Day nine
One payment missed. Not a default, not arrears in any meaningful sense, no letter yet. A customer who has paid on time for three years and didn't this month.
The dialler connects and the agent says: "Hi, I'm calling about your account — there's a payment of two hundred and forty outstanding and I'd like to take that today."
Fifteen seconds in, and the whole thing has been decided. Either she pays, or she says she'll pay Friday to end the conversation, or she doesn't answer the phone again — and the third outcome, which is the most expensive one available, is produced almost entirely by the first fifteen seconds.
The money in this call is small. What's actually at stake is whether this person picks up in November.
The first call is an information call
Almost all collections training begins later — at genuine arrears, where the conversation is a negotiation and the customer has already decided how they feel about you. The first missed payment call is a different object and it's usually handled by whatever script the dialler is attached to.
Three features define it.
Nothing serious has happened yet. One payment. That's a data point with several completely different explanations, and treating it as the start of a collections process forecloses all of them.
You don't know what kind of problem this is. A card that expired, a pay date that moved, a job that ended, a relationship that ended. These require entirely different responses and the call's only real job is to find out which one it is.
And the second call is the one that matters. Whether this person answers next time is set here. An operation with a good first-call approach and a mediocre everything-else outperforms the reverse, because engagement is the constraint on every subsequent option.
What to say
1. Say who you are and why, immediately and plainly. No "courtesy call," no vagueness about the purpose. Ambiguity at the opening reads as a trap the moment the real subject arrives, and it costs you the rest of the call.
2. Ask what happened before asking when they'll pay. "I wanted to check in — the payment didn't go through this month and that's not like this account. Is everything alright?" It's a real question, it takes four seconds, and it produces the information the entire case depends on.
3. Then be quiet. The most common failure is an agent who asks and immediately fills the pause with the balance and a request. The answer arrives in the silence or not at all.
4. Establish blip or change. A card that expired is administrative. Hours cut at work is structural. Everything downstream — what's offered, who handles it, how fast it escalates — depends on that distinction, and it is available in the first ninety seconds if anyone asks.
5. Don't take a promise they can't keep. "I'll do it Friday" said quickly, to end a call, is worth nothing and costs something: a broken promise makes the next conversation harder for both of you. "Is Friday realistic, or would the end of the month be more likely?" is a better question than most agents are encouraged to ask.
6. Say what happens next, accurately. What the actual consequences are, in what order, on what timescale — within whatever the firm permits and requires you to say. No implied threats, no vagueness that the customer will fill in with something worse.
7. Write down what they told you. Not "customer will pay Friday." What they said about the reason. If the next call starts from zero, the customer learns that explaining anything here is wasted effort, and stops.
The thing the metric gets wrong
Most first-contact operations are measured on promises secured. It is the wrong number and it produces the wrong call.
An agent measured on promises will take one from anybody, including the customer who cannot make it, because the promise closes the call and scores. The operation records a good day and inherits a broken promise, a worse second conversation and a customer who has learned that this line is somewhere you get pressed for a date.
The number that matters is promises kept, and the difference between the two is a direct measure of how honest the first call was. Most lenders hold both figures and report the first.
Four ways it goes wrong
The demander, who opens with the balance. Efficient, and it converts a routine call into an adversarial one before anyone has spoken.
The vague greeter, whose ambiguous opener collapses the moment the purpose becomes clear.
The promise-taker, who accepts any date to close the call.
The no-noter, who resolves the call and records nothing useful, so contact two begins from nothing and the customer repeats themselves.
Why this isn't trained well
Collections training starts too late. The curriculum is built around negotiation, arrears stages and settlement — all of it relevant at month four and none of it relevant on day nine.
The first call is treated as a dialler output. It's automated, scripted and volume-driven, so it's designed for efficiency rather than for the one thing it's actually good for, which is finding out what kind of problem this is.
The measurement rewards the wrong behaviour. Promises secured is easy to count and moves in the wrong direction.
And peer role play produces an explainer. A colleague playing a customer who missed a payment gives a clear, complete reason in the first exchange. Real ones say "sorry, I'll sort it" and hang up as quickly as they can — and getting past that, in about twenty seconds, is the entire skill.
What early arrears training can rehearse
A simulation can hold a customer who wants the call over, gives a minimal answer, and only discloses the real situation to an agent who asks well and waits — and it can score the classification the agent reached rather than the promise they got. Foretell AI supplies the counterparty configuration, transcripts and rubric-based scoring; contact rules, permitted wording, forbearance options, vulnerability policy and all regulatory requirements stay with the lender.
Four to build:
- The administrative one, a genuine card or date problem, where the call should be thirty seconds and warm.
- The minimiser, who says it’s nothing and is at the start of something.
- The one who’ll promise anything, testing whether the agent tests the date.
- The one with a real change of circumstances, who mentions it once, in passing, late in the call.
Design caution. Contact frequency and timing, what may be said about consequences, forbearance options and treatment of customers in difficulty are governed by the lender's policy and by rules that vary by jurisdiction. Nothing here states any requirement; modules must use the firm's own permitted wording and vulnerability policy, and the exercise rehearses conversation quality only.
Designing the module
Pass one — the opening. Score whether the agent identified the purpose plainly and whether the first substantive question was about cause or about money.
Pass two — the silence. Score the pause after the question, and whether the agent filled it.
Pass three — the promise. Score whether a date was tested for realism, and whether the classification — administrative or structural — was correct.
Rubric on observable behavior: Was the purpose stated in the first sentence? Was cause asked before amount? Length of the pause after the question. Was the date tested? Was the situation classified correctly? Was the customer's own explanation recorded?
Cause-before-amount is the measure to build on. It's binary, it's visible in the first twenty seconds of any recording, and in most operations it's close to zero because nothing in the script asks for it.
The operator case
Promises kept versus promises secured is the whole argument in one comparison. You already hold both numbers. The gap is a measure of how many dates were taken rather than agreed, and it's usually large enough to end the discussion about whether the first call matters.
Second-call contact rate is the real leading indicator. Whether people answer the phone the next time is more predictive of recovery than anything achieved on day nine, and almost nobody reports it by agent.
Early classification routes the case correctly. Distinguishing an administrative miss from a change in circumstances in the first call is what allows the right treatment early — and it costs one question.
And the script is the intervention. This is not primarily an individual skill gap; it's a designed opening that asks for money before it asks what happened. Changing the first two sentences changes the behaviour of the whole floor.
For lending and customer operations programmes, this is a clean illustration that the earliest conversation in a process is frequently the highest-leverage one and the least designed — attention and investment tend to arrive later, where the options are worse.
Frequently asked questions
What should you say on a first missed payment call? Identify yourself and the purpose plainly, ask what happened before asking for money, wait for the answer, and establish whether this is a one-off or a change in circumstances.
Should you take a payment promise on the first call? Only one the customer can realistically keep. Testing the date — "is Friday realistic, or is the end of the month more likely?" — produces fewer promises and more payments.
Why do customers stop answering collections calls? Usually because of how the early contacts felt. A first call that opened with a demand teaches people that answering is unpleasant and achieves nothing, and disengagement is the most expensive outcome available.
What should early arrears contact be measured on? Promises kept rather than promises secured, and second-call contact rate. Counting promises alone rewards taking dates that were never going to hold.
The short version
Two hundred and forty pounds is not what this call is about. It's about whether she picks up in November.
Say who you are and why. Ask what happened. Then stop talking long enough for an answer to arrive. Work out whether this is a card that expired or a job that ended, because everything after this depends on which. Don't take a date she can't make.
And write down what she actually said, so that the next person doesn't make her say it again.
Foretell AI lets lenders build conversational simulations — including early arrears contact, minimal-disclosure customers and first-call classification like the one above — with configurable counterparties, transcripts, recordings, and rubric-based evaluation. If you report promises secured and hold promises kept, we're happy to walk through how other lenders have structured it.