing, retry, and the treatment of customers in difficulty.
CADi ProofEXTERNAL ATTESTATION
Regulatory evidence, any regulator, any jurisdiction.
CADi PromptTHE PROMPT EVENT
Causal prompt-injection detection for LLM applications.
CADi AgentTHE AUTONOMOUS AGENT
Causal governance for multi-agent enterprise systems.
CADi PlayerTHE ONLINE GAMING PLAYER
Protection for the player, the community and the game across Care, Integrity and game protection.
CADi CultureORGANISATIONAL ALIGNMENT
Board-grade evidence that decisions track stated purpose and values.

03HOW WE RUN THE COMPANY

Our work turns do-calculus, counterfactual estimation and identifiability into something a business can run. We start from your data, producing analysis you can act on before CADitouches a live decision.

We run CADi on the discipline we sell: Ai produces, enforcement gates check every claim, and a human approves. People own accountability; machines own production.

Two rules are non-negotiable: a human owns irreversible acts, and vulnerability always outweighs efficiency.

04WHAT WE ARE WORKING ON NOW

Payments first. CADiPay is the Truth Engine’s first application, shadowing the live system and pricing opportunity cost before touching the decision path.
A working demonstrator, honestly labelled. This site runs simulated transactions through the live pipeline and causal graph. Nothing routes on real data.
The evidence layer. Regulator-readable packs and audit trails make governed decisions defensible.
The category conversation. We ask what a better credible choice could have changed, not only what the system chose.