Three weeks ago a customer came in for an oil change and declined $840 of recommended work: rear brakes, a serpentine belt, and a cabin filter. The recommendation is sitting in the DMS. Nobody has called her.
When someone finally does, the call usually opens with "I'm calling to follow up on the service we recommended," which is the sentence most likely to end it.
Why this conversation is hard
This is the largest pool of identified, quantified, unbooked revenue in most dealerships. The work has already been inspected, priced, and documented. The customer has already been told about it. Nothing needs to be sold from scratch.
And it goes uncalled, or it gets called badly, for a reason that is easy to miss. The call feels like chasing money. Whoever makes it knows the customer already said no once, so they open apologetically, ask a closed question, and accept the first refusal.
There is a second problem. The person making the call frequently does not understand the work well enough to explain why it mattered. Reading "rear brake pads at 3mm" off a screen is not an argument. A customer cannot act on it, and it sounds like an upsell because it functions like one.
Turnover makes this permanent rather than temporary. With dealership turnover in the mid-forties annually and higher on the phones, the person calling this customer today may have been hired after the recommendation was written.
How the agent is built
The persona is a customer who declined for a real reason, and the reason is configurable, which is what makes this scenario reusable. She declined because the timing was bad, or because she did not trust the recommendation, or because she planned to get it done cheaper elsewhere and has not yet.
Those three produce three completely different calls, and most training treats them as one.
Behavior starts cool rather than hostile. She is not annoyed to hear from the store. She is unconvinced, and she has had three weeks to become comfortable with her decision.
Proactive sharing is set low so she does not volunteer that the brakes have started making a noise. If the advisor asks a good open question, that information exists. If they ask "are you ready to schedule that service now," it never surfaces.
Objections and priorities carries the real sequence: "I got a quote for less," "it's still driving fine," "can it wait until my next oil change," and the one that requires actual product knowledge, "do I really need that or is it just something you recommend."
The workflow half records the DMS path: locating the prior repair order, pulling the declined line items with their measurements, checking whether pricing still holds, and booking into the correct slot.
What gets scored
Whether the call opened with a reason for the customer rather than a reason for the store. Whether the technical finding was translated into a consequence she could act on, so "3mm" became "you have a few thousand miles before this stops being a pad replacement and starts being a rotor replacement."
Whether the original decline reason was surfaced and addressed rather than talked past. Whether the price objection was answered on the work itself rather than by discounting reflexively. And whether a specific appointment was asked for, with a day and a time, rather than an invitation to call back.
The workflow side scores whether the declined items were actually retrieved before the call rather than improvised, and whether the outcome was dispositioned properly so the next attempt is not a cold start.
What the run shows
The good run is built on one thing: the advisor read the repair order before dialing. It shows up in the first fifteen seconds, and the customer can hear it.
The bad run that teaches the most is the polite one. The advisor is friendly, mentions the recommended service, accepts "I'll think about it," says "no problem, just give us a call," and logs it as contacted. Every part of that call sounds acceptable.
She was never asked why she declined. The noise she has started hearing never came up. The store will not call again, and she will go somewhere else when the brakes get bad enough.
That call is scored as a pass by any tool that only listens to tone. It is a scored failure here, because closing behaviors and discovery behaviors are graded explicitly and the disposition step is graded too.
What a manager does with it
A fixed operations director gets the diagnosis, not just the outcome. When the whole team's discovery scores are low on this scenario, the problem is the opening question and it can be fixed in one meeting. When translation scores are low, the problem is product knowledge, and that is an AI Tutor built from the store's own service menu rather than a coaching conversation.
The two failures look identical on a booking report. They need completely different responses.
This is also the scenario that most justifies training to a dealer principal, because the revenue is already identified and sitting in the system. The gap between what was recommended and what was booked is a number the store already tracks.
And because the same scorecard runs on real follow-up calls, the director can show that the behaviors rehearsed on Tuesday are the behaviors used on Thursday. Practice proves capability. The live call is the only thing that proves it changed.
Bring one week of declined-service line items and five recorded follow-up calls. We will score the calls and build the agent from your own declines.
Frequently Asked Questions
.svg)
It feels like chasing money. Whoever makes it knows the customer already said no once, so they open apologetically, ask a closed question and accept the first refusal. Turnover makes it permanent rather than temporary, since the person calling today may have been hired after the recommendation was written.
The reason is a setting on one agent. She declined because the timing was bad, because she did not trust the recommendation, or because she planned to get it done cheaper elsewhere, and each produces a different call from the same scenario.
Whether the declined line items were actually retrieved before the call rather than improvised, whether pricing was checked for whether it still holds, whether the booking went into the correct slot, and whether the outcome was dispositioned properly so the next attempt is not a cold start.
Not on its own, and the scorecard tells you which problem you have. Low discovery scores are an opening-question problem that a meeting can fix. Low translation scores are product knowledge, which is an AI Tutor built from your own service menu rather than a coaching conversation.
With a number the store already tracks. The gap between what was recommended and what was booked is identified, inspected and priced work sitting in the DMS, which makes this the easiest training case in the building to put in front of an owner.
Table of Contents
Talk to Sales
Have questions about training and enablement for your sales, CS, support, or leadership team? Let's talk.
Talk to Sales
