How Dealerships Show Up in ChatGPT, Claude and Google AI
2026 Sold Out 2027January 5, 6, and 7·JW Marriott Atlanta Downtown·Early Bird through Nov 20 Register Now →

Claude for Dealers / Learn / Showing Up in AI Search

Visibility · Google and LLMs

How Car Dealerships Show Up in ChatGPT, Claude, Perplexity, and Google AI Answers

A shopper asks an AI assistant where to get a Honda serviced on a Saturday near them. It names three stores. Whether yours is one of them is not luck, and it is not the same thing as ranking on Google.

AI assistant
You

Which Toyota dealers near Marietta have Saturday service hours and good reviews?

Claude

Three options with Saturday service:

  • Dealer A: service open Sat 7 to 5. Reviews mention fast oil changes. Source: dealer site, maps listing
  • Dealer B: Sat 8 to 4, shuttle offered. Source: dealer site, review site
  • Dealer C: Sat 8 to 2. Source: directory

Your store: open Saturdays, not mentioned. Hours exist only inside an image on your site.

Illustration. AI assistants answer from what they can read and verify. A store whose hours live in an image, behind a script, or behind a blocked crawler does not make the list.
The short answer

AI assistants recommend dealerships they can read, verify, and cite. To show up, a store needs seven things: AI crawlers allowed in robots.txt and not blocked by the firewall, an llms.txt file, complete AutoDealer structured data, pages that answer real shopper questions in plain text, the same name, address, phone, and hours everywhere, recent reviews and third-party mentions, and a monthly prompt test to see what the assistants actually say. Most of this is work your website vendor controls, so the first step is asking them.

Key facts
  • Ranking on Google and being named by an AI assistant are related, not identical
  • Blocked AI crawlers are the most common problem and the fastest fix
  • Hours, departments, and services must exist as text, not images
  • Assistants lean on third-party sources: maps, reviews, directories, OEM locators
  • You can test your own visibility today with eight prompts
  • Much of the fix list is controlled by your website platform

Two scores, not one

When we audit a store we separate two questions. They have different fixes and different owners.

Visibility: do assistants mention you?
  • Are you named when shoppers ask for a dealer, a service, or a model in your market
  • Driven by third-party sources, reviews, content, and entity consistency
  • Measured by running real prompts
  • Owner: marketing
Agent readiness: can assistants read and use your site?
  • Can AI crawlers get in
  • Is the content text, structured, and current
  • Can an agent find hours, inventory, and how to book
  • Owner: website vendor and IT
The two-score model. A store can be well known and unreadable, or perfectly readable and unknown. You need both.

The 7 checks

1. Let the AI crawlers in

Open yourdealership.com/robots.txt. Look for these user agents. If any are disallowed, the matching assistant cannot read your site.

CrawlerWho it belongs toWhat it does
OAI-SearchBot, ChatGPT-User, GPTBotOpenAISearch results, live lookups, training
Claude-SearchBot, Claude-User, ClaudeBotAnthropicSearch results, live lookups, training
PerplexityBot, Perplexity-UserPerplexityIndex and live lookups
Google-ExtendedGoogleControls use in Gemini products
Applebot-ExtendedAppleControls use in Apple AI features

The trap: robots.txt can say "allowed" while your CDN or firewall blocks the same bots at the edge. Bot protection settings on the hosting platform often do this by default. Ask your website vendor, in writing, whether AI crawlers are blocked or challenged at the firewall.

2. Publish an llms.txt

An llms.txt file at your site root is a plain-text summary of who you are and which pages matter, written for AI systems. It is a young convention and not every assistant uses it, but it costs an hour. Include the store name, address, phone, hours by department, brands, services, and links to your key pages. You can see ours at learnclaudefordealers.com/llms.txt.

3. Complete your structured data

Your home page should carry AutoDealer schema with name, address, phone, geo coordinates, and opening hours. Service and parts should be marked up as departments with their own hours. Vehicle pages should carry Vehicle or Car markup with price, mileage, and VIN. Test any page with Google's Rich Results Test. Most platforms ship part of this. Few ship all of it.

4. Build pages that answer questions

Assistants quote pages that answer a question directly. Most dealer sites have none. Write short, factual pages or sections for the questions shoppers actually ask:

Put the answer in the first sentence. Use real text, not a graphic. The marketing guide covers producing this content in your voice.

5. Make your entity consistent

Your legal name, the name on the sign, the name on your maps listing, and the name on the OEM locator are often four different strings. Same for phone numbers and hours. Assistants cross-check. Conflicts lower confidence. Pick one version and fix the rest: business profile, OEM locator, major directories, review sites, social profiles, and your own footer.

6. Earn third-party citations

When an assistant names a dealer, it usually cites someone other than the dealer: a maps listing, a review site, a local news story, a chamber page, a forum thread. Recent, specific reviews matter. "Fast oil change on a Saturday" helps you show up for that exact question. Local coverage and community pages count. Your own site saying you are the best does not.

7. Test it monthly

Run the same prompts in ChatGPT, Claude, Perplexity, and Google's AI mode every month. Record whether you are named, what is said, and which sources are cited.

1. Best [brand] dealer near [city]
2. Where can I get a [brand] serviced on Saturday near [city]
3. Which dealers near [city] have a [model] in stock under [price]
4. Is [your dealership name] a good place to buy a car
5. What fees does [your dealership name] charge
6. [Your dealership] vs [competitor]: which is better for service
7. Who buys used cars for cash near [city]
8. What are [your dealership name]'s service hours
  1. 01Run8 prompts, 4 assistants, same day each month
  2. 02RecordNamed or not. What was said. Sources cited
  3. 03TraceWrong answer? Find the source it came from
  4. 04FixCorrect the source, add the missing page
  5. 05RepeatChanges show up over weeks
The monthly prompt test. Prompts 4, 5, and 8 are the ones that surface wrong hours, old names, and fees you did not know were being quoted.

Who owns each fix

FixDealerWebsite platform
robots.txt and firewall bot rulesRequest itControls it
llms.txtWrite the contentMust host the file
Schema markupAudit itControls it
Answer pagesWrite themProvides the page tool
Entity consistencyOwns itFooter and contact page only
Reviews and citationsOwns itNone
Monthly prompt testOwns itNone

Disclosure: Gray Scott co-founded PromptGraph, a platform that measures AI visibility for dealer groups. Everything on this page can be done by hand, and you should do it by hand at least once so you know what any tool is telling you.

Common questions

Is AI search optimization different from SEO?

It overlaps and extends it. Fast, crawlable, well-structured sites with good content help both. AI visibility adds three things SEO mostly ignored: AI crawler access, consistency across third-party sources that assistants cite, and content written as direct answers.

How do I know if my dealership website blocks AI crawlers?

Check yourdealership.com/robots.txt for disallow rules against GPTBot, OAI-SearchBot, ClaudeBot, Claude-SearchBot, PerplexityBot, and Google-Extended. Then ask your website vendor whether their firewall or bot protection blocks or challenges those crawlers, because edge blocking does not show up in robots.txt.

Does llms.txt actually work?

It is an emerging convention, not a guaranteed ranking factor, and support varies by assistant. It takes about an hour, it cannot hurt, and it forces you to write down the facts about your store in one clean place.

Should I block AI crawlers to protect my content?

For a dealership, no. Your content is hours, inventory, services, and reasons to visit. You want that repeated everywhere. Blocking makes sense for publishers who sell content. Dealers sell cars and service.

How long until changes show up in AI answers?

Live lookup answers can change within days once a page is readable. Answers drawn from a model's training or from third-party sources take weeks to months. That is why the test is monthly.

Can Claude help me do this work?

Yes. Claude can audit your robots.txt, draft your llms.txt, write answer pages in your voice, and generate schema for your developer. Our guide to building dealership agents includes an SEO analyst role.

Build Your Visibility Analyst Live

Attendees build an SEO and AI visibility agent and run it on their own store. Atlanta, January 5, 6, and 7, 2027.

Register Now →