Short answer: if sales promised a first service, a loaner, or a "we'll take care of that noise," service has to see it before the customer is at the advisor's window. AI can assemble that handover from the deal, the CRM note, and the RO. It should not invent a promise, and it should not text the customer a version sales would not sign.
Fixed ops is the lane. Parts holds are the wait. This is the handoff between two departments that already think the other one dropped it.
Where the promise goes to die
A salesperson writes "first oil change on us" in a sticky note, a text, or a CRM field nobody in service reads. F&I adds a product. The customer arrives at 7:15 a.m. six months later and quotes a sentence the advisor has never seen. The advisor is not wrong. The store is.
BDC bots make this worse when they "confirm the appointment" with a script that does not include the exception. The customer hears a second promise. Service hears a normal first visit. You pay for that gap in CSI and in the next trade.
What a handover has to carry
- What was sold, including products that change what service should expect.
- Anything promised in person or by text that is not in the deal jacket.
- Delivery issues already noted: noise, dash light, missing piece.
- Who the customer should ask for, if that still matters.
What AI may do
- Pull deal, CRM notes, and any open service recommendation onto one page for the advisor.
- Flag a first-visit customer whose deal includes prepaid maintenance or a "we'll handle it" note.
- Draft the internal brief. Not the customer text.
What it must never do
- Create a we-owe that is not in the record. If sales did not write it down, the tool must say "nothing in the file," not guess.
- Tell the customer service already knows. That is a promise that the advisor will be ambushed if the brief was wrong.
- Let a BDC agent close the loop unsupervised. Confirming time and place is one job. Repeating a we-owe is another.
The questions to ask
- Can service see sales notes without a scavenger hunt?
- Does anything go to the customer without an advisor?
- Where does a we-owe live after the tool is off?
- Who is accountable when the brief and the customer disagree? A person. Design it that way.
CDK, Reynolds, and the CRM are how stores already keep this. Naming them is not a connector claim. Customer data in those systems is still Safeguards territory.
A realistic first step
Sit with one advisor for a morning of first-visit customers. For each one, ask what they wish they had known from sales. Then see whether that fact exists anywhere in your systems. The ones that exist and did not travel are the handover problem. The ones that never got written down are a sales-process problem. Do not buy software for the second.
If you want the first problem assembled for the lane, with advisors still on every customer sentence — talk with us. If a better we-owe sheet is enough, we will say so.
Related: AI for dealership fixed operations · AI for dealership service: track the part, do not promise the car. · our approach for dealerships.