AI Services by Engagement Stage
The catalogue tells you what each service costs. This tells you what is inside it, what sits outside it, and how to tell which one you are actually ready for.
Start with a conversation
Most people arrive with a workflow they have stopped defending. The report that takes a week. The spreadsheet three people maintain because the system will not. The handover that only works because one person remembers it.
A first conversation is free and it is not a pitch. We listen, play back what we heard, and say whether AI is the right tool for it. Sometimes it is not, and we say so.
Scoping: two depths, and how to pick
Everything paid starts with scoping, because building the wrong thing quickly is still building the wrong thing. There are two depths and the difference is how much ground has to be covered before anyone can commit.
An AI Working Session suits one workflow and one team. Half a day to a full day, your people in the room, on your actual work. We separate what AI can move now from what needs a process fix first, then scope and cost a contained build. If you go ahead with that build, the session fee comes off it.
An Operations Assessment suits several teams or several systems, where the diagnosis is itself real work. Stakeholder interviews, a systems and data audit, opportunities ranked by payback and effort, and a costed roadmap you can take to a decision.
The honest test: if you can name the one workflow, book the working session. If you would have to name six, and they argue with each other, book the assessment.
AI diagnostic scope · AI Working Session · Operations Assessment
Building: one thing, working, with one team
A Pilot build puts one working tool into production with one team. Not a prototype, not a demo. Built on your data, inside your own tenancy, on fixed scope and fixed dates.
Most of it is already done before we start. These are patterns we have shipped before, so what you pay for is the part that is genuinely yours, not a platform someone else is renting you.
When the whole operation is the problem rather than one workflow, that becomes a 90-day transformation: the data plumbing, the integrations and the lifecycle of a record end to end, then your team trained to run it. No licence, no lock-in.
After the build
Two things tend to follow. Some organisations have a body of knowledge that never reaches the people who need it, and want it searchable with answers cited back to the source document. Others want us close by while their team takes over, which is an embedded retainer, not a support contract.
Training runs alongside all of it. Most of our clients meet us that way first, and teams that have been trained make better build clients because they already know what is hard.
Private RAG over your library · Embedded retainer · Hands-on workshops
What we say no to
We do not take work where AI is the answer someone has already decided on and the question is still missing. We do not sell seats in a platform. We do not stay longer than the work needs, because the point is that your team can run it without us.
If what you actually have is a process problem wearing an AI costume, we will tell you that in the first conversation, at no cost.
Tell us where it hurts
A free first conversation. We play it back and tell you straight whether AI fits, and where.
Tell us where it hurts