The first build.
First I work out how your week actually runs. Then the one process that costs you the most, built as working software and handed to somebody on your team who owns it. First working result inside three weeks of starting.
Not an AI rollout.
If what you need is your whole workforce trained and using AI, that's a change management programme and I don't do it. I will say so in the first conversation and point you at someone who does. What I do is narrower and it's for one person: the principal, and the week they actually have.
Three working sessions with you, plus a look at the work itself in between, on days you pick. About five hours of your time in total. Most of what gets built after this is not what was asked for at the start.
Every recommendation carries a price and an elapsed time, and a view on what order to do them in. Including, sometimes, that nothing should be built.
Fixed scope, fixed price, no second discovery phase. The most expensive process in your month, working, on systems you already pay for.
A named person inside your organisation owns it at the end. I don't become a dependency you can't remove.
This is the operating system I run my life on. Yours would look like your work, not mine, and it gets handed to somebody on your team at the end.
The first build.
read, then build
$8,600A map of how your week actually runs, with a cost and an elapsed time against every recommendation. Three sessions, plus a look at the work in between.
One nominated workflow off that map, fixed scope and fixed price. No second discovery phase, no open-ended number.
Where people are already using AI unsupervised, and in most organisations some are, you find out early what's going where. It falls out of the looking anyway.
The rest of the map, documented well enough for your own people to work through, plus two sessions of office hours. Normally $1,900.
Slow if your data is a mess. I'll tell you that in week one, not week six.
the guarantees
both in writingIf it ends without at least three recommendations you could act on without ever speaking to me again, there is no fee.
If what I build isn't actually running in your week when we finish, I keep working at no charge until it is, or I refund you. Your election.
The record-keeping stays switched on, the decisions stay yours rather than mine, and you give me one hour a week while we're running, in person, not a delegate, capped at six weeks. Those three aren't conditions I invented to get out of the guarantee. They're the three things that decide whether any of this survives contact with your actual month.
One principal and their office. If the person whose week we are rebuilding is not the person in the room, it doesn't work, and I'd rather say that now than in week four.
Larger pieces follow only if you want them: a full workflow build (from $10,400), the whole operating system for a principal and their office, meaning capture, triage, decision record, commitments and follow-through, from $24,000, or a standing engagement from $7,500 a month with a cap on how many run at once. The full schedule is issued on request: ask for NRL-SF-2026-01.
Most of what gets sold as AI transformation is a tool nobody adopts, bought because the board asked. The thing that decides whether any of it survives isn't the model. It's whether the record-keeping stays on and whether the decisions stay with the person whose judgement they are. That's why those two are conditions on the guarantee rather than advice in a slide. I build the means of observation. You keep the judgement.
Want to talk it through?
And if what you actually need is your whole team trained rather than one week rebuilt, say so and I'll tell you plainly. It's not what I do.