FCA authorised · FRN 724952 0117 325 0027Quote & buy →
Aipedia: Bookkeepers
AI credit control tools send payment reminders automatically on a client's behalf, so a wrong chaser can go out in the client's name before anyone checks. For bookkeepers offering credit control, that creates financial, reputational and data protection risks.
Part of: AI in Bookkeepers
In short
Some bookkeepers may run credit control for clients using software that sends reminder emails and texts, predicts late payers and, in some tools, writes the message with generative AI. Errors include chasing customers who have paid, quoting wrong amounts, sending one customer's details to another and an aggressive tone that damages the client's relationships. Where significant decisions about individuals are made solely by automated means, UK GDPR Articles 22A to 22D require safeguards. ICB s.14 requires absolute confidentiality. Some resulting losses are a PI matter; others, such as lost goodwill or data leaks, may not be, depending on the wording.
Credit control modules sit on top of the sales ledger. They schedule reminders by due date, escalate tone over time, predict which customers are likely to pay late and suggest actions such as stopping supply. Generative AI may draft each message so it reads as personal. For a small business, a bookkeeper running this service can bring in cash faster. The tool, though, only knows what the ledger tells it, and the ledger may be wrong (see AI bank reconciliation).
The ICB rules are in its Professional Conduct Regulations. No bookkeeping body has issued AI guidance on credit control.
Illustrative scenario: a bookkeeper runs automated credit control for a wholesale client. A customer's payment was posted to the wrong account, so the tool sends three escalating reminders and then an automatic "account on hold" notice. The customer, a long-standing buyer, has paid in full and moves its business elsewhere. In the same week, a generative AI draft includes a different customer's outstanding balance in the email. The client complains of lost sales and a data breach.
Errors in automated chasing are visible to the client's customers, so they cause harm quickly and are hard to undo.
Generative AI raises a separate accuracy problem. A model asked to write a firm but polite reminder may invent a late payment charge, a deadline or a consequence the client never agreed. If the customer is an individual, a threatening or inaccurate message can lead to a complaint against the client. The bookkeeper who set up and ran the process is then the obvious person for the client to blame, even if the client approved the original template.
It depends on the loss. A claim that the bookkeeper negligently managed the ledger or chasing process, causing financial loss, may fall within PI, subject to terms. Lost profits from a customer who left can be hard to prove and may be argued as too remote. A misdirected email that discloses personal data is often a cyber or data breach matter; see data breach liability for professional firms. Chasers that include bank details are a target for fraudsters impersonating the sender; see wire fraud and compromised email. Check that credit control is declared as a service on your proposal.
Apex arranges PI for bookkeepers who offer credit control, and a broker can review how PI and cyber cover fit together for this service. See the AI in bookkeeping hub and AI-written client emails.
Sending a reminder is unlikely to be a significant decision in itself. Automatically putting an individual customer on stop or refusing credit could be, and then the safeguards in UK GDPR Articles 22A to 22D apply. Business customers who are not individuals fall outside these rules.
The client is usually the data controller, but it may look to the bookkeeper who ran the process. Whether the loss is covered by PI, cyber or neither depends on the policies and the facts.
Yes. Set out what you will send, how often, who approves escalation, which customers are excluded and who is responsible for the accuracy of the ledger. Without that, disputes about a wrong chaser turn on memory rather than agreed terms.
You can, but keep the figures and account details out of AI drafting. Use templates for the facts and let AI vary only the wording, then sample check what goes out.
Tell us about your firm and the work you do. A broker will review how your current PI treats AI-assisted work, or arrange new cover. No fee for a quote. Or call 0117 325 0027.
Start a PI quote Request a callbackApex Insurance Brokers Limited is authorised and regulated by the Financial Conduct Authority. Registered in England and Wales, company number 07014570. This page is general information about how AI is used in the profession and how insurance may respond. It is not legal, regulatory or insurance advice on your circumstances, and it does not guarantee that cover will be available or on what terms.
