Small first, and priced second.

No deck. We ask about one live project — its phase, what is slow, what has already been tried — and you get a straight read on where this pays in a firm shaped like yours. If the honest answer is that it doesn’t pay yet, that is the answer you get, and it costs you half an hour.

• One live file and whoever actually owns the number
• A grounded read, including where AI is the wrong tool
• Either a readiness audit, or nothing — both are fine

The audit is not a technology review. It works backwards from goals you already have, finds what is actually blocking them, and maps that to KPIs somebody on your board already owns — because a project measured against a number nobody owns is how most of the 95% happens. You end with targets and a recommended direction, and sometimes that direction is a tool we don’t sell.

• Map the process as it runs, not as the org chart says
• Rank the leverage points by what they are worth, not by ease
• Targets, a direction, and a scope you can say no to

One leverage point, built, in days rather than quarters. It is scoped against a KPI and a date instead of a feature list, and the date carries our risk as well as yours: miss it and we take 20% off. That clause exists because it changes what we are willing to promise, which is the only kind of guarantee worth reading.

• One workflow, end to end, in production — not a pilot
• Tied to a named KPI and a named date
• Miss the date and the fee drops by 20%

Only after the first three have gone well. Thirty years of your firm’s judgement — why you passed on that site, what that ward always asks for, which contractor’s numbers you halve — held as a system your people query and keep adding to. The point is not automation. It is that the knowledge stops leaving the building when someone retires.

• Your data, your schema, your instance
• Grows from the work rather than from a migration project
• The most expensive thing we do, and the last