Quick Summary
AI visibility competitor research answers one question your rank tracker cannot. When buyers in your market ask ChatGPT, Gemini, Perplexity or Google AI who to buy from and where to service, who gets named, who gets cited, and who is absent?
Our own market audits show why this needs measuring rather than guessing. The engines cite 121 to 457 distinct domains per market on just four buyer questions, the winning store captures under a quarter of those citations, and the same store can be cited in 45% of one engine's answers and 95% of another's. The dealers doing this research are finding the gaps and building the pages that turn AI answers into traffic and leads.
What You Should Know
For GMs
- Your buyers ask AI assistants who to trust, and somebody in your market is collecting those recommendations. This research tells you who, with counts instead of hunches.
- Your real AI competitors include Cars.com, DealerRater, local press and quick-lube chains, not just the rival rooftop. The winnable gaps are usually against the third parties.
- The free version costs three minutes: run the Saggy scan and read who the engines cited instead of you.
For Marketing Directors
- Add named-versus-cited to your competitor reporting. A rival who is mentioned everywhere but cited nowhere can be passed by becoming the source the engines quote.
- Sample every engine three times before drawing conclusions. The 45%-versus-95% disagreement we measured means one engine's answer is not your market.
- The cited sources are your content brief. Open what the engines quote on each lost question and build the page that beats it.
For Dealer Principals
- One in four new-vehicle buyers now shops with AI assistance, and AI answers have fewer seats than a results page. Presence in them is a market-share fight you are in whether you measure it or not.
- Rank-1 stores in our audits held under a quarter of their market's AI citations. Nobody owns this channel yet, which is exactly why moving early is cheap.
- Judge the research by the loop it feeds: gaps found, pages published, score re-measured monthly. A report that does not end in pages is a decoration.
“The spreadsheet that changes a GM's mind is never our pitch. It is the one from their own market audit, the one where the engines cited a quick-lube chain forty times on the service question and their store zero. After that, nobody asks why we measure competitors.”
Ryan Boyle
Director, A3 Brands
A buyer in your market opens ChatGPT tonight and types a question you will never see in any rank tracker. Something like where should I buy a Subaru near Prescott. The engine answers with names.
Either your store is in that answer or a competitor is collecting the lead you paid twenty years of reputation to earn. Classic competitor research tells you who outranks you on Google. It says nothing about who the AI engines hand your buyers to, and that is now a different list.
We measure this for a living, so this guide comes with our own numbers. It covers what AI visibility competitor research is and the questions and counts that make it real research instead of a vibe check. Then it shows what our audits found across dealer markets, and how the findings become pages that pull traffic and leads back from the engines.
What AI visibility competitor research is
Classic competitor research compares rankings. You pick rival rooftops, check who sits above you for your money keywords, and study their pages. Useful, and incomplete, because AI answers do not work like a results page.
When a buyer asks an AI assistant who to buy from, there are no ten blue links. The engine composes one answer. It names a handful of businesses and cites a handful of sources. Everyone else is invisible.
AI visibility competitor research measures that answer directly. It asks the engines the questions your buyers ask, then counts three things. Which businesses get named. Whose websites get cited as sources. And who owns the citations you assumed were yours.
The output is a map of your market as the engines see it, which turns out to be a different map than Google's page one. A store can rank second organically and never appear in an AI answer. We wrote about that mismatch in rankings look fine but leads are down, and the research in this guide is how you find out whether it is happening to you.
Why this research matters now
The buyer behavior has already crossed the threshold where this stops being optional.
Cox Automotive's 2025 Car Buyer Journey Study found 19% of all vehicle buyers and 25% of new-vehicle buyers used AI tools or AI-generated overviews while shopping. That was the study's first year measuring it.
The click side moved too. Pew Research monitored 900 US adults through a month of searching and found that when an AI summary appears, clicks on the traditional results below it drop from 15% to 8%.
Put those together and the competitive question changes shape. The fight is no longer only over positions on a page. It is over presence inside an answer, and an answer has far fewer seats. Somebody in your market is taking those seats right now, and the entire point of this research is that you should know their names before you spend another content dollar.
25%
Of New-Vehicle Buyers Use AI While Shopping
Cox Automotive's 2025 Car Buyer Journey Study found a quarter of new-vehicle buyers already use AI tools or AI overviews during the purchase, in the study's first year of measuring it.
Your real competitors are not who you think
Ask a GM who their competitors are and you get the same-brand store two towns over, maybe the mega-group's rooftop across the highway. The engines have a longer list.
When we run a full audit on one dealership market, four buyer questions surface between 121 and 457 distinct cited domains. The store's true rivals for AI visibility include businesses no dealer thinks about. Cars.com, AutoTrader, CarGurus and DealerRater collect citations on buying questions.
Local newspapers get quoted on dealership stories. Quick-lube chains and repair shops win the service questions, a fight we covered in why quick-lube chains outrank service departments.
Even the winners hold less ground than they assume. In our audits, stores that ranked first for AI citations in their own market captured only 15% to 24% of them. Three quarters of the market's AI visibility went to somebody else, mostly third parties.
That is the first practical payoff of the research. It replaces a mental list of two rival rooftops with the real roster of who the engines trust in your market, and the real roster is where the winnable gaps are.
A Dealer Market, as the Engines See It
121-457
Domains Cited Per Market
Distinct domains the engines cited answering four buyer questions in our audits
≤24%
Citation Share of the Winner
Rank-1 stores captured 15-24% of their market's citations; third parties took the rest
45% vs 95%
Same Store, Two Engines
Per-engine cited rates for one store in the same week, which is why single-engine checks mislead
The four questions to research
Buyer questions cluster into four shapes, and you can build all four from your marque and your city.
- ●best {marque} dealer in {your city}
- ●where should I buy a {marque} near {your city}
- ●best place to service a {marque} in {your city}
- ●which {marque} dealership near {your city} has the best reviews
The third one earns special attention. Service is where dealership gross profit lives, and it is the question where dealers lose to chains most often. A store can win the sales questions and send every oil change to a quick lube without knowing it.
These four are the same prompts our AI visibility score methodology runs, and they are deliberately plain. Buyers do not type marketing language. They type the question, and the research only works when you ask what they ask, word for word, without leading the engine toward your own name.
The two numbers to count
For every business that appears in an answer, count two different things, because they mean different futures.
Named means the answer recommends or mentions the business. Names come from whatever sources the engine read, so a store can be named often while its website contributes nothing. That visibility is rented. It rests on third-party content the store does not control, and it can vanish when a list article gets rewritten.
Cited means the engine links the business's own website as a source. Citation is owned visibility. The engine trusts that domain enough to quote it, which makes the position durable and, more importantly, clickable. A citation carries the buyer to the store's site. A bare mention carries them nowhere.
The gap between the two is the diagnosis. A competitor who is named everywhere but cited nowhere is coasting on lists, and you can pass them by becoming the source. A competitor whose own domain gets cited has content the engines trust, and their pages are the ones worth studying.
Count both per engine. The engines disagree more than anyone expects, and the disagreement is itself intelligence about where your next win is cheapest.
How to run it yourself
The manual version costs an afternoon and no money.
Ask every engine, not one.
Run each of the four questions in ChatGPT, Gemini and Perplexity, plus a plain Google search to catch the AI Overview. One engine is a sample of one and will mislead you.
Repeat each question three times.
The engines vary their answers between runs. A single run treats one roll of the dice as the truth. Three runs per engine per question shows you what is stable and what was luck.
Count names and citations separately.
A spreadsheet with one row per business and columns for named and cited, per engine, is the whole apparatus.
Open what the engines cite.
This step is where the strategy comes from. The cited sources are the content that wins your market, and reading them shows you exactly what the engines consider quotable.
When we ran this exercise on our own industry, we found the sources were mostly list articles the contestants had written themselves, a finding we published in full in our measured list of automotive SEO companies.
Our 15-minute AI search check covers the quick version if an afternoon is more than you have.
The DIY Research Loop
Ask Every Engine
Four buyer questions in ChatGPT, Gemini, Perplexity, plus Google for the AI Overview.
Sample Three Times
Engines vary between runs. Three samples per question separates signal from luck.
Count Named vs Cited
One spreadsheet row per business, columns per engine. Citations of their own domain matter most.
Read the Sources
Open what the engines quote. The cited pages define the bar your content has to clear.
What our own research shows
Three findings from our own measurement work, each one a reason this research pays.
The engines disagree about the same store.
In one August audit set, a store was cited in 45% of one engine's answers and 95% of another's. A dealer who checked only ChatGPT would have reached the opposite conclusion from one who checked only Perplexity. Every single-engine impression of your market is wrong somewhere.
The sources are narrower than the answers.
Engines name many businesses but lean on few sources. In our industry experiment, two list articles appeared in nearly every one of 39 answers. Whoever writes the page the engines quote controls what the engines say, which is the entire strategic insight of generative engine optimization compressed into one sentence.
Scale confirms it.
Our 2026 Dealership AI Search Study read 113,094 prompts across 661 US franchise dealerships and four engines. The pattern holds everywhere. Editorial content that answers buyer questions gets cited. Inventory pages and brochure copy do not. The dealers winning AI answers are the ones publishing the pages the engines need, and most markets still have nobody doing it.
From research to traffic and leads
The research is worthless if it ends in a spreadsheet. Here is the conversion path we run for clients.
Pick the questions you lost.
The audit tells you which of the four questions name competitors instead of you, and which engine. That is your priority list, ordered by lead value. Service questions usually pay fastest because the competition is thinner and the visits repeat.
Study the winning source, then beat it.
If the engine cites DealerRater on the reviews question, your review volume and your reviews page are the fight. If it cites a chain's oil-change page on the service question, a service page with your real prices, your real hours and your real answers is the fight. The cited source defines the bar. Clear it.
Publish answers, not brochures.
Across every measurement we run, the engines quote pages that state prices, name numbers, and answer one question all the way down. They skip pages that say welcome to our state-of-the-art facility.
Watch the score move, then the leads.
Citations show up first, traffic from AI referrals second, leads third. The stores we run this loop for have picked up 93% more leads in 60 days with content built from exactly this kind of gap analysis. The research finds the gap. The pages close it. The engines do the distribution.
Tools that do it for you
The manual method works. These are the faster versions, and two of the three are free.
The Saggy AI analyzer runs the whole measurement in about three minutes. It reads your website and builds the four buyer questions for your marque and market. Then it asks the engines across roughly 200 live responses and hands back your score with the competitor view built in.
You see your rank among every domain the engines cited in your market, and who holds the share you are missing. Run it free, once per store per day.
The Competitor DNA report goes deeper on the rivals themselves. It maps who is taking your traffic on Google and in AI answers, then shows what they win on and where they are exposed. Request it free and it arrives within 24 hours.
GALAXY Competitor DNA is the full research layer we build for clients, part of the GALAXY system that profiles your brand, your buyers and your competition before any page gets written.
Start with the free scan. It converts this whole article from an argument into your own numbers, and the numbers are more persuasive than we are.
The Same Measurement, Three Depths
The free Saggy scan gives you the three-minute market read. The free Competitor DNA report maps the rivals and their exposure within 24 hours. GALAXY Competitor DNA is the full research layer behind client strategies. All three run on real engine responses rather than estimates, and the first two cost nothing.
Cadence, and the traps to avoid
Run it monthly.
The engines update constantly, competitors publish, review counts shift. A quarterly check misses the month a rival's new service pages started collecting your citations. Monthly is the working rhythm, and it is why we track client scores through AI visibility monitoring instead of framing one audit.
Four traps catch almost everyone who starts this research.
The single-engine conclusion.
ChatGPT said the competitor's name, so panic. One engine, one run, means nothing. Sample all of them, repeatedly, before believing anything.
The vanity comparison.
Measuring yourself against the rival rooftop while ignoring the directories and chains that hold the real citation share. The research must include every domain the engines cite, not just the ones with a showroom.
Chasing names instead of citations.
Getting mentioned in more lists feels like progress and rests on content you do not control. The durable position is your own domain in the sources.
Doing the research and skipping the pages.
The most common failure. The audit is a map. Only published answers move the score, and only a moved score moves leads.
Key Takeaways
- ✓AI visibility competitor research measures who the engines name and cite when buyers ask where to buy and service, a different list from who outranks you on Google.
- ✓The market is wider than any dealer assumes. Our audits surface 121 to 457 cited domains per market, and rank-1 stores capture only 15% to 24% of citations.
- ✓Count named and cited separately. Mentions are rented visibility built on third-party lists; citations of your own domain are owned and durable.
- ✓The engines disagree about the same store, as far apart as 45% versus 95% cited rates, so single-engine checks reach wrong conclusions.
- ✓The cited sources are the strategy. Whoever writes the page the engines quote controls the answer, and most dealer markets still have nobody doing it.
- ✓Research converts to leads through pages. Pick the lost questions, study the winning source, publish a better answer, and watch citations then traffic then leads follow.
- ✓Run the measurement monthly. One in four new-vehicle buyers already shops with AI assistance, and the answers change under your feet.
Founder & President, A3 Brands
Tim spent a decade distributing products to 3,000+ dealerships, ran the Internet Sales department at Baker Automotive Group, and served as Acura's Field Program Manager and Digital Strategist at Shift Digital before founding A3 Brands — one of very few SEO agencies built exclusively for car dealerships.
Frequently Asked Questions
What is AI visibility competitor research?
How is it different from normal competitor SEO analysis?
Who are a dealership's competitors in AI search?
How do I see who ChatGPT recommends in my market?
Does AI visibility competitor research generate leads?
How often should the research be refreshed?
Sources & References
- Cox Automotive 2025 Car Buyer Journey Study (released January 2026) — 19% of all buyers and 25% of new-vehicle buyers used AI tools or AI-generated overviews while shopping; 2,300 consumers surveyed fall 2025
- Pew Research Center (March 2025) — Clicks on traditional results drop from 15% to 8% when an AI summary appears; 900 US adults, 68,879 searches
- A3 Brands market audit data (August 2026) — 121-457 distinct cited domains per dealer market on four buyer questions; rank-1 stores at 15-24% citation share; 45% vs 95% per-engine divergence for one store
- A3 Brands 2026 Dealership AI Search Study — 113,094 prompts across 661 US franchise dealerships and four engines; editorial answer pages win citations, inventory and brochure pages do not
Find Out Who AI Sends Your Buyers To
The free Saggy scan runs the four buyer questions across the engines. It shows you every domain cited in your market, your rank among them, and which questions you are losing. Three minutes, no credit card.