How we engage

From AI idea to production, one decision at a time

Most AI work stalls between the demo and production. We break it into three engagements, each a defined piece of work with a deliverable, so you decide on evidence at every step instead of signing up for an open-ended program.

Step 1

AI Use Case Discovery Workshop

A working session that turns your list of AI ideas into one use case worth building, and a plan to put it into production.

Who it's for: Teams with more AI ideas than decisions: customer service, search or commerce leaders under pressure to use AI, or IT leaders asked where to start.

What you get

A ranked shortlist

Use cases scored on value, data readiness and effort, agreed by business and IT.

One use case chosen

A named owner on your side and a success measure set before anything is built.

A build proposal

Scope, architecture and the cost of the next step, in writing.

How it works

1Prepare. Your goals, systems and data reviewed with the people closest to the work.
2Workshop. Business and IT in one room. Ideas scored and ranked.
3Assess. The data behind the top ideas checked, architecture sketched.
4Propose. A written proposal for one use case, ready to go to a proof of concept.

Step 2

AI Proof of Concept

A working prototype of one AI use case, built on your data in Google Cloud, so you decide on evidence rather than a demo.

Who it's for: Teams with a use case picked, such as a customer service agent or site search, who need to show it works before asking for budget.

What you get

A working prototype

Built on Gemini in your own Google Cloud environment, not a vendor sandbox.

Measured results

Tested against success criteria and a test set agreed on day one.

A clear decision

Go or no-go on a pilot, with the business case behind it.

How it works

1Scope. Use case, success test and data in Google Cloud agreed.
2Build. Gemini grounded on your data and the prototype built, with your team involved.
3Test. We run it against the agreed test set and tune what falls short.
4Decide. A readout of what worked, what didn't, and what a pilot would take.

Step 3

AI Pilot

One AI use case running in production, on live data, with real users, so you see what it does for the business before you scale it.

Who it's for: Teams with a proven prototype, or an AI project stuck short of production, ready to put it in front of customers or staff.

What you get

Live in production

Running on live data and connected to the systems your team already uses.

Usage measured

Adoption, accuracy and business impact tracked against an agreed baseline.

Ready to scale

Security reviewed, monitored, and a plan for the next use cases.

How it works

1Plan. Users, integrations and success measures agreed. Owner named.
2Build. Connected to your systems, hardened, and through security review.
3Launch. Live with a defined group of users on real data, monitored and tuned.
4Review. Results against the baseline, handover to your team and a plan to scale.

Not sure where to start?

The AI readiness assessment scores your data, platform, use cases, governance and operations in about four minutes. For data and platform work, a scoped assessment produces a migration plan and a cost model you can take to your CFO.

Score your AI readiness

Funding

Each of these engagements may be eligible for funding. We check eligibility before you commit to anything, and you see what is covered, in writing, first.

Start with a scoped engagement, not a proposal.

A workshop to identify where the value is. An assessment to plan the work properly. A pilot to prove it in your environment before you commit. Each one is a defined piece of work with a deliverable, not an open-ended statement of work.

Book a working session