The AI Readiness Map
Is your business ready for AI?
Fifteen questions about one process you already run. It draws that process back to you, draws the version a system would leave behind, and prints the arithmetic underneath both. It can also tell you not to buy anything, and for most businesses that is what it says.
The formula, in full
How is the readiness score calculated?
Every figure this page produces comes from the arithmetic below and from nothing else. There is no sector multiplier, no benchmark we cannot source, and no number borrowed from a client we are not allowed to name. The score is five dimensions worth twenty points each, and the money is four lines of multiplication.
hours_lost = affected_people × hours_each
recoverable_share = 0.55
− 0.10 if the work touches 4 or more systems
− 0.10 if a previous attempt stalled
+ 0.05 if a previous attempt worked
clamped to 0.25 … 0.7
hours_back = hours_lost × recoverable_share
annual = hours_back × your_hourly_cost × 46 weeks
shown as annual × 0.6 … annual × 1.4In words: we multiply the number of people who touch the process by the hours each of them spends on it, take the share we think is realistically recoverable, multiply by the hourly cost you told us and by 46 working weeks, and then show the result as a range from 0.6 to 1.4 times that figure rather than as a single number.
The 55 per cent starting point is AGMM's working assumption, not a measured client average. We have not run enough measured before-and-after studies to publish an average, and any firm quoting one at you should be asked where it came from. 46 working weeks is a convention, not your calendar. The range exists because the honest width of an estimate built from recollection is wide, and a single number would imply a precision this does not have.
When you finish the map, three reasons this number might be wrong are printed under it, generated from the answers you actually gave.
The limits, published
What this score cannot tell you
Nobody else in this field publishes this list, which is the main reason it is here. A diagnostic that will not say what it cannot see is a sales form with a progress bar on it.
- Whether the hours you told us are real. They are a recollection of your own week, and hours people estimate about themselves are rounded. That single input moves every pound figure on the page, and it is the reason the audit is fourteen days of measurement rather than a longer version of this form.
- Which process is actually your constraint. This prices the one process you answered about. It has no way of knowing whether a different one is costing you more, and in about half the businesses we talk to it is.
- What a build would cost you. Recoverable hours are not a budget. The price of a build depends on what your systems will let us connect to, and no form can see that.
- Whether your data is good enough. Two answers here ask about records and both are self-reported. Records that are inconsistent enough that no system could read them reliably are their own piece of work, and they are the most common reason a build takes longer than it should.
- Whether your team will use it. The fifth set asks, and the answers are worth having, but an owner predicting how their own team will react to a bad fortnight is the least reliable measurement on this page.
- Anything about revenue. Every figure here is time and the cost of time. Work that gets displaced onto other people when this process goes wrong is not counted, and neither is a single pound of revenue.
Common questions
Questions about the readiness assessment
Is the AI readiness assessment free?
Yes, and it does not ask for an email to show you the result. Fifteen questions, a score out of 100, a map of the process you described and the three to five things to do before any AI work. The email is only asked for afterwards, if you want the steps sent as a four-week plan.
How long does the readiness assessment take?
About five minutes. Two setup answers and then five sets of three questions. Your answers are kept in your own browser so a reload does not lose them, and nothing is sent anywhere until you choose to send it.
How is the readiness score calculated?
Five dimensions worth twenty points each: repetition, data, systems, process, and people and appetite. The hours and the money use a published formula: affected people multiplied by hours each, multiplied by a recoverable share that starts at 55 per cent and moves with your answers, multiplied by your own hourly cost and 46 working weeks. The whole thing is printed on the page.
Where does the 55 per cent figure come from?
It is AGMM’s working assumption, not a measured client average, and the page says so in those words. It moves down ten points when the work touches four or more systems, down another ten when a previous attempt stalled, and up five when one worked. It is clamped between 25 and 70 per cent.
Will it ever tell me not to buy anything?
Yes. Unless you score in the top two bands and can confirm that the owner and the person who runs the process can each give three hours in the next fortnight, the result says do not buy the audit yet and gives you the steps instead. Where your answers support it, the recommendation is to use Claude or ChatGPT with a written down process rather than build anything.
What does the AI Constraint Audit cost?
The audit is £2,000 for fourteen days and the fee is never deducted from a build, so nobody in the room has a financial reason to conclude that a build is the answer. The 30-minute fit call before it is free.
Which industries does the readiness assessment cover?
Twelve, and each one has its own question bank rather than a shared set of generic questions: construction and £1m plus contractors, accountancy and professional services, logistics and distribution, legal, manufacturing, recruitment, wholesale, property and lettings, private healthcare and clinics, agencies, e-commerce and retail operations, and multi-site service businesses. The niche changes the wording and the process it draws. It never changes the arithmetic.
If you want a person to look at it
The fit call is thirty minutes and it is free
If the map says you are ready, the next step is fourteen days of measurement: the AI Constraint Audit, £2,000, fixed, and the fee is never deducted from a build. If it says you are not, the steps are the next step and they cost nothing but the hours.
Either way, a call is free and we will tell you in it if we think you should not buy anything. That answer costs us the sale and it is the reason the rest of this page is worth reading.
Book the free fit call · What everything costs · The work we can show you