A shopper asks an AI assistant, “Which dealership near me will buy my truck without making me buy another vehicle?” Your store does that every day. But imagine the answer names two other rooftops and leaves yours out. There’s no missed call to review, no unanswered text, no overdue task for the BDC manager. You never got the opportunity to mishandle this customer.
That possibility deserves attention without a sweeping claim about how shoppers now choose dealerships. When someone asks an assistant where to buy, sell or service a vehicle, the answer can narrow their choices before the first measurable contact. How often that happens in your market—and whether it changes your results—requires evidence beyond an occasional screenshot.
Your CRM measures the survivors
We already accept that a CRM doesn’t capture everyone who considered the store. A customer can drive past, read reviews or abandon a website without identifying themselves. AI recommendations add another selection point: the shopper may receive a short list explaining which dealership fits the request. Your inventory page or service scheduler might never get a look.
Your lead report still matters; it just cannot explain every opportunity that never reached the store.
A closing ratio tells you what happened to recorded opportunities. It cannot tell you whether your store was considered for Saturday brake work, a used SUV under $25,000 or an outright vehicle purchase. If recorded leads soften while closing stays steady, recommendation visibility belongs on the diagnostic list. It does not automatically belong at the top: inventory mix, pricing, advertising changes and broken forms still deserve inspection.
The exposure differs by department. Sales might lose a retail prospect; service might miss someone who assumes franchise stores won’t work on an off-brand vehicle. The used-car department might miss an owner ready to sell outright. That last one matters twice: you lose an acquisition opportunity and may need to replace the unit elsewhere, with auction fees and transportation expense. None of that proves a lost deal, but it explains why the used-car manager should care.
AI dealership recommendations need to be accurate, too
I’d argue that accuracy deserves as much attention as visibility. An assistant naming your dealership for a service you don’t provide is not a marketing win. Neither is an answer that gets your service hours wrong or directs a seller to a trade-only form. You might earn the click and still lose the customer—or create an argument at the advisor’s counter.
A screenshot showing your name proves only that you appeared in that answer. Another shopper may receive different guidance, and the same question may produce a different result later. Treat a favorable mention as something to inspect, not a position your store now owns.
There’s an attribution trap here, too. If the shopper later searches your dealership name and calls, the recorded source may describe the final step rather than the original recommendation. Asking “What first put us on your list?” can add context, especially when staff record the customer’s own words. Memory still isn’t a clean tracking system. Don’t turn a plausible influence into a precise revenue claim.
Run a 24-answer audit, not a ranking contest
Start with four customer questions tied to business you actually want and can handle. Use two AI assistants and ask each question in three fresh conversations, keeping the stated location consistent. That produces 24 answers. Save the wording, date, responses and any linked pages so the next check is comparable. If your store doesn’t offer one of the services below, substitute a question that matches your business.
- Who will buy my vehicle locally without requiring a replacement purchase?
- Where can I get Saturday service for my vehicle’s make?
- Which dealerships nearby sell used SUVs under $25,000?
- Where can I schedule service for an off-brand vehicle?
Repeated fresh conversations do not make this small sample representative of local shoppers’ experiences, and pooling assistants and departments can hide which problem needs fixing.
Keep the results separated by assistant and customer question, even if you also prepare a storewide summary. For each answer, check three things: whether it names your store, accurately describes the relevant offering and gives the customer a working next step. Follow the link or contact instructions yourself. A page loading successfully isn’t enough if the seller must claim to be buying a replacement vehicle before submitting a request.
For a simple internal measure, divide answers meeting all three conditions by all answers tested. In a hypothetical audit, your store might appear in nine answers, with four containing a material error or unusable next step. That leaves five usable recommendations out of 24, roughly 21%. It is not market share, a reliable comparison between departments or a number you can multiply by sales volume to estimate lost deals.
Assign each problem to someone who can resolve the customer confusion. Your website team can clarify the outright-purchase policy; fixed ops can confirm hours and supported makes; marketing can reconcile public business listings. Give each correction an owner and a review date. If an answer promises Saturday brake work, verify that customers can actually request that work—not merely that the building is open. Clearer information won’t guarantee a recommendation, but it can prevent wasted calls and disappointed arrivals.
Before shifting budget, compare how often your store appeared with how often it was named accurately and offered a working next step. Fix those customer-facing gaps first, then repeat the check to see what still needs attention.