Find out where AI would genuinely pay back in your business, what it is worth in pounds against what it costs, what to fix before you automate, and which three things to try first.
Where does our time go?The recurring tasks, from the people who do them.
What could AI take on?Matched to common AI patterns, with cautious starting assumptions.
What is it worth, against what it costs?Ranges in pounds, with payback in the low case.
What do we do first?Three pilots, each with a way to prove it.
Most AI advice starts with the tools. This starts with the recurring work your team already does, puts a cost on it, and only then asks where AI fits. Choose this to value AI task by task across your teams. To check readiness first, see the AI Readiness Assessment; for whole processes across people and systems, the Workflow Automation Map.
Founder pricing. The method, the calculation spreadsheet with 19 built-in checks, and a complete worked example are built and tested. The first client for this product gets the founder price, in return for permission to measure the results and, if they are good, write them up as a case study. After that, the price goes up. Book a call to take the founder place.
The recurring work across your teams: how often, how long, who does it and what that time really costs, in a spreadsheet with 19 built-in checks, summed up on one page for your team or board.
Cautious by design: for each task, what AI could realistically take on, as a low-to-high range on cautious starting assumptions, set against set-up and running costs shared fairly between the tasks that use each tool. Payback uses the low case, so these are numbers you can defend.
Pilot now, plan, bundle or leave for now, on a value-against-ease matrix you can explain to your team. Tasks that shouldn't exist go to Simplify first, so you don't pay to make waste faster.
Each with a baseline, a success measure, an owner and a clear rule for when a person checks the AI's work.
Every task involving personal data is flagged, and anything sensitive is held back until the right safeguards are in place.
Guided and done for you: in your own accounts, it drafts the interview invitations and one reminder for a person you name to send, logs the replies and drafts task rows from the interview notes for checking, then reminds the pilot owners to measure their baselines and prepares the re-run after the pilots. Nothing leaves your business, and no decision changes, without one of your people approving it.
A fictional industrial distributor with 23 staff. We mapped 26 recurring tasks across 9 teams, taking about 5,200 hours a year.
On cautious starting assumptions, AI could release £23k–£45k of that time each year. Three pilots came first: the MD's inbox, quotations and customer questionnaires, each paying back within 5 months of released time in the low case. Four tasks were flagged to simplify before anything is automated, including re-keying web orders that should arrive in the system automatically.
Short interviews with the people who do the work, using the interview guide in the kit, capture each recurring task, how often and how long.
Which AI pattern fits, how clear the rules are, what a mistake would cost, what data is involved, and whether it is waste.
Savings as ranges, set against the cost of the tools; every task lands on the matrix with a next step.
Measure a baseline, agree how success will be judged, and decide who checks the AI's work.
The agent reminds the pilot owners to measure their baselines and, two weeks after the pilots end, prepares the re-run of the map with your measured results beside the assumptions, for your people to approve.
The same method and the same spreadsheet in every option. You choose how much of your own time goes in.
For owners comfortable with a spreadsheet and an AI assistant.
For owners who want a second pair of eyes on the data and the analysis.
For owners who would rather hand over the data and get the answer.
It is built on the open Agent Skills standard and designed for Claude (any plan, with code execution switched on). A paste-in version is included for ChatGPT, Gemini and Microsoft Copilot; it has not yet been tested on them.
The calculations live in the spreadsheet, not in the AI, so the numbers don't depend on which assistant reads them.
Book a short call to talk through your business and which way of buying the AI Opportunity Map suits you.
Book a call