Collection Call Scripts: the Rule Behind Every Line

A collection call script is the structured language a collector uses on a call: how to identify yourself, what has to be disclosed, how to verify you are speaking to the right person, and how to move toward a payment arrangement without saying something the law prohibits.

The tension in writing one is real. The disclosures are non-negotiable and they make the opening sound formal, while the recovery depends on the call not sounding like a threat.

Below are six script templates covering the calls that actually happen, the rules each line exists to satisfy, and what changes when an AI voice agent delivers them.

These are templates, not legal advice. Your compliance counsel should approve the final language, because requirements vary by state, by account type, and by client contract.

TL;DR

• A third-party collection call must identify the collector, disclose that it is an attempt to collect a debt, and stay inside the permitted hours.
• Regulation F presumes a violation above seven calls about a debt in seven consecutive days, or a call within seven days of a phone conversation about that debt, counted per debt.
• Voicemails should follow the limited-content message format, which deliberately does not mention a debt.
• Never reveal the debt to anyone other than the consumer. The wrong person answering is the highest-risk moment on the call.
• A dispute or a request to stop contact changes the account state immediately, however casually the consumer phrases it.
• First-party scripts carry fewer mandatory disclosures and should sound like service, because that is what they are.

What every collection call script needs

Eight elements, in this order, cover what a third-party collection call has to do. They are a starting structure, not a guarantee. What your counsel approves is what governs.

1. Right-party verification before anything about the account is said.
2. Identification of the collector and, on a third-party call, the required debt collector disclosure.
3. A plain statement of why you are calling, once verification is complete.
4. A pause. The consumer's response determines everything after this point, and scripts that talk through it lose the call.
5. The resolution path: pay in full, arrange a plan, or identify what is blocking payment.
6. Handling for the three off-script events: dispute, hardship, and a request to stop calling.
7. Confirmation of whatever was agreed, in specific terms and amounts.
8. Documentation of what was said, not just the outcome code.

Scripts fail most often at step four and step six. Everything else is compliance drafting, and those two are where recovery is won or lost.

The rules the script has to respect

Six constraints shape the language of a third-party call.

• Identification and purpose. The collector must identify itself and disclose that the communication is from a debt collector attempting to collect a debt, with the initial communication carrying the fuller disclosure.
• The validation notice. A validation notice with the required debt information goes out in or within five days of the initial communication, and a written dispute within the validation period requires collection on the disputed amount to stop until it is verified.
• Call frequency. Regulation F presumes a violation where a collector places more than seven calls about a particular debt within seven consecutive days, or calls within seven days of having had a telephone conversation about that debt. Both presumptions are rebuttable and both count per debt rather than per consumer.
• Hours and place. No calls before 8 a.m. or after 9 p.m. in the consumer's local time, no calls at a time or place the collector knows is inconvenient, and no workplace calls where the employer prohibits them.
• Third-party disclosure. The debt cannot be revealed to anyone but the consumer. When a collector calls someone else to find the consumer, 12 CFR 1006.10 requires the collector to identify themselves, say they are confirming or correcting location information about the consumer, and name their employer only if asked, and they must not say the consumer owes a debt. The collector generally contacts that person only once, unless the person asks for another contact or the collector has reason to believe the earlier answer was wrong or incomplete, and nothing sent to them, such as a postcard, or any symbol on it may indicate debt collection.
• Conduct. No threats, no misrepresentation of the debt, the consequences, or who you are, and no false urgency.

The CFPB's Debt Collection Rule FAQs are the primary source for the frequency presumptions and the communication rules, and they are more precise than any summary. State rules sit on top, and several states are stricter than the federal baseline.

Script 1: first outbound contact, third party

The goal is verification, disclosure, and a listening pause, in under 30 seconds.

• "Hello, may I speak with [consumer name]?"
• If anyone else answers: do not state the reason for the call. Ask only when the consumer can be reached, then end.
• Once the consumer is on the line: "[Consumer name], my name is [agent name] calling from [company name]. This is an attempt to collect a debt and any information obtained will be used for that purpose."
• "This call is regarding an account with [creditor name], with a balance of [amount]. I wanted to see how we can get this resolved for you."
• Then stop and let them speak.

Three notes on why it is written this way. Verification comes first because the disclosure itself reveals the existence of a debt.

The balance is stated as fact without characterizing it, since anything that implies a consequence you cannot impose is a misrepresentation.

And the close is an open question rather than a demand, because the most common blocker is affordability, not refusal, and you cannot solve for it until they say it.

If an AI voice delivers this script, two lines attach ahead of the collection disclosure. 47 CFR 64.1200(b)(1) requires an artificial-voice message to state the identity of the business responsible for the call at the beginning, and several states, California among them, require the call to say that the voice is AI generated. The same applies to the first-party script further down.

Script 2: voicemail, as a limited-content message

A voicemail should not mention a debt, and the limited-content format exists precisely so it does not.

• "Hello, [consumer first name]. This is [agent first name] with [business name that does not indicate the collection business]. Please call me back at [phone number]. You can reach me [suggested times to call back]."

That is the entire message. A limited-content message must include four things: a business name for the collector that does not indicate it is in the collection business, a request that the consumer reply, the name of at least one natural person the consumer can contact, and a telephone number to reply on. It may also include a salutation, the date and time of the message, and suggested times to reply, and it may include nothing else. The consumer's name is not one of the four required elements, which is why it appears above only inside the salutation.

It must not state that the call concerns a debt. The business name is required rather than optional, and it has to be one that does not itself indicate the caller is in the collection business, which is why collectors use a non-indicating trade name for this message.

Regulation F's voicemail business-name rule (12 CFR 1006.2(j)) and the TCPA's artificial-voice identification rule (47 CFR 64.1200(b)(1)) can pull in opposite directions for an agency whose name signals collections, so settle your naming approach with counsel before recording this message.

The reason for the format is the third-party disclosure rule. Anyone can hear a voicemail, so a message that names the debt has disclosed it to whoever plays it back.

Script 3: inbound call from a consumer

The consumer called you, so the job is verification and speed, not persuasion.

• "Thank you for calling [company name], this is [agent name]. How can I help?"
• "Before we discuss the account, can you confirm your full name and [second identifier]?"
• After verification: "This is a communication from a debt collector. I'm looking at an account with [creditor name] and a balance of [amount]."
• "What would you like to do about it?"

Inbound calls convert far better than outbound ones, because the consumer has already decided to engage. The way to lose one is to make them wait or to repeat a script they already heard on a letter.

Script 4: payment arrangement

Confirm the specifics out loud, then confirm them again in writing.

• "Based on what you've told me, would [amount] on [date] work, and then [amount] on the same date each month?"
• "To confirm: [amount] on [date], then [amount] monthly until the balance of [total] is paid. Is that right?"
• "How would you like to make that payment?"
• "You'll receive written confirmation of this arrangement at [email or address]. Is that still correct?"

Two things to avoid in this part of the call. Do not promise an outcome you cannot deliver, such as what will happen to the consumer's credit file once the balance is paid.

And do not push past a stated affordability limit. A plan the consumer cannot meet breaks in month two and produces a broken promise rather than a recovery.

Script 5: dispute, hardship, or a request to stop calling

All three end the collection conversation and start a different one.

Consumers rarely use the legal words. "That's not mine", "I already paid that" and "the amount is wrong" are disputes. "Stop calling me" is a cease request.

• Dispute: "Thank you for telling me. I'm noting the dispute and pausing collection on the disputed amount while we look into it. The formal route is a dispute in writing to [address]. If you send that within the validation period shown on your notice, we have to stop collecting on that amount until we send you verification."
• Cease request: "Understood. I'm recording your request. Can I confirm you want us to stop contacting you about this account?" Then end the call.
• Hardship: "I'm sorry, that sounds difficult. Let me see what options we have." Then route to whoever is authorized to discuss them.
• Attorney representation: "Can you give me your attorney's name and contact information?" Then end the call and stop direct contact.

Each of these has to change the account state immediately rather than at the end of a shift, because the next automated dial is what turns a handled objection into a violation.

Script 6: first-party, pre-charge-off

This is a service call, and it should sound like one.

• "Hi, is this [customer name]? This is [agent name] calling from [creditor name] about your account."
• "I'm calling because a payment of [amount] was due on [date] and we haven't received it. I wanted to check whether something went wrong on our end."
• "Is there anything about the charge you'd like me to explain?"
• "We can take the payment now, or set up a plan. Which works better?"

No debt collector disclosure appears here, because the creditor collecting its own debt generally is not one under federal law. That is a legal position rather than a style choice, and it should be confirmed for your specific arrangement.

The tone difference is the point. A first-party call is a customer conversation with a balance in it, and treating it as a collection call damages a relationship the creditor still wants.

Lines to cut from any script

Six patterns that show up in old scripts and should not survive a review.

• Implied legal consequences. Anything suggesting suit, garnishment or arrest that you are not authorized to pursue and do not intend to.
• False urgency. "This offer expires today" where it does not, or a deadline invented to force a decision.
• Credit file promises. Statements about what will happen to a credit report are frequently wrong and always risky.
• Shame framing. Any line implying the consumer is dishonest or irresponsible. It suppresses recovery as well as creating exposure.
• Talking over a dispute. Continuing the collection pitch after the consumer has questioned the debt.
• Guessing. Answering a question about the account when you do not have the detail in front of you. Say you will confirm and follow up.

The artificial-voice rules, which are not the FDCPA rules

An AI-generated voice is an artificial voice under the TCPA, and that raises a consent question before any FDCPA question is reached.

The FCC ruled on February 8, 2024 that calls using AI-generated voices are "artificial" within the meaning of the Telephone Consumer Protection Act (FCC 24-17). Four things follow for a collections calling program, and they apply to a creditor and an agency alike.

• Consent. An artificial-voice call to a wireless number needs the called party's prior express consent. The FDCPA analysis never reaches this, so it is a separate review, with the consent record kept at the account level.
• A numerical limit on residential lines. Artificial and prerecorded debt collection calls to a residential line run under the exemption at 47 CFR 64.1200(a)(3)(iii), which allows no more than three such calls in any consecutive 30-day period. Going past that needs prior express consent. On a landline this binds before the Regulation F frequency presumption does.
• An opt-out on every call. The same exemption requires an automated, interactive voice or key-press opt-out mechanism on each call, under 47 CFR 64.1200(b)(3).
• Identification. 47 CFR 64.1200(b)(1) and (b)(2) require an artificial-voice message to state the identity of the business responsible for the call at the beginning of the message, and to give a telephone number during or after it.

State law adds to this. California's AB 2905, effective January 1, 2025, requires a call that uses an AI-generated voice to say so.

None of this prohibits the technology. It sets requirements the configuration has to meet, and they sit next to the FDCPA rather than inside it. Retell's TCPA compliance playbook for voice AI outbound goes through the consent side in detail.

How scripts change when an AI makes the call

The script stops being guidance and becomes configuration, which is both the advantage and the risk.

A human collector improvises around a script and produces variance. A voice agent executes it, which means the disclosure lands identically on every call, the hours and frequency rules are enforced by the system rather than remembered, and every call produces a record of what was actually said.

Four things have to be built rather than written.

1. Right-party verification as a hard gate. Nothing about the account is said until identity is confirmed, and the wrong person answering must trigger the location-information path with no mention of a debt.
2. Dispute and cease recognition in ordinary language, not keywords. "I already paid this" has to land the same way a formal dispute does.
3. Frequency and hours enforced in the dialing logic, per debt, on a rolling window, with state rules layered on top.
4. A transfer path that carries context, so a consumer routed to a person does not restate everything.

On the platform side, the pieces that matter are the transfer behavior, the per-call record from post-call analysis, and scoring calls against your own criteria with AI quality assurance so compliance can find drift across thousands of calls instead of sampling a handful. The debt collection industry page covers how this sits alongside the collections platform you already run.

One boundary worth stating plainly. A voice agent executes the script your compliance team approves. It does not decide what the disclosure should say, it does not make a program compliant, and a change to a prompt that touches disclosure language is a compliance change regardless of how small it looks.

Frequently asked questions

What should a debt collector say at the beginning of a call?

Ask for the consumer by name, and say nothing about the account until identity is confirmed. Once verified, a third-party collector identifies itself and discloses that the communication is an attempt to collect a debt and that information obtained will be used for that purpose.

How many times can a collector call?

Regulation F presumes a violation where more than seven calls are placed about a particular debt within seven consecutive days, or where a call is placed within seven days of a telephone conversation about that debt. The presumptions are rebuttable and count per debt, and some states set stricter limits.

Can a debt collector leave a voicemail?

Yes, using the limited-content message format: a business name that does not indicate the collection business, a request to call back, the name of a person to contact, and a phone number. The message must not indicate that the call concerns a debt.

What do you say when someone disputes the debt?

Acknowledge it, note the dispute on the account, pause collection on the disputed amount, tell the consumer how to submit the dispute in writing, and do not continue the collection pitch. A written dispute within the validation period requires collection on that amount to stop until verification is sent.

Do first-party collection calls need the debt collector disclosure?

Generally not, because a creditor collecting its own debt is usually not a debt collector under the FDCPA. The determination is fact-specific, and state law and unfair practices rules still apply, so confirm the position for your arrangement before removing any language.

Can an AI voice agent deliver a collection call script?

It can, and it will deliver it more consistently than a person. The obligations do not change: the disclosures, hours, frequency limits, dispute handling, and third-party disclosure rules all still apply, and they have to be configured by whoever owns compliance.

Run the approved script on every call.

Retell is a Customer Experience AI Platform for Autonomous Customer Relations. It delivers the language your compliance team approved on every call and transfers to a collector the moment a call needs one. Unlike scripted bots and controlled demos that work on the happy path, Retell is built for repeatable production reliability. The hardest moments must work again and again, not just once. In Cekura's 414-call benchmark on a regulated Medicare workflow, Retell passed 22 of 23 scenarios on all three attempts, leading six platforms with 95.7% workflow accuracy and 95.7% strict end-to-end reliability, across scenarios covering interrupted disclosures, consent refusal, changing intent, corrected information, prohibited advice, sensitive data and failed-transfer recovery.

Prove it on your own calls before you sign anything. Run a pilot on your own call volume.

This article is for general information only and is not legal advice. Debt collection and calling rules vary by state and change over time, so have qualified counsel review any script, disclosure, or calling program before you use it.