A 30-day AI diagnostic, one day of your time

The question people ask me right before saying yes is almost never about the price. It is: how much of my team’s time is this actually going to take?

It is the right question, and it is also the one most providers answer badly, because an honest answer sounds like an admission of weakness. The head of a mid-sized company has no reserve of person-days sitting idle. If a diagnostic requires department heads to free up two days a week for a month, it will not happen, and rightly so.

So here is the figure, before anything else.

Three milestones, and nothing else to prepare

Over thirty days, what your organisation has to give is three fixed meetings, plus a few hours of interviews with the people doing the work.

Framing, day zero. Ninety minutes with you. We set the scope, the processes to look at, the people to meet. This meeting decides the value of everything that follows: too wide a scope produces a flat map, too narrow a scope misses the process that actually costs the most.

Checkpoint, day fifteen. One hour. I show the map as it stands, you correct what is wrong. It is the most profitable of the three milestones, and the one most comparable engagements leave out: a dead end dropped here never gets costed for nothing, and a misunderstanding gets corrected before it contaminates the roadmap.

Readout, day thirty. Two hours with the leadership team. Costed roadmap, discarded use cases with the reason they were discarded, decision. This is not an end-of-engagement presentation, it is the moment you make the call.

Between the first two milestones sit the interviews: forty-five minutes per person, five to eight people depending on the size of the company. Not a full-day group workshop. Short conversations, at their desk, on their real files, because the process described in a meeting room is never the one that actually runs.

Total for you: roughly one cumulative day over the month. Half a day per person interviewed.

The rest is my job. Collecting volumes, measuring real end-to-end lead times, writing the case files, costing, ranking, drafting.

Why this exact balance

There is a reason for this split, and it is not a commercial one. There are two ways to get a diagnostic wrong, and they are symmetrical.

The first is to demand too much. A diagnostic that takes two days a week from department heads produces a document nobody reads: everyone has had enough of the subject before the readout, and the deliverable lands in an organisation already tired of it. The real cost of that diagnostic is not its price, it is the month of output it consumed.

The second is to talk to nobody. A diagnostic run purely with the leadership team and two data extracts produces a document that is wrong. It describes the official process, the one in the procedures manual, not the one someone actually runs every morning with its workarounds, its parallel spreadsheets and its manual re-entry. A business case built on that process does not survive first contact with development.

The right balance is the only one that goes after ground truth without freezing the company. Three leadership meetings, short interviews at the desk, and all the analysis carried by the provider.

The side effect I do not sell

One thing happens every single time, and it appears on no proposal.

The people interviewed already know, to a large extent, what is wrong. They have a precise idea of what wastes their time, of where they redo by hand what the tool should have done, and of the steps nobody can justify any more. This is not a consultant’s discovery. It is information that already existed inside the company, and that nobody had ever asked for while writing the answers down.

It is also what settles the question of resistance to change before it is even raised. An automation decided elsewhere, on a process described once in a meeting, runs into the people who execute it. A map built from what they said, and shown to them before it is costed, changes their status: they are no longer the object of the project, they are its source.

Three meetings, a few interviews, a documented decision at the end. That is all it takes from you, and deliberately all it can take.