The 40-Question Email: How to Spot AI-Written Messages

The pattern is remarkably consistent:

• Abnormal length. 800–1,500 words for what used to be a two-line inquiry.
• The FAQ format. Ten, twenty, thirty bulleted questions, neatly grouped under bold headings.
• Credential demands. Licenses, registrations, certifications, "your legal framework" — asked with the confidence of a regulator and the context of none.
• Process interrogation. "Explain in detail how you work," step by step, as if onboarding you to their compliance department.
• The nonsense question. One technical-sounding query that doesn't apply to anything. Our recent favourite: a debtor asking about our "blockchain escrow protocols." We collect invoices. There is no blockchain. There is no escrow. There was, until that email, no protocol.
• The dead giveaway. Unfilled placeholders — [Your Name], [Company], [Insert relevant detail] — or the model's own signature pasted in whole.

This behaviour shift is visible in usage statistics. Anthropic's Economic Index — real AI usage matched to workplace tasks across 121 countries — shows the machinery behind it (May 2026 snapshot):

Half of observed AI use is full delegation — the output sent as produced, which is exactly how [Your Name Here] survives into your inbox. And email is one of the top things being produced.

But the stats only explain the how. The why is more interesting.

For the entire history of business correspondence, one rule held: long, detailed letters signalled effort, and effort signalled seriousness. A prospect who asked thirty careful questions was a prospect doing homework. A debtor demanding your credentials was a debtor with a lawyer.

AI broke that rule in about eighteen months.

Today a thirty-question email signals one thing only: someone typed "write a thorough due diligence email to a debt collection agency" and pressed enter. The homework took ten seconds. Frequently, they haven't read their own questions — which is how a B2B collection agency gets asked about blockchain escrow.

Three genuinely new behaviours sit underneath this:

Simulated diligence. People use AI to look rigorous — to themselves as much as to you. Sending the questionnaire feels like doing due diligence. Reading the answers is optional and often skipped.

Interrogation as armour. For debtors especially, the question-flood is the newest stalling tactic: bury the creditor in "legitimate compliance concerns" and every week spent answering is a week not paying. It's the infinite reply loop's aggressive cousin.

Cost asymmetry as a weapon. The sender pays seconds; the responder pays hours. Whoever answers AI-generated volume with human-generated effort loses by design.

The sociological headline: written length has been decoupled from human effort — permanently. The new signal of a serious counterpart is the opposite: a short, specific, personally-written message. Brevity is the new credibility.

The same email species requires two different treatments, because the intent differs.

If it's a lead (prospective client)

Don't punish them — behind some questionnaire bombs is a genuine client who simply outsourced the typing. But don't feed the machine either.

1. Answer three questions, not thirty. Pick the three that actually matter for their case (typically: cost, process, jurisdiction) and answer those crisply.

2. Redirect to voice. "The fastest way to cover all of this properly is a 15-minute call — here's a booking link." A real prospect takes the call. A tyre-kicker's chatbot doesn't.

3. Skip the nonsense with grace. "Some of your points don't apply to invoice collection — happy to explain which and why on the call." One sentence. Zero hours lost.

If it's a debtor

Different animal. The debtor's questionnaire is a stalling instrument: every question you answer buys them a week and produces another round.

1. Classify it as a delay tactic, not a dispute. A genuine dispute names an invoice, a defect, an amount. A credentials interrogation names none of these. Log it accordingly.

2. Answer once, minimally, in writing. Confirm your legal standing in two sentences. Attach nothing. Then restate the debt, the amount, and the deadline — unchanged.

3. Never let questions move the clock. The escalation calendar was set the day the invoice fell overdue. Thirty bullets don't amend it. This is the same iron rule as with AI-written payment excuses: only money moves the date.

4. Save the specimen. A debtor who interrogates instead of disputing has, usefully, put in writing that they have no actual dispute. That email is evidence. File it with a small smile.

Three rules for the new era:

Short is the new serious. Two specific sentences from a human now outrank twelve hundred generated words. Write accordingly — and judge accordingly.

Match effort to effort. Human messages deserve human answers. Generated volume gets a structured, minimal, one-time response. Never an hour of your day.

Behaviour over prose — always. The same rule as with payment excuses: what counterparts do (call, pay, name an invoice, show up) is signal. What they send is, increasingly, just weather.

The questionnaire bomb is what due diligence looks like when it costs nothing to fake: maximum volume, minimum attention, and a placeholder where the human used to be.

You can't stop them coming. You can stop paying for them — with your hours, your patience, or your escalation calendar.

And if the sender is a debtor: those thirty questions just confirmed, in writing, that they don't have a dispute. They have a delay strategy. We know exactly what to do with those.

Contact Us — Free Review →