Car shopping has quietly become an AI research task. A Cars.com survey of 936 shoppers, fielded November 4 to 10, 2025, found that 44 percent of consumers have already used an AI-powered tool to shop for a vehicle, and 97 percent of those AI users say the technology will influence their next purchase decision. For a brand selling or marketing cars, that means a growing share of research happens inside a chat window before a shopper ever lands on a dealer site. Here is what actually earns a mention there, and what does not.
Source: Cars.com's AI in Car Shopping Consumer Survey, fielded November 4 to 10, 2025 among 936 respondents, linked in the article.
How many car buyers are actually using AI tools now?
Close to half, and the share who trust what it tells them is catching up fast.
In the same Cars.com survey, 71 percent of AI users said they have at least a moderate amount of trust in AI tools for unbiased, accurate vehicle information, and 73 percent said AI turning a plain-language question into a targeted list of results saves them time. Two-thirds said they want an AI-powered personal shopping assistant for their next purchase. This is not a niche behavior confined to early adopters. It is becoming a default first step, in the same way a Google search was a default first step a decade ago, which is exactly the shift covered in how AI browsers are changing brand visibility.
What do buyers actually ask an AI assistant about a car?
Mostly comparison and research questions, not yet the final purchase.
The Cars.com data shows 59 percent of users treat an AI tool's answer as a starting point for further research rather than a final answer, while 30 percent say it already gives them a satisfactory answer on its own. That split matters for what a brand should optimize for. Early in the journey, an assistant is being asked to shortlist models, compare trims, or explain a feature. Later, it is being asked to recommend a specific dealer or verify a price. A brand that only shows up for the first kind of question, the broad category research, and never for the second, the specific comparison that leads to a transaction, is capturing attention without capturing the sale.
Does being named by an AI assistant actually drive traffic to a dealer?
Yes, when the assistant cites a specific source rather than giving a generic answer.
Cars.com's survey found that 41 percent of shoppers who used AI in their search said they would next visit the dealership or automaker website the AI tool specifically cited. That is the mechanism that matters most for this entire topic: being the named, cited source converts into a real visit for roughly two out of five of these shoppers. An assistant that answers from memory, with no citation at all, sends that visit nowhere. The practical goal is not to appear in an AI's training data in some vague sense. It is to be the specific page or source an assistant names and links when a shopper asks a comparison question, the same mechanic covered for other categories in how to get cited by ChatGPT and how to appear in Perplexity.
What are buyers worried about when AI gets involved in car shopping?
Bias, mostly, and it is a reasonable worry.
Sixty-three percent of respondents in the Cars.com survey said they are concerned AI tools might recommend cars in a biased manner, and two-thirds said car-shopping and review sites like Cars.com remain their most trusted source for objective information, ahead of the AI tool itself. That distrust of a single AI answer, paired with continued trust in independent review and marketplace sites, is the opening for earned media. A dealership group or automaker that has been independently reviewed, covered, or cited by a trusted third-party site gives an AI assistant a credible, attributable source to repeat. A brand with no independent coverage anywhere gives the assistant nothing trustworthy to cite except its own marketing copy, which the assistant has less reason to repeat as fact.
What should a dealership group or automaker actually do about this?
Give assistants something specific, sourced, and current to cite, the same discipline that works in any other AI-visibility category.
A few concrete places to start:
- Publish real, dated specifics an assistant can quote directly: incentive terms, inventory counts by model, delivery timelines, trade-in policies, and warranty terms, with a visible last-updated date. Vague marketing language gives a model nothing to extract.
- Earn independent coverage and third-party reviews rather than relying only on owned content. The trust gap in the Cars.com data runs in favor of independent sites, so a mention in one of them is worth more to an assistant's confidence than the same claim on your own homepage, a dynamic explained in more depth in whether PR actually helps SEO.
- Keep structured, current data about your actual inventory and pricing easy for a crawler to find and parse, since an assistant answering a specific-model question needs current facts, not a brand story.
- Track whether you are being named at all. If you cannot tell whether ChatGPT, Gemini, or a comparison tool mentions your dealership or model lineup when a shopper asks, you cannot fix what is missing, a gap covered in tools for tracking AI visibility.
Do marketplace and review sites still matter once an AI assistant is doing the summarizing?
More than ever, because they are often the source the assistant is summarizing.
Cars.com itself reported, citing its own Semrush data from September 2025, that it holds roughly double the AI citations of its closest competitor marketplace, a figure worth treating as a vendor's own claim about its own product rather than independent research, but directionally consistent with the rest of its survey: shoppers say they trust marketplace and review sites more than a single AI answer, and an assistant has to pull its answer from somewhere. That "somewhere" is increasingly a handful of marketplaces and review sites with deep, structured, frequently updated vehicle data. A dealership's own listing on those platforms, kept accurate and complete, is now doing double duty: it serves the human shopper browsing the marketplace directly, and it is raw material for the AI summary a different shopper sees in a chat window. Letting a listing go stale does not just cost a click on the marketplace. It can get repeated as fact by an assistant weeks later.
Is this different for a new model launch versus an existing dealership?
Yes, the timing and the ask are both different.
A new model launch is a news event, and it should be pitched and covered like one, following the same sequencing used for announcing any product launch: embargoed access for reviewers, real specifications at launch, and a publish date that lets outlets with road tests and comparisons get ahead of the assistant's next crawl. An existing dealership or marketplace is playing a different, slower game: consistent, current inventory data and a track record of independent reviews that accumulate over months, not a single news moment. Both strategies end at the same place, being the source an assistant trusts enough to name, but a launch needs a news hook and an existing operation needs sustained, verifiable detail.
The direct answer, for anyone who read this far
Nearly half of car shoppers already use an AI tool somewhere in their research, and when that tool cites a specific dealer or automaker by name, roughly two in five of those shoppers go visit that exact source next. Winning that mention is not about tricking a model. It is about giving it current, specific, independently verified facts to repeat, because buyers already say they trust independent review sites over a single AI answer on its own. If you want help building the earned coverage and current, citable content that makes an assistant name your brand instead of a competitor's, request a quotation.
