None of them are about technology. They are about how your business runs today, because AI can only speed up work you can already describe.

1. Do you know which jobs made money?

Ask yourself how long after a job finishes you know what it actually made. If the answer is months, or only when the year-end accounts come back, start there. AI can pull costs, valuations and timesheets together far faster than a person, but only if they are recorded somewhere it can reach. If your job costs live in three spreadsheets and the QS’s head, you will just get the wrong numbers faster.

Ready if: within a month of a job closing, you can say what it made against what you priced.

2. What still leans on a handful of key people?

Write down every price, sign-off and decision that waits on a named person in a normal week: you, a director, the senior estimator. That list is where the business stalls when one of them is on site, on holiday or off sick. It is also where AI earns its keep: drafting the quote, chasing the missing number, preparing the decision so the right person only makes the call. But it can only follow rules someone has written down. If the rules exist only in people’s heads, writing them down is step one.

Ready if: you could hand a new starter the rules for at least one of those decisions on a single page.

3. Where does the same information get typed twice?

Follow one job from enquiry to final invoice and count how many times the same details get handled again: into the quote, the order, the programme, the valuation, the invoice. Every repeat costs hours and is another chance to get it wrong. These handovers are usually the quickest wins, because the information already exists.

Ready if: you can name the two or three handovers where repeat work costs your team the most time.

4. Is your information in the business, or in people’s inboxes?

Drawings, rates, supplier prices and job files often live in personal inboxes and on individual laptops. AI works from records. If the information walks out of the door when someone leaves, the first job is getting it somewhere the business owns. That is worth doing whether or not you ever use AI.

Ready if: the files for a live job sit in one shared place that anyone working on it can find.

5. Who would own it?

Every change needs a named person with time in their week to run it, check it and say when it isn’t working. Not “the office”, and not you. A tool without an owner gets used for a few weeks and then quietly dropped.

Ready if: you can name that person today, and they would agree.

What your answers tell you

Three or more confident answers and you are ready to pick a first workflow and put a number on it. Fewer than three and the fix starts with the basics: job costing, written rules, one place for files. That is still progress, and it is far cheaper to learn it now than after someone has sold you software.

If you want a scored view rather than a gut feel, Are we ready for AI? asks 24 quick questions about your team, systems and how work gets done, and takes about five minutes. You find out whether the business is ready for AI or needs work first, what to sort first and where to go next, with a three-page report to keep.