As of 13 August 2026, AI can chase a late-paying business customer.
This still needs a person who signs their name to it.
Can you do it?
5 minutesto a draft.
30 minutesto something you’d act on.
Cost, all in£0
Skill neededchat-fluent
Who has to check ityou
What the alternative costsThe supplied commercial data gives no price for a solicitor or debt-recovery service.
If this goes wrong, you can damage the customer relationship, make an unsupported legal threat or weaken your position in a later dispute.
What to actually do
Hand it to a person
The route this page recommends
A person who owns the outcome does this end to end, worth it when the failure is dear.
Use a tool built for this
Second choiceDo it yourself
The distant thirdA chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.
How to actually do it
- Open the invoice, customer account and signed contract or purchase order, and gather the invoice number, amount, issue date, due date, payment record and exact payment terms.
- Paste the relevant facts and correspondence into the prompt, removing unnecessary personal data and keeping the contract wording exact.
- Ask the chatbot to produce the reminder sequence, call script and checking table without sending any message.
- Compare every amount, date, name and contractual reference in the drafts against the invoice, ledger and contract before editing the wording.
- Remove any legal threat, interest claim, penalty or escalation statement that is not clearly supported by the contract or advice from your solicitor.
- Send the first approved message through your normal business email or credit-control system, record the date and response, then use the later drafts only if the agreed interval passes.
Prompt
Act as a careful business communications assistant for a UK company. Draft a payment-chasing sequence for this overdue business invoice using only the facts I provide. Do not invent contract terms, dates, payment rights, interest, penalties or legal consequences. Do not threaten court action, debt collection, reporting, winding-up proceedings or any other escalation unless the supplied documents clearly support the wording, and flag anything that needs a solicitor instead. Prepare: 1. A short first reminder email. 2. A firmer follow-up email. 3. A final escalation email that remains factual and professional. 4. A short call script and voicemail version. 5. A table showing when to send each message, what fact it relies on and what I must check before sending. Keep the tone commercially firm but preserve the customer relationship. Ask the customer to confirm the payment date or explain any genuine dispute. Separate the undisputed invoice balance from any disputed amount. Include no personal data beyond the names and contact details I provide. Mark every uncertain point as CHECK BEFORE SENDING. Company: [company name] Customer: [customer business name] Customer contact: [name and role, if needed] Invoice number: [invoice number] Invoice date: [date] Due date: [date] Amount and currency: [amount] Payments received: [details] Relevant contract or purchase order terms: [paste exact relevant wording] Previous communications: [paste relevant messages] Known reason for non-payment or dispute: [details or 'unknown'] Preferred outcome: [payment date, payment plan, dispute response or other] Messages already sent and their dates: [details] Do not send or publish anything. Return drafts and checks only.
Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.
What it gets wrong
- AI cannot know whether the customer has a genuine contractual dispute unless you provide the relevant evidence.
- AI cannot decide whether a threat of court action, statutory demand or debt collection is legally appropriate.
- AI cannot take responsibility for a message that breaches the contract, privacy rules or debt-recovery requirements.
- AI cannot replace a solicitor when the debt is disputed, commercially important or heading towards formal proceedings.
Even on a YES, the friction has a name: legal accountability, judgement under ambiguity and stakes of error.
How we scored this
Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.
| Axis | Score (0–2) |
|---|---|
| Output | 2 |
| Inputs | 2 |
| Verification | 1 |
| Liability | 1 |
| Effort delta | 2 |
| Total | 8 / 10 |
The methodology and its thresholds are published in full.
FAQ
- Can AI write a payment reminder email?
- Yes. It can turn your invoice, payment history and contract terms into a clear sequence of reminder emails, provided you check every fact before sending. Do not let it invent interest, penalties or legal consequences.
- Can AI legally chase an unpaid invoice?
- AI can help draft the message, but it cannot carry the legal responsibility for sending it. This is not professional advice, and a solicitor should handle a serious dispute or formal escalation.
- What information does AI need to chase a late-paying customer?
- Give it the invoice number, amount, issue date, due date, payment record, relevant contract terms, earlier messages and any known reason for non-payment. Remove unnecessary personal data and paste important contractual wording exactly.
- Should I threaten court action in a payment reminder?
- Only if you have checked that the wording and proposed action are appropriate for the contract, dispute and current circumstances. Ask a solicitor before making a formal legal threat, especially where the customer disputes the debt.
Nearby answers
- Can AI help me appeal a private parking charge for my business?YES
- Can AI help me appeal an HMRC decision for my business?PARTLY
- Can AI help my business challenge a business rates bill?PARTLY
- Can AI tell me if my dispute belongs in the small claims court?PARTLY
- Can AI help my business defend a breach of contract claim?NO
- Can AI help my business defend a wrongful dismissal claim?NO
Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.
The newsletter
AI news, new answers and product picks, straight to your inbox.