Someone Else Answered: Third-Party Disclosure and the Location-Information Exception

The riskiest three seconds on a collections floor happen before anyone has said anything about a debt. A number is dialled, somebody picks up, and the collector has to decide — in real conversation, with no time to look anything up — whether the voice on the line is the person they are allowed to talk to.

Get it wrong in the direction of caution and you lose a contact. Get it wrong in the other direction and you have disclosed a debt to a stranger, which is one of the few things in this business that is both a statutory problem and the kind of thing consumers tell people about.

This is an orientation to what the rule actually says about that moment, and to what a call record has to be able to show afterwards. It is not advice about your own obligations — that belongs to your counsel, and the last section of this piece is explicit about where the line sits.

"Consumer" is wider than the debtor, and that surprises people

Regulation F defines the term for the purposes of its communications rules, and the definition is broader than most floors assume. Under § 1006.6(a), for that section the term consumer includes:

• the consumer's spouse;
• the consumer's parent, if the consumer is a minor;
• the consumer's legal guardian;
• the executor or administrator of the consumer's estate, if the consumer is deceased; and
• a confirmed successor in interest, as that term is defined in Regulation X and Regulation Z.

That list has an immediate operational consequence. A spouse who answers is not a third party for the purposes of § 1006.6. Neither is the executor of a deceased consumer's estate. A floor whose training says "only ever speak to the named debtor" is being more restrictive than the rule, which is a defensible choice — but a floor that treats every non-debtor voice as a hard stop should at least know it is making a business decision rather than following a requirement.

The mirror image also matters: an adult child, a sibling, a housemate, a receptionist and a colleague are all outside the list. So is a parent, unless the consumer is a minor.

The third-party prohibition, and the six people it does not cover

§ 1006.6(d)(1) states the prohibition plainly: in connection with the collection of any debt, a debt collector must not communicate with any person other than the consumer, the consumer's attorney, a consumer reporting agency (if otherwise permitted by law), the creditor, the creditor's attorney, or the debt collector's attorney.

Note the phrase doing the work. The prohibition is on communicating in connection with the collection of any debt — which is why the practical question on a live call is not only who is this but how much have I already revealed. A collector who says "I'm calling from [agency] about an important account matter" to an unidentified voice has arguably already communicated in connection with the collection of a debt, depending on what the agency's name conveys.

The rule then lists four exceptions in § 1006.6(d)(2): acquiring location information as provided in § 1006.10; the prior consent of the consumer given directly to the debt collector; the express permission of a court of competent jurisdiction; and communication as reasonably necessary to effectuate a postjudgment judicial remedy.

Only one of those four is a routine, everyday exception on a dialling floor, and it is the first.

The location-information exception is a script, not a licence

This is the part most worth reading in the original, because § 1006.10 is unusually specific about what a collector may and may not say. It defines location information narrowly — the consumer's place of abode and the telephone number at that place, or the consumer's place of employment — and then sets out the form of the communication.

A collector communicating with someone other than the consumer to acquire location information must:

1. Identify himself or herself individually by name. Not the agency. The person.
2. State that he or she is confirming or correcting the consumer's location information.
3. Identify the employer only if expressly requested. Volunteering the agency name is outside the exception.
4. Not state that the consumer owes any debt. No hedge, no implication, no "it's about an overdue balance".
5. Not communicate by postcard, and use no language or symbol on an envelope or in mailed contents indicating that the sender is in the debt collection business or that the communication relates to collecting a debt.
6. Stop contacting anyone but the attorney once the collector knows the consumer is represented with regard to the debt and knows or can readily ascertain the attorney's name and address, unless that attorney fails to respond within a reasonable period.

And then there is the frequency limit in § 1006.10(c), which floors forget more often than any of the six above: a collector acquiring location information must not communicate more than once with that person, unless that person requests it, or unless the collector reasonably believes the earlier response was erroneous or incomplete and that the person now has correct or complete information.

Read as a QA problem, that is a striking list, because every single element of it is audible. Items one through four either happened in the first fifteen seconds of the call or they did not. There is no interpretation required, no state of mind to infer, no context to reconstruct. It is the most mechanically checkable thing on a collections call — and on a floor reviewing two percent of its calls, nobody is checking it.

A limited-content message is a voicemail construct, not a live-call script

A related definition is worth keeping straight because the two get conflated in training decks. The limited-content message in § 1006.2(j) is defined as a voicemail message for a consumer. Its required content is a closed set: a business name for the debt collector that does not indicate the collector is in the debt collection business; a request that the consumer reply; the name or names of one or more natural persons the consumer can contact to reply; and a telephone number or numbers to use. Its optional content is also a closed set — a salutation, the date and time of the message, suggested dates and times to reply, and a statement that if the consumer replies they may speak to any of the company's representatives or associates. And it includes no other content.

Two things follow for a QA programme. First, that construct does not describe what a collector says to a live human who answers unexpectedly; the location-information rules do. Second, because the permitted content is a closed list, a voicemail intended as a limited-content message either matches the list or it is something else — which makes it, like the location script, unusually easy to score from a transcript.

What the call record has to be able to show

Here is the QA checklist the two sections above imply. Each row is a thing a reviewer can confirm from a transcript and a timestamp, which is exactly why it belongs in the binary half of a scorecard rather than the graded half — the scorecard design piece sets out that split in full.

What to check	Where it happens	What good looks like in the record
Who the collector believed they had	First fifteen seconds	An explicit confirmation, quoted, before any account detail
Whether the debt was mentioned pre-confirmation	Before right-party confirmation	No reference to a balance, account or collection purpose
Individual name given	Opening line of a location call	The collector's own name, spoken, not the agency's
The confirming-or-correcting statement	Opening line of a location call	The stated purpose, in the collector's own words
Employer volunteered unrequested	Anywhere in the call	Agency name absent unless the third party asked first
Attorney representation surfaced	Anywhere in the call	The moment it was said, and what the collector did next
Second contact with the same third party	Across calls, not within one	A reason on the record, or no second contact at all

Six of those seven are answerable from one call. The seventh is not, and it is worth being blunt about why: a once-only frequency limit is a counting problem across attempts, and attempt counts live in the platform that placed the calls, not in a review record. That is the same structural boundary we described for the seven-in-seven call frequency presumption — the review record can prove what was said on each contact, and the dialler is the only system that can prove how many contacts there were.

The failure modes we actually see

Reading transcripts rather than policies changes your sense of where this goes wrong. Four patterns recur, and none of them is the dramatic one people expect.

The agency name in the opening. By far the most common. A collector opens with the agency's name out of ordinary professional habit, to a voice that has not been identified yet. Whether that conveys a debt depends on the name, which is precisely the kind of ambiguity you do not want sitting at the front of every call.

The helpful third party. Someone answers, says "she's not here, can I take a message, what's it about?" — and the collector, being a decent person answering a direct question, tells them. The pull of ordinary conversational cooperation is strong, and it points the wrong way.

Voicemail drift. A message that began as a tightly-scoped script acquires a sentence over months of use — a reference to an account, a deadline, an "important matter regarding your file". Nobody decided to change it. It drifted, and the closed content list means drift is the whole problem.

The workplace number nobody flagged. A number in the file turns out to be a switchboard, so the person who answers is a colleague. Worth noting a real limit of ours here: our record ties calls to collectors and times, not to the account in your platform, so we cannot tell you that a number was flagged as a workplace line. Your system of record knows that; the review record knows what was said to whoever answered.

Training changes behaviour; the record is how you know it did

The reason to review this category on every call rather than a sample is not that collectors are careless. It is that the correct behaviour here runs against two strong instincts — name your employer, and answer a direct question — and instinct reasserts itself quietly over months. That is a drift problem, and drift is invisible to a sample.

Complete screening of the language-anchored categories is what our Voice tier does: every call transcribed with segment timestamps, screened against a collections rubric, and each finding reported with the second it occurred and the line that triggered it. Findings are reported per dimension rather than as one blended score, deliberately — a warm, efficient, well-handled call with a third-party disclosure inside it should not average out to a pass. The features page covers what each capability does.

Three constraints to state plainly, because they bear on this category specifically. Review is next-morning by design, so nothing here helps a supervisor intervene while a call is happening. Audio comes off the collector's computer, so a desk phone that does not route through the machine produces nothing to review. And a flag is a pointer for a human, never a verdict — the person who can be held responsible for the decision is the one who makes it.

The third-party disclosures we caught were never the collector being cavalier. It was always somebody being polite. A voice asks you what the call is about and you answer, because that is what people do. You cannot train that out with a poster; you train it out by showing someone their own call.

— Compliance manager, agency collections, 11 years, name withheld by request

Where this stops

Nothing here is legal advice, and this article deliberately stops short of telling you what your obligations are or how a regulator or court would view any particular call. It summarises published text so you can read the text.

The regulation is at the Electronic Code of Federal Regulations, the Bureau publishes it in interactive form with its commentary at consumerfinance.gov, the underlying statute sits with the Federal Trade Commission, and the rulemaking history that produced the current text is on the Federal Register. State law adds a separate layer on recording the call in the first place, which our coverage page treats as categorisation rather than conclusion, and which is a question for your counsel rather than for us.

As of September 2026, if you want this category screened on every call rather than sampled, that is the Voice tier at $99 per monitored collector per month with supervisors free. The Essentials tier at $39 covers desktop activity and the daily report without voice capture, and pricing has the rest. If you would rather establish first whether your dialler and your handsets can produce reviewable audio at all, ask us — that answer is free and occasionally it is no.

Frequently asked questions

Who counts as a third party on a collections call?

Under § 1006.6, anyone who is not the consumer, the consumer's attorney, a consumer reporting agency where otherwise permitted by law, the creditor, the creditor's attorney, or the debt collector's attorney. The definition of consumer for that section is wider than the named debtor — it includes a spouse, a parent if the consumer is a minor, a legal guardian, the executor or administrator of a deceased consumer's estate, and a confirmed successor in interest. An adult child, a housemate or a colleague is outside that list.

Can a collector leave a message with whoever answers?

Not about the debt. The routine exception for contacting someone other than the consumer is acquiring location information under § 1006.10, and that exception comes with a specific form: the collector identifies himself or herself individually by name, states they are confirming or correcting the consumer's location information, identifies the employer only if expressly requested, and does not state that the consumer owes any debt.

What is a limited-content message and does it apply to a live call?

It is defined as a voicemail message for a consumer with a closed set of required content — a business name that does not indicate the sender is in the debt collection business, a request to reply, the name of one or more natural persons to contact, and a reply telephone number — plus a closed set of optional items and no other content. Because it is defined as a voicemail message, it does not describe what a collector says to a live person who answers unexpectedly.

How many times can a collector contact a third party for location information?

§ 1006.10(c) says not more than once with that person, unless the person requests further contact, or unless the collector reasonably believes the earlier response was erroneous or incomplete and that the person now has correct or complete information. Auditing that limit needs attempt counts from the system that placed the calls — a review record can show what was said on each contact but not how many contacts there were.

Can call review catch third-party disclosure automatically?

The language-anchored parts, yes, and unusually well: whether the collector gave their own name, whether the stated purpose was given, whether the agency name appeared unrequested, and whether a balance or account was mentioned before the right party was confirmed are all present or absent in a transcript at a specific second. Whether a particular disclosure crossed a legal line is a judgement for a human, and the flag exists to put that judgement in front of one.

Why does this category need every call rather than a sample?

Because the correct behaviour runs against two ordinary instincts — naming your employer and answering a direct question — so it decays quietly rather than failing loudly. A slow drift across a floor is exactly the pattern a small sample is worst at detecting, and it is usually months old by the time a complaint surfaces it.