Start with the work. Not the AI.
The useful first question is not which model to buy. It is which process still depends on memory, retyping, and whoever happens to be at the desk.
Most AI conversations start in the wrong place. A vendor names a model. A slide names a transformation. The room nods. Nobody has named the job that already hurts on a Tuesday.
Spirality starts the other way around. We assemble AI around the business — which means the process comes first, and the model is a part, not the product. If you cannot point at the work, you are not buying a system. You are buying a demo with a kickoff date.
Name the process before you name the tool
A process you can name has inputs, exceptions, write-backs, and a person who currently owns the failure. Invoices that do not match. Leads that die in a spreadsheet. Support replies that are the same eight answers, copied by hand. That is the map. Tools come after the map.
If the pain is “we need AI,” pause. Ask which queue, which document, which follow-up. Use-case patterns exist so a buyer can recognise their own work without a glossary. If none of them fit, the work can still be in scope — it just has not been written up yet.
Three depths. One honest choice.
Once the process is named, depth is the next decision. Not the model. Depth is whether you need clarity, a built system, or someone to keep the system alive.
- Consulting alone — figure out what to do, and what not to do.
- Consulting plus a build — design and put the system in.
- Consulting, build, and aftercare — keep running it: support, servers, monitoring, the unglamorous part.
The wrong depth wastes more money than the wrong model. The aftercare question is how we talk about that without turning it into an upsell script.
What “good” looks like without invented numbers
We will not promise a percentage. We will not invent a case study metric to make a page feel finished. Good looks like this: the process is clearer, the exceptions have an owner, and the system writes back to tools you already use. People still decide the hard bits.
How to use this cluster
Read what to bring to the first conversation if you want a practical brief. Read why demos fail at a desk if you are being sold an agent. Read organisational memory if every new tool in your company starts from zero.
Then talk to us about one process. Not a transformation programme. One process.
Keep reading.
Bring one process to the first conversation
A useful first call is not a tour of models. It is one process, the depth you think you need, and an honest list of what you do not know yet.
Read the note
Demos aren't desks
An agent that dazzles in a slide deck still has to survive your exceptions, approvals, and Tuesday afternoon volume.
Read the note
The aftercare question
Consulting, build, or build-and-run — the wrong depth wastes more money than the wrong model.
Read the note