Collection check

How to spot a fake debt collection letter

A letter, email or text claiming you owe money to a collection agency or a bailiff is easy to send and hard to judge: the amount is usually small enough to just pay, and the deadline is short. In the Netherlands the national fraud desk reports recurring waves of false messages sent in the name of bailiffs and collection agencies, sometimes addressing the recipient by name (Fraudehelpdesk, false messages in the name of bailiffs). The pattern repeats: an unspecified outstanding payment, a demand to pay within two days or face a seizure, and either a payment link or no payment option at all.

A real collection claim is checkable: it names the original creditor, says where the claim comes from, and shows how the amount is built up. When that is missing, the message is pressure rather than a bill.

Direct answer: five checks on a collection notice

1. Do you recognise the debt? Look through your own records, statements and mailboxes for the creditor and the amount. A claim you cannot trace back to anything is an assertion, not an arrears position. Keep the message, the envelope and any sender address before deleting anything.
2. Demand the validation information. Under the US Fair Debt Collection Practices Act a collector must give you validation information either at first contact or within five days of it: their name and mailing address, the name of the creditor you owe, how much you owe including interest, fees, payments and credits, what to do if you do not think the debt is yours, and your rights (FTC, Debt Collection FAQs). A sender who will not name the creditor in writing is not validating a claim.
3. Read the pressure tactics and the timing. Legitimate collection has limits: contact not before 8 a.m. or after 9 p.m. unless you agree, no contact at work once you say you are not allowed to receive calls there, no more than seven calls within seven days, and no contact by email or text once you ask them to stop (FTC). Repeated same-day calls, threats of immediate seizure or arrest, and a payment link in an unexpected message point the other way.
4. Call back on a number you found yourself. Use the number on the collector's own website or on an earlier document, never the number printed in the suspect message. A number that rings through to somebody else — the national fraud desk in the Dutch cases — proves nothing about the sender (Fraudehelpdesk).
5. Do not pay through the link, and put your dispute in writing. Ask for a written breakdown with the original creditor, and reply only through a channel you have verified. If the debt is not yours, dispute it in writing; in the US you can also send a letter asking the collector to stop contacting you, keeping a copy and using certified mail with a return receipt so you have proof it arrived (FTC).

Why AI makes this cheaper to run

Fake collection messages are not new. What changed is the cost of production: a language model writes a flawless letter with the right tone, and a logo with matching branding takes minutes to copy. The mistake is therefore no longer visible in the language or the design, but in whether the debt and the sender exist. That is why the checks above target the claim itself rather than how convincing the message reads.

What a detector does and does not see

A collection notice is usually text, sometimes a scan or a screenshot of a letter. You can run the text through the text detector and a scan through the image detector. That is triage: it helps you decide which message to verify first, and it produces a signal you can keep in a report. It says nothing about authenticity. Templates carry little personal writing style, short letters are too short for a reliable signal, and a fully generated letter can score low. So a low score never proves that the debt exists.

What a collection notice cannot tell you

• whether the debt exists and whether you are the right addressee;
• whether the sender is a collection agency or bailiff, even when the logo and layout match;
• whether a named creditor ever instructed anyone to collect;
• whether a threat of seizure is legally correct — enforcement runs through a court and an appointed officer, not through a payment link;
• whether it is fraud; only an investigating authority can establish that.

Treat the message as a prompt for your own verification, never as evidence that you owe something.

If you already paid

• Contact your bank immediately if you paid through a payment request, a link or a transfer; speed decides the chance of a block or a recall.
• Keep the message, the envelope or sender address, the payment details, the link and the timestamp.
• Report it to the national fraud desk or, in the US, to the FTC and your state attorney general; the FTC Debt Collection FAQs list the reporting routes (FTC).
• Check whether details you supplied are being used elsewhere, and change any password you were asked for in those conversations.

More on this family of documents is in the guides on checking a suspect invoice, spotting a fake payment proof and spotting a fake delivery message. To check a text or a scan on its own, use the free checker.

Frequently asked questions

Is a collection letter with a payment link automatically fake?

Not automatically, but it is the combination that goes wrong most often: an unspecified claim, a two-day deadline and a payment link or payment request. Ask for a written breakdown naming the original creditor, and verify the sender through a channel you found yourself.

Does a high detector score prove the letter is fake?

No. The score describes patterns in a text or a file and says nothing about whether the debt exists or who sent the message. A low score does not prove the message is genuine either.

Should I reply if I do not recognise the debt?

Do not pay on the strength of the message alone. Ask for the validation information in writing, dispute the claim if it is not yours, and keep a copy of everything you send. If the claim is genuinely yours, paying the collector directly through a route you verified is still safer than following a link.