What a sponsor should ask before approving an automation or AI programme

1. Which business outcome is this programme for?

Choose the outcome before you choose the technology. An innovation project can serve six different objectives, such as protecting market share, improving operational efficiency or retaining staff. Most companies pursue more than one, and each needs its own strategy. Our Registered Management Consultants agree the objective with you before any design work begins. Our approach page shows how.

2. Can the data support the decision?

Automation is only as good as the data behind it. Before we build, we test each decision in scope for data coverage, freshness and trust. The result may be a recommendation not to automate a decision yet. An integrator that never gives you that advice has not looked at your data.

3. What do we commit to, and when can we stop?

You do not buy the whole programme at once. After the site assessment and preliminary design, we issue a fixed-scope proposal that sets out the full scope, cost and sequence. Robotics and assurance engagements then run in six phases, and each phase ends at a gate with a written test or result. You buy the next phase only after the gate is passed.

4. Who is accountable for what the AI decides?

For any system that contains an AI model, we define what it may decide on its own, what is logged, who reviews it and how it is withdrawn. In our agentic AI work, people keep the decisions that carry risk. In the quoting desk, code sets every price and a manager approves commitments. In the voice banking agent, a policy engine can hold an action and a person can take over. Our practice is aligned to ISO/IEC 42001:2023. Our assurance service can prepare your own organisation for the same standard.

5. What happens in month thirteen?

On final payment, you own the delivered system outright, including the source code and configuration. Handover includes documentation and training, so your team can run the system without us. If you want ongoing support, the proposal states the support scope and response time before you sign. Our FAQ covers ownership and continuity in full.

What we can show today

These results are recorded in our case studies as measured or observed. For most of our other projects, the case studies say that no result was quantified. We do not publish a financial return for any project yet. If you need one, we will agree with you in the scoping stage which measure the project is tracked against.

• Raymang Eggs, robotic carton removal line. Unboxing fell from 24 hours to 1 hour for the same stock. An Enterprise Development Grant funded the line.
• PSA Marine, biometric fuel management system. In continuous use since 2019. The second refresh was completed in 2025.
• AETOS, contactless temperature scanner. More than 1,200 units deployed. Listed as a Class B medical device after certification by the Health Sciences Authority in November 2021.
• PestBusters, industry transformation programme. The service changed from routine trap checking to rodent risk management. The client was acquired before phase three.

What we cannot yet show

Our agentic AI case studies cover 15 confidential clients. Most are demonstrations on test data, and each case study states its own status. They report test results, such as scenarios passed and attacks resisted, and they report failures we found and fixed. They are not evidence of business value in production. Treat them as evidence of how we build and test, not of what a deployment has returned.

We are a team of around 15 people, in operation since 2015. Ask us on the scoping call how a programme of your size would be staffed.

How to start

Book a scoping call. We tell you the cost of the first two stages on that call, and we reply within 24 hours.