10 Checks to Find Out Why Your Competitor Shows Up in AI Answers (and You Don't)
In our study, a brand's own website supplied 2.9% of the citations behind AI answers, so the page that put your competitor there is usually someone else's. But one answer is a weak signal, and each engine names different rivals. Check that the gap is real first. Here are 10 checks in the order to run them, each with a free test.
When an AI engine names your competitor and not you, the reason is usually on pages you don't own, or in what the model already associates with your category. Your own website is rarely the deciding page: across the 149,912 citations in our GEO study, a brand's own site supplied 2.9% of them.
Before you change anything, check that the gap is real. One answer is a weak signal, and each engine keeps its own list of rivals. The ten checks below are in the order worth running them: three that test the gap, four that find the pages AI is reading, and three about your own site and brand. Each says what a positive result looks like and how to run it without paying for a tool. When you want the full spreadsheet version, our competitor audit picks up where this leaves off.
The 10 Checks at a Glance
| # | Check | If this is the cause, you will see | Run it yourself |
|---|---|---|---|
| 1 | Ask more than once | The names change from run to run | Ask again on another day, in a fresh session |
| 2 | Ask every engine | The rival leads on one engine only | Ask ChatGPT, Perplexity, Gemini, Google and Claude the same question |
| 3 | Read the question | The answer names no product, or asks for something only the rival offers | Check the answer names products; look for a price, place or size only the rival fits |
| 4 | Open the pages it cites | The cited pages name the rival and not you | Open every source under the answer and note the brands on each |
| 5 | Check the best-of lists | The rival is on a cited list and you are not | Search each cited list for your name |
| 6 | Check the communities | The cited threads recommend the rival | Read the Reddit threads and forum posts the answer cites |
| 7 | Check that a page names them | The rival is named, but no cited page mentions it | Compare the names in the answer with the pages cited under it |
| 8 | Check AI crawlers can fetch your site | A search crawler is refused what a browser gets | Read your robots.txt, then fetch your homepage as each crawler |
| 9 | Check AI knows what you are | AI describes you wrongly, or never ties you to the category | Ask each engine what your company does and who it competes with |
| 10 | Check whether they are simply bigger | The rival has the press, reviews and reference pages you lack | Look each of you up on Wikipedia, in the news and on review sites |
First, Check the Gap Is Real
Few guides start here, but these three checks decide whether there is anything to fix.
1. Ask more than once
Ask the same question again on another day, in a fresh session with no chat history, and write down who is named each time. Two runs are the minimum. Across our tracking runs, a brand named on a question, on the same engine, in two runs in a row was named on the next run 96.1% of the time, and one missing from two runs in a row came back only 3.4% of the time. A brand that had just appeared stayed 54.3% of the time, close to a coin toss. Count each engine separately.
Chance a brand is named on the next run
By what the last runs showed for the same question on the same engine. Measured across consecutive tracking runs of the brands Ranqo tracks.
- Named in the last two runs
- 96.1% named next run
- Named in the last run only
- 54.3% named next run
- Missing from the last two runs
- 3.4% named next run
A brand named twice in a row is almost certain to be named again, and one missing twice almost never comes back on its own. One sighting in between is close to a coin toss.
Scores built from many answers move too: between consecutive runs with nothing changed, an overall visibility score moved 14.5 points or more about one time in twenty. If the rival is named in both of your runs and you are named in neither, you have a gap. If the names come and go, you have noise, so keep asking before you change anything.
2. Ask every engine
Put the same question to ChatGPT, Perplexity, Gemini, Google's AI Overviews and Claude, or whichever of them your buyers use. Engines disagree about who your competitors are. Of the 2,317 company names ChatGPT, Perplexity and Gemini returned as competitors in our tracking, 73.8% were named by one engine only and 12.5% by all three.
How many engines named each competitor
2,317 distinct company names that ChatGPT, Perplexity and Gemini returned as competitors, July 6 to August 22, 2026. Each name is counted once per tracked brand.
- Named by all three engines
- 12.5%
- Named by two engines
- 13.7%
- Named by one engine only
- 73.8%
Even the top three rivals differ: between any two engines, a brand's top three overlapped by a median of 50%.
If the rival leads everywhere, keep going. If it leads on one engine only, the cause is that engine's sources, so run check 4 on that engine's answers.
3. Read the question
Two things in the question itself can explain a gap. First, make sure the answer names products at all. About one answer in seven names no vendor, and a brand appeared in only 4.3% of those, against 32.5% of the answers that did name vendors. A brand rarely wins a question like that, so it is a poor one to track.
Second, look for a constraint the rival meets and you don't: a price, a location, a company size, an integration. Then the answer may simply be right. Ask again without the constraint and see who comes back.
Then, Find the Pages AI Is Reading
If the gap holds up, the next question is where the rival's name comes from. Most of the time it is a page.
4. Open the pages it cites
Most engines list their sources with the answer. Open each one and note which brands it names. In our study, a brand's own site supplied 2.9% of citations and other companies' own sites 75.2%, with the rest coming from media, communities, video, reference, review and other sites. So the page that put your rival in the answer is usually someone else's, very often another company's own site.
A positive result here is the most useful one you can get, because it hands you a list of pages to work on. The audit shows how to log them across runs and turn them into targets.
5. Check the best-of lists
Among cited pages that carry content, rather than a homepage or a product page, ranked lists were the most cited format, with 35.7% of those citations. Search each cited list for your name. If the rival is on it and you are not, that list is a specific gap you can fix, and getting onto one has its own playbook.
6. Check the communities
ChatGPT leans on Reddit more than on any other site. In our matched runs from July 6 to August 22, 2026, Reddit was 17.5% of ChatGPT's citations, and 63.9% of its answers linked at least one Reddit thread. Perplexity's share was 7.7% and Gemini's 3.0%. Read the threads the answer cites: who gets recommended, and does anyone mention you? If they recommend the rival and never mention you, that is the gap, and working the subreddit is how to close it.
7. Check that a page names them
Sometimes the rival is named but none of the cited pages mention it. Then the name most likely came from what the model already knows, rather than from a page it just read, and there is no single page to get onto. Set those answers apart from the rest, so you don't pitch pages that had nothing to do with it, and look at checks 9 and 10: that kind of gap closes slowly.
Last, Check What AI Knows About You
8. Check AI crawlers can fetch your site
This is rarely the cause, but it is the cheapest check. Cloudflare reports that under 1% of its sites block search bots, against 17% that block training crawlers in some way. The difference matters. Blocking a training crawler such as GPTBot affects what future models learn, while blocking a search crawler such as OAI-SearchBot or PerplexityBot keeps your pages out of the answers being written now.
Read your robots.txt, then fetch your homepage as each crawler, because a firewall can refuse a bot that robots.txt allows. Our free AI crawler inspector runs that comparison against a normal browser, and our crawler check covers the rest.
9. Check AI knows what you are
Ask each engine, in a fresh session, what your company does and who it competes with. If the description is wrong, or puts you in another category, a question about the category won't bring you up, and correcting it comes first.
Being known by name is not enough on its own. In our study, engines recognised a brand in at least 94% of first-run answers to questions that named it. Asked about the category instead, ChatGPT, Gemini, and Perplexity named the same brands in 18.7% to 23.9% of answers. Claude and Grok ran on a smaller set of better-known brands, so their figures (Claude 51.5% and Grok 12%) are not comparable. If you are named only when the question names you, AI knows you but doesn't yet treat you as an answer, which is what entity work addresses.
10. Check whether they are simply bigger
Size shows up plainly in our study. In their first tracking run, brands with global stature averaged 72.9% visibility on questions that didn't name them, and niche brands averaged 11.4%. If your rival has a Wikipedia article, steady press coverage and years of reviews, a good part of the gap is that, and it won't close in a week. It closes slowly, and on other people's pages: the lists, threads and coverage from checks 4 to 6. How challengers climb covers the order to work on them.
What You Saw, and What It Most Likely Means
Most gaps have more than one cause. Read your results against this table, then start with the cause you can act on soonest.
| What you saw | Most likely means | Checks |
|---|---|---|
| The rival is named in some runs, and you in others | Normal variation; not a gap yet | 1 |
| The rival leads on one engine only | That engine's sources | 2, then 4 |
| The answer names no product at all | Nothing to win on that question | 3 |
| The cited pages name the rival, not you | You are missing from the pages AI reads | 4 to 6 |
| A cited best-of list includes the rival | Get onto that list | 5 |
| The cited threads recommend the rival | Join the conversation AI reads | 6 |
| The rival is named, but no cited page mentions it | The model's memory; a slower fix | 7, 9, 10 |
| A crawler is refused what a browser gets | AI cannot read your pages | 8 |
| AI describes your company wrongly | Fix the facts on the pages it reads | 9 |
| You are named only when the question names you | AI knows you, but not as an answer to the category | 9, 10 |
Once you know the cause, our six-step guide takes one question from here to a published page, and judges the result by the answers it was meant to change, using the protocol for telling whether a change worked.
How Ranqo Runs These Checks
Ranqo asks your tracked questions every week on each engine your plan covers, so the answers behind the first two checks build up on their own. A new brand needs its second run before check 1 means anything.
The gap, engine by engine
The Competitors page's Head-to-head tab puts you beside one rival, counts the prompts each of you leads, and lists the prompts where the rival was named instead of you or ranked above you, engine by engine. The Prompts page's Tracking tab shows, for every question, whether each engine's latest answer named you and which competitors it named most.
The pages behind it
Sources lists every site and page cited in your answers, with each page's format, so the ranked lists stand out. The last card on Head-to-head lists the pages and threads AI cited in answers that named the rival and left you out, and Outreach ranks those communities and publications and drafts a reply or a pitch when you ask. Every answer stays on its prompt's insights page with the brands it named and the pages it cited, which is where check 7 happens.
Your site, and the fix
Site Access fetches your homepage as each AI crawler and as a browser, and flags any crawler refused what the browser got. When you work a gap through an Action Center recommendation, it takes one reading when you press Re-measure now and judges it against measured noise bands: a move inside the normal week-to-week swing is never shown as a lift in points. You see the answers won instead, or No detectable change.
See where your rival is ahead, and why
Ranqo tracks your buyers' questions every week on ChatGPT, Perplexity, Gemini and the other engines your plan covers, shows the prompts each rival wins and the pages behind them, then turns the gaps into prioritized actions and content. The step-by-step version of this diagnosis is our competitor audit.
Start a 7-day trialWritten by
Nisha Kumari
Nisha Kumari is Co-Founder at Ranqo, where she leads growth strategy and client acquisition. With a background in digital marketing and financial management, she specializes in SEO, Generative Engine Optimization, and helping brands build visibility across AI platforms.
Share this article
Related articles
How to Audit a Competitor's AI Visibility (and Find the Pages Doing the Work)
Most competitor audits stop at a spreadsheet of who got mentioned, which tells you a rival is winning without telling you why. Only 2.9% of AI citations point at a brand's own domain, so the page carrying your competitor almost always belongs to someone else. Here is the seven-step audit that finds it, and the part no audit can fix.
Every AI Engine Names Different Competitors for Your Brand
Build a competitor list from one AI engine and roughly three quarters of it is that engine's opinion alone. Across 2,317 company names returned for 26 brands, only 12.5% were named by all three engines. The ordering is worse: in most cells the leading rival is ahead by a single mention. Treat the set as a set.
Your Website Is Only 2.9% of AI Citations. Here's Where the Other 97% Comes From.
Only 2.9% of the 149,912 AI citations we measured pointed to a brand's own website. Yet most AI-visibility advice still starts with 'optimize your pages.' In AI search your own domain is the floor, not the lever; the other 97% is earned on third-party pages. Here's the full source breakdown, and how to earn your way in.