CPD lab · Non-finance, Controls
Spot the fake invoice: a requester's check
Ten invoices and emails reach your department. Decide which are fine, which need a check first and which are scams, with your AI tool, before you approve anything.
Comes with Raising and approving spend: procure-to-pay for requesters and Spotting fraud and deepfakes: the staff version, and with the year. 29 labs open now.
The brief
You are Maya Hendry, centre manager at Wrenfield Community Leisure, a UK leisure trust. You are a budget holder: you ask for goods and services, and finance pays the invoices only after you approve them. It is Monday 16 November 2026. Ten invoices and emails have reached you over the past two weeks.
The pack has the ten messages, the invoice details, the list of suppliers we hold on file, the order log with what was delivered, our rules for approving invoices and some notes from finance and a colleague. Every company, person and figure is invented, so you can paste it into any AI tool. You do not need accounting knowledge.
Use your AI tool to check every item against the rules and the order log, then decide for each one. Pay nothing, reply to nobody and change nothing: your job is the table.
Decision, one per item. APPROVE: every test in the rules is met, so finance can pay it. CHECK_FIRST: the supplier is one we use and the invoice looks real, but one fact does not match or has not been confirmed, so you or finance must ask the supplier or a colleague before it is approved. REPORT: the item bears the signs of a scam, so it is not paid and goes to finance. The signs are: it did not come from the email address on our supplier list (even by one character); it asks for payment in a hurry or in secret, or from outside the system; or it is from a company we have never bought from for something nobody ordered.
Give one line for each of the 10 items, in the table columns item_ref, decision and reason. The reason is one sentence saying what you found and who you would contact.
Waving through a scam costs money, but so does stopping genuine invoices: suppliers stop deliveries when they are not paid. Use CHECK_FIRST and REPORT only where the evidence supports it.
Then write a short review note: what the AI got wrong, what you checked, and who you would tell.
Objective: To be able to use an AI assistant to check invoices and payment requests against what was ordered and delivered, recognising the common signs of invoice fraud, and to refer doubtful items to finance.
What's in the pack
Every name, figure and email in the pack is invented. Use your own AI tool: Claude, ChatGPT, Copilot or Gemini.
emails-and-invoices.txt
The ten emails and invoices received, 2 to 13 November 2026
invoice-details.csv
The ten items in a table: who sent them, the total, the bank details shown
supplier-list.csv
The suppliers on file: email domain, phone number and bank details
order-log.csv
What we ordered, the agreed price, and what was delivered or done
approval-rules.txt
How we approve invoices, for budget holders
finance-and-colleague-notes.txt
Notes from finance and from the facilities supervisor
The traps
The pack has 8 deliberate traps: cases where AI tools often get it wrong and a person must check. We don't list them here. Finding them is the job, and your feedback names each one you missed.
What you submit
- Your result, as a table with the columns item_ref, decision, reason.
- The prompt you used.
- A short review note, 150 to 350 words: what the AI got wrong, what you checked, and what goes to a person.
How it's marked
- Your result, checked line by line against the answer key60%
- Your review note, marked on traits40%
The review note's traits:
- Scam signs spotted
- Genuine invoices approved
- Checking first and reporting kept apart
- AI errors found and corrected
- Invoice and bank data kept out of personal accounts
Pass at 80. Under 80, read the feedback and try again. Then write your reflection and copy the entry for your CPD record.
Price
- Raising and approving spend: procure-to-pay for requesters, £39. The role track and its labs, this one included. No end date.
- Spotting fraud and deepfakes: the staff version, £39. The role track and its labs, this one included. No end date.
- Everything for a year, £199. Every course, role track and lab for 365 days.
- Monthly, £19.99 a month, cancel any time. Every lesson, plus 8 marked tasks a month, across every course, role track and lab.
Buying for a team, or want a lab built on your own process? CPD for employers.