We Asked ChatGPT About 15 Volvo Dealers: How Do Car Dealers Show Up in ChatGPT Answers?

Adriana Dikih

SEO/GEO Researcher

SEO

/

Car dealers show up in ChatGPT answers mostly through their own websites and review sites. When a shopper asks about a dealership by name, ChatGPT names it in almost every answer and cites the dealer's own website in most of them. When the shopper asks what other people think of the store, ChatGPT turns to review sites like DealerRater, Birdeye, and CarEdge.

To find out, Honest Digital asked ChatGPT the same 62 shopper questions about 15 Volvo dealerships in 15 U.S. markets in August 2026, then asked them again in November 2026. We read every answer: 1,094 from each run. Between the two runs, ChatGPT cut its sources from about nine per answer to four, and dealer websites and review sites made the cut most often.

Overview

Here are the 8 takeaways from this study:

  • Your website is the source ChatGPT cites most. The dealer's own site was cited in 82% of answers, and it held up best when ChatGPT cut its links from nine per answer to four.

  • ChatGPT reads a handful of your pages. The homepage, About, specials, and service pages lead. Some pages that were never meant to be sources, like sitemaps and thank-you pages, get cited too.

  • Shoppers in nearby towns don't always find you. ChatGPT named the store in 61% of nearby-town answers and cited its website in 29%.

  • The question decides the source. Your site answers price, service, and inventory questions. Review sites answer trust and "which dealer is best" questions.

  • You're named almost every time, but cited less on reputation. On questions about what people say about you, review sites replaced the dealer's own site in November.

  • Price details come from inventory pages, specials, and CarEdge. Most dealers publish their doc fees and payments, but deep in vehicle pages.

  • ChatGPT tells shoppers what it can't find. About one answer in six says so, most often about service wait times, credit requirements, and trade-in values.

  • ChatGPT is saying more with fewer sources. November answers ran about a fifth longer while citing less than half as many links.

How We Measured It

We used our prompt tracking tool to ask ChatGPT 62 questions about 15 Volvo dealerships across the country. The questions cover the whole shopping journey: trust and reputation, price and financing, service, inventory, what makes the store different, competitors, and real shopper situations ("It's my first time buying from a dealership...").

This post compares only the questions asked in both runs: 1,094 answers in each.

For each answer, we recorded:

  • whether ChatGPT named the dealer;

  • which websites ChatGPT cited;

  • what it said it couldn't find.

Unless we say otherwise, the numbers below come from the August run's 62 shared questions (930 answers), and each finding ends with what changed in November.

What Changed Between August and November

Between the two runs, ChatGPT itself changed more than the dealers did. On the 1,094 answers that repeat an August question, it cited fewer sources, ended almost every answer with a question back to the shopper, and named the store just as often.

On the 1,094 repeated answers

August

November

Store named

91%

93%

Dealer's own site cited

77%

65%

Review site cited

66%

53%

Marketplace cited

56%

24%

volvocars.com cited

31%

12%

Other dealers cited

44%

15%

Links cited per answer

9.4

3.8

Ended with a question for the shopper

80%

93%

  • Fewer links. About four per answer instead of nine. ChatGPT dropped listings and rival dealer pages first, which is most of why marketplace and other-dealer shares fell.

  • Your site lost ground to review sites. The dealer's own site was cited less on every kind of question, most on brand trust (74% to 49%) and inventory (94% to 76%).

Finding #1: Your Website Is the Source ChatGPT Uses Most

The store's own site was cited in 82% of answers to the 62 shared questions, and dealer sites made up a third of every link ChatGPT gave. ChatGPT also named the store in 95% of answers, and four answers in five did both. When an answer cited a dealer's site, it linked to about three of its pages. Across all 62 questions, ChatGPT cited about 35 different pages per dealer. The homepage was cited most, followed by About, specials, and service pages (see Finding #2).

The other sources it leaned on:

Source

Share of all answers citing it

The dealer's own website

82%

Any review site (DealerRater, Carfax, Birdeye, Yelp, BBB, and others)

65%

Any marketplace (Cars.com, CarGurus, Kelley Blue Book, Capital One, and others)

52%

DealerRater

41%

Other dealerships (Volvo or other brands)

38%

Cars.com

33%

volvocars.com

33%

Reddit

3.5%

In November, with fewer links per answer, nearly every source showed up less. Dealers' own websites held up better than anything else (see What Changed, above). 

What this means for you: your website is already in the conversation. The work now is making sure the pages ChatGPT reads give it something specific to say: specific facts, clear policies, and proof.

Finding #2: The Pages ChatGPT Reads (and the Ones It Shouldn't)

These were the dealer pages ChatGPT cited most often:

Page type

Share of all answers

Homepage

47%

About or dealership info

32%

Specials and offers

29%

Service and parts

24%

New inventory

19%

Contact, hours, and directions

17%

Used inventory

15%

Certified pre-owned

12%

Value promise pages

10%

Reviews and testimonials

8%

A few cited pages were never meant to be sources. One store's sitemap page was cited in 30 answers, across trust, service, inventory, financing, and shopper-situation questions. Another store's account registration page was cited in 21 answers, on questions ranging from reputation to service to financing.

A third store's form thank-you page was cited in 11 answers, nearly all of them price questions. Two answers even linked to a test version of a dealer's website, one about ordering a special trim and one about booking a test drive.

What this means for you: your About page does more work than you think. Make it specific: when the store opened, who owns it, how many people work there, awards, and what you do differently. And ask your website provider to keep test sites and thank-you pages out of search.

Finding #3: Shoppers in Nearby Towns Don't Always Find You

We also asked the questions the way a shopper in a nearby town would, such as "I live in [town], whats the best Volvo dealer near me?" or "I want to lease a XC90 in [town], which dealer has the best offers?" Each store got three towns, between three and 28 miles from the store in a straight line.

  • ChatGPT named the store in just 61% of those answers.

  • It cited the store's own website in only 29%.

  • It cited marketplaces like Cars.com and CarGurus in 85%.

In November, on the same questions, ChatGPT named the store a little more often (68%) but cited its website less (21%).

What this means for you: your conquest market is where AI visibility drops off. If you want shoppers from the next town over, give ChatGPT a reason: pages that speak to those towns, delivery or pickup options, and the specific reasons people make the drive.

Finding #4: The Question Decides the Source

Type of question

Dealer site cited, Aug → Nov

Review sites cited, Aug → Nov

Inventory ("Do they have a good selection?")

94% → 76%

52% → 39%

Fixed Ops ("How is the service department?")

90% → 85%

53% → 41%

Price and financing

90% → 82%

47% → 28%

Brand trust and reputation

74% → 49%

87% → 81%

Competitor ("Which Volvo dealer is best?")

57% → 45%

96% → 82%

When a shopper asks about inventory, financing, or service, your website does most of the talking. When they ask whether you can be trusted, or which dealer is best, review sites take over. On those questions, DealerRater alone was cited in 81% of competitor answers. 

In November, trust answers leaned even harder on reviews. The dealer's own site showed up in 49% of brand trust answers, down from 74%. Which review sites matter also shifted. DealerRater fell from 42% of answers to 12%, while Birdeye (20% to 28%) and CarEdge (23% to 25%) became the two most-cited sources overall.

What this means for you: there are two jobs here, and they need different owners. Your website team owns the price, service, and inventory answers. Your reputation work (review volume, responses, and the profiles on DealerRater, Birdeye, CarEdge, Google, Carfax, and Cars.com) owns the trust answers.

So where did Reddit go?

Earlier this year, we ran a smaller study on one dealership: Reddit showed up in 86% of questions, and a single thread about that store in 72% of questions. Across these 15 stores, it showed up in 3.5% in August and 2% in November.

The first store had a popular Reddit thread written about it, and ChatGPT surfaced it again and again. Most dealers don't have one. When we asked ChatGPT "What does Reddit say about [store]?", every August answer said there was little or no Reddit discussion of that store. Five answers still linked to Reddit, but to general threads (Volvo reliability and a local subreddit), not threads about the store. In November, seven of 15 linked to Reddit, again to general threads.

Timing matters too. In August 2026, ChatGPT sharply cut how often it cites Reddit. Qwairy measured Reddit's share of ChatGPT citations falling 95% between August 1 and August 14. Almost all of our August answers were collected on August 5 and 6, just before the drop, so it doesn't explain this study's low Reddit numbers. It does mean Reddit likely matters even less in ChatGPT today.

What this means for you: Reddit matters a lot when a thread about your store exists. Search for your store name on Reddit. If a thread exists, treat it like a review profile you can't edit. If none exists, your review sites and your own pages are doing that job instead.

Finding #5: Named Almost Every Time, Cited Less on Reputation

When the shopper names your store, ChatGPT names it nearly every time. Two kinds of questions are softer. When the shopper asks which dealer to pick, the store is still named in most answers, just not all of them. When they ask what other people say about you, the store is named, but its own site is cited much less. Below are the exact questions we asked, with [store] standing in for each dealer's name and [city] for its city. We asked each question once for each of the 15 stores in each run, so every count is out of 15 answers, shown August → November.

"Which dealer" questions: named in most answers, not all

  • "Can you rank the Volvo dealers in [city] from your top recommendation to the worst?" (store named: 11 → 11 of 15; own site cited: 5 → 3 of 15)

  • "Which Volvo dealer is the best in [city]?" (store named: 12 → 11 of 15; own site cited: 9 → 8 of 15)

  • "Who is the top Volvo dealer in [city]?" (store named: 13 → 11 of 15; own site cited: 7 → 9 of 15)

  • "Which Volvo dealer in [city] has the best prices?" (store named: 13 → 11 of 15; own site cited: 12 → 9 of 15)

  • "Which Volvo dealer in [city] has the best deals?" (store named: 12 → 12 of 15; own site cited: 12 → 11 of 15)

Questions where the store is named, but its site isn't cited

  • "What does reddit say about [store]?" (store named: 15 → 14 of 15; own site cited: 9 → 0 of 15)

  • "Has [store] been mentioned in the news?" (store named: 15 → 15 of 15; own site cited: 11 → 2 of 15)

  • "What do social media sites say about [store]?" (store named: 15 → 15 of 15; own site cited: 14 → 3 of 15)

  • "What do customers say about [store]?" (store named: 14 → 15 of 15; own site cited: 4 → 4 of 15)

  • "What kind of reviews does [store] have?" (store named: 15 → 15 of 15; own site cited: 9 → 4 of 15)

  • "Is [store] a trustworthy dealer to buy from?" (store named: 15 → 15 of 15; own site cited: 9 → 5 of 15)

What this means for you: your own pages need to answer the questions shoppers ask about you, not just the ones they ask you. Keep your review profiles current, because that's where ChatGPT goes for these answers.

Finding #6: When ChatGPT Gets Specific About Price, Here Is Where the Numbers Come From

Most money answers stayed general. Only about one in three answers to price and financing questions included an actual dollar figure or interest rate. The rest gave advice like "ask for the out-the-door price."

When ChatGPT did get specific, here is what it said and where it found it. Examples are quoted from real answers.

What ChatGPT said

Prompt that produced it most often

Example from an answer

Where it came from

How often

A specific car and its price

"Are cars overpriced at [store]?"

"A 2025 Volvo XC90 Plus with about 18,465 miles was listed around $52,000."

The dealer's own inventory pages, often alongside Cars.com listings of the same cars. Every dealer site (15 of 15) shows vehicle prices on vehicle pages or listings.

23 answers, 9 dealers

A grade for the dealer's pricing from CarEdge

"Are cars overpriced at [store]?"

"CarEdge gives it an A ... grade (99/100) ... No mandatory dealer add-ons were found ... The documentation fee was $425."

CarEdge, a site that grades dealers on verified out-the-door price quotes. ChatGPT named CarEdge in the answer. CarEdge pages are not on dealer sites. ChatGPT cited the store's own CarEdge page for 11 of the 15 stores. For the rest, it cited CarEdge pages for other businesses, sometimes another brand's store in the same town.

19 answers, 5 dealers

Monthly lease payments

"Are the lease prices fair at [store]?"

XC40: "advertised lease example around $543.86/month for 36 months"

Dealer specials pages and volvocars.com offers. Nine of the 15 dealer sites publish lease payments, on specials or vehicle pages.

19 answers, 12 dealers

Service prices and coupons

"Is it expensive to service my car at [store]?"

"Oil change: about $199.99 before any applicable discounts."

Dealer service and coupon pages. Six of the 15 dealer sites publish service prices.

18 answers, 11 dealers

A named pricing policy

"Are cars overpriced at [store]?"

A store's no-haggle pricing program, or an offer to meet or beat a written quote from another dealer

Dealer websites. Five of the 15 dealer sites have a named pricing policy.

14 answers, 6 dealers

Doc fees

"Are used car prices fair at [store]?"

"Their listings show a $995 processing fee on some vehicles." (Fees quoted ranged from $169 to $995.)

CarEdge's dealer reports, or the fee printed on the dealer's own listings. Fourteen of the 15 dealer sites publish their doc fee, mostly in vehicle-page fine print.

13 answers, 6 dealers

Promotional financing rates

"Does [store] have any special offers or custom financing?"

"1.99% APR for up to 72 months on select 2026 XC90 and XC60 models."

volvocars.com national offers, and dealer certified pre-owned specials. All 15 dealer sites show a promotional rate, usually on vehicle pages or the certified pre-owned page.

12 answers, 10 dealers

Credit score guidance

"What credit score do I need to lease or finance at [store]?"

"750+ credit score: Usually puts you in the best position for top lease programs"

No dealer published this. ChatGPT gave a general rule of thumb. Four of the 15 dealer sites mention financing for all credit types, but none gives credit score guidance.

17 answers, 14 dealers

Outside sources didn't replace your site. In the 19 answers that quoted a CarEdge grade, ChatGPT also cited the dealer's own website in 17. In the 23 answers that named specific cars and prices, it cited the dealer's site in 22 and Cars.com in 11.

CarEdge was cited in about one answer in four, in every store's results, mostly for a letter grade on the store's pricing and whether it adds dealer add-ons.

What this means for you: ChatGPT will quote your numbers, but only the ones it can find. Most dealers in this study already publish fees, payments, and prices, deep in their vehicle pages. Test bringing a few of those numbers up to pages ChatGPT already reads, on one or two stores, and measure the answers before and after rolling it out. Also look up your store on CarEdge, because ChatGPT often quotes its grade when it judges your pricing.

Finding #7: ChatGPT Tells Shoppers What It Can't Find

In 16% of August answers, ChatGPT said outright that it couldn't find the information. The most common gaps:

  • service wait times ("How long is the wait at [store] service?");

  • minimum credit score or approval requirements ("What credit score do I need to lease or finance at [store]?");

  • news coverage of the store ("Has [store] been mentioned in the news?");

  • number of employees, and how long the store has been in business ("How many employees does [store] have?" and "How long has [store] been in business?");

  • whether customers come back to buy again ("Does [store] have repeat customers?");

  • trade-in values ("...Does [store] offer good prices on trade-ins?");

  • how long finance approval takes ("How long does finance approval take at [store]?").

The first five gaps showed up for most of the 15 stores, not just a few. In November, it happened in 16% of the repeated answers, and the list barely changed: trade-in values (17 answers), employee count (14), and Reddit discussion, service wait times, minimum credit score, and finance approval time (13 each) led it.

What this means for you: this is the easiest list in this report to act on. Every gap is a question shoppers ask, and AI can't answer. A well-written About page, a financing FAQ, and a service page with typical wait times close most of them.

Finding #8: ChatGPT Is Saying More With Fewer Sources

In November, ChatGPT cited about four links per answer instead of nine. It didn't talk less. On the repeated questions, answers got about a fifth longer, and seven in ten were longer than the August answer to the same question for the same store.

On the 1,094 repeated answers

August

November

Median words per answer

240

282

Average words per answer

246

299

Answers over 400 words

6%

23%

Links cited per answer

9.4

3.8

Words per cited link

about 26

about 79

Ended with a question for the shopper

80%

93%

Answers grew most on open-ended questions. Real shopper situations ("It's my first time buying from a dealership...") rose from a median of 377 words to 498; questions about what makes the store different rose from 265 to 374; and trust questions rose from 250 to 324. Service, inventory, and price answers barely changed.

What this means for you: ChatGPT now writes more from fewer sources, so each page it does cite carries more of the answer. It also ends almost every answer by asking about model, budget, and how the shopper wants to buy. Make sure the pages it already reads answer those follow-ups too: which models you stock, how trade-ins work, and what financing looks like.

What We Are Not Claiming

  • These are two moments. Each question ran once in August 2026 and once in November 2026. AI answers change from run to run and month to month, and we can't fully separate changes in ChatGPT from changes dealers made in between.

  • Location wasn't controlled. Where a question is asked from changes the answer. We did not test from inside each store's market.

Where This Leaves Your Store

Most dealers already show up in ChatGPT. The question is what it says when it gets there. Start here:

  1. Ask ChatGPT about your store. Use your exact name, then your name with your city and state. Is it talking about you?

  2. Fix the facts it can't find. Build or update your About page, a financing FAQ, and a service page with real details.

  3. Make your money answers easy to find. Most dealers already publish their doc fee and payments, deep in vehicle pages. Test moving them onto pages ChatGPT already reads, on a store or two first.

  4. Treat review profiles like web pages. DealerRater, Birdeye, CarEdge, Google, Carfax, and Cars.com decide your trust and "best dealer" answers, and which ones ChatGPT leans on can change in a few months.

  5. Speak to your nearby towns. Give shoppers outside your city a reason to choose you.

Want to know how ChatGPT describes your store today, and which of these fixes would move it most? Book a Search-to-Sale Audit.