Your competitor is cited by ChatGPT because the engine can do three things with their brand that it cannot do with yours: retrieve their pages, extract a direct answer from them, and corroborate that answer across the rest of the web. Somewhere in that chain there is a break on your side, and the seven checks below will find it — or you can skip the manual work and run a free AnswerMonk audit, which probes your category across three engines and returns your appearance rate, a competitor leaderboard, and a citation-source breakdown in 3-8 minutes.
In July 2026 we ran the literal question "why is my competitor cited by ChatGPT and not me?" across three engines. Not one answer named a tool or vendor. The citations went to small personal blogs. If you typed that question and landed here, you have already experienced the problem firsthand: the engines do not yet have a good answer to it. What follows is our attempt at the definitive one, grounded in our 2,994-probe July 2026 study across 19 categories and the strongest external research we know of.
The seven checks, one sentence each:
- Retrieval: the engine may be answering from memory, where nothing you publish today can reach it.
- Bing indexation: ChatGPT Search leans on Bing, and plenty of sites that look healthy in Google are thin or absent there.
- Third-party sources: engines cite listicles and review platforms heavily, and your competitor is on them while you are not.
- Extractability: their page answers the buyer's question in a liftable format; yours makes the model do the work.
- Entity consistency: the web describes your competitor one way and describes you three different ways.
- Mention footprint: they get talked about off their own site; you mostly talk about yourself.
- Freshness: their cited pages are recently updated, and stale pages quietly fall out of answers.
Work through them in order. The first two rule out whole classes of problem in minutes, and in our experience most teams find their break somewhere in the first four.
Check 1: The engine never retrieved anything
If ChatGPT answered from memory, no page you publish can appear in that answer — so before auditing any content, confirm that retrieval actually happened.
Symptom: the answer that names your competitor carries no source links at all. No citations, no "searching the web" step, just a confident paragraph naming brands.
Likely cause: memory-only answering is far more common than most people assume. In our 2,994-probe July 2026 study across 19 categories, 30.1% of all answers cited zero sources — 900 of 2,994 were written purely from the model's training memory. In those answers, whoever was prominent in the training data wins, and your website is never consulted.
How to verify: ask your buyer's question twice, once generic and once specific, and watch whether citations appear. The clearest retrieval trigger in our data is specificity — a city name in particular. In our July 2026 probe run, identical vet-clinic questions asked without a city name got memory-only answers 81% of the time; add "Toronto" and that fell to 0% — every answer retrieved and cited live sources. The same flip showed up for Austin gyms (75% vs 0%), London plumbers (55% vs 6%), and Sydney accountants (55% vs 0%).
Fix: you cannot force retrieval, and you cannot edit a model's memory on any useful timescale. What you can do is recognize that real buyers usually ask specific questions — a city, a budget, a use case — and specific questions trigger retrieval. Win the retrieved answers first; the memory-only answers should follow slowly as your off-site footprint grows (Check 6) — that part is inference, not something we have measured. And if your own manual tests keep hitting memory-only answers, the diagnosis may simply be that you tested prompts nobody actually types.
Check 2: You are thin or absent in Bing
ChatGPT Search leans on Bing's index, not Google's: Seer Interactive's analysis found ~87% of ChatGPT Search citations align with Bing's top organic results.
Symptom: you rank respectably in Google, your competitor does not obviously outrank you there, and yet ChatGPT keeps citing them and skipping you.
Likely cause: a decade of SEO habits pointed at Google. Many sites have never verified themselves in Bing, have sitemap or crawl issues Bing never worked around, or block Bingbot in ways nobody noticed. Note what is not the cause: authority. Industry analyses find ~90% of pages ChatGPT cites rank on Google page 3 or beyond (position 21+) — domain authority is not the gate. Being present and relevant in the index that matters beats ranking first in the index that does not.
How to verify: search site:yourdomain.com in Bing and check that your key commercial pages appear. Then do the same for your competitor. Five minutes, and it catches one of the most common breaks we see.
Fix: verify your site in Bing Webmaster Tools, submit your sitemap, enable IndexNow so updates are picked up quickly, and clear any crawl errors or stray noindex directives. This is the cheapest fix on the list, which is exactly why it is worth doing first. For the broader mechanics of how answer engines source pages, see GEO vs SEO.
Check 3: You are missing from the sources engines actually cite
Engines mostly cite third parties, not vendors — listicles take roughly 41% of commercial-intent AI citations (industry citation analyses), and in our July 2026 live probes G2 and Capterra were recommended in nearly every answer about getting mentioned by ChatGPT.
Symptom: the citations under your competitor's mention are not their own website. They are "best X" roundups, review platforms, and directories, and your brand appears on none of them.
Likely cause: when an engine answers a commercial question, it prefers a neutral page that has already done the comparing. If the roundups in your category list your competitor and not you, the engine faithfully repeats that omission. The pattern extends beyond listicles: Reddit alone accounts for roughly 20-24% of Perplexity's citations (industry citation analyses), so a competitor who shows up in genuine community threads has a source you simply do not have.
How to verify: ask the real buyer question in ChatGPT and Perplexity, collect every domain they cite, and search each one for your brand. The list of cited-domains-where-you-are-absent is your gap, stated precisely.
Fix: pitch the specific listicles that were cited, not listicles in general. Claim and populate your review-platform profiles if you sell software. Participate honestly in the communities that keep appearing in citations. We cover the outreach mechanics in how to get traffic from ChatGPT.
Check 4: Their page is extractable and yours is not
In our crawl of pages that win AI citations, pages with FAQ blocks were 3.4x more likely to be cited, pages served with llms.txt 2.4x, and pages with question-form headings roughly 2x.
Symptom: both sites are indexed, both get retrieved, and the engine still quotes their page instead of yours.
Likely cause: extraction. Their page states the answer in two sentences directly under a heading that matches the question. Yours buries the same answer across four paragraphs of brand narrative. The model is assembling an answer under time pressure; it takes the liftable version. And do not assume your own site cannot be the cited source — in our 200-probe Dubai dental set, engines averaged 11.8 citations per answer and the most-cited sources were clinics' own websites (nicolasandasp.com, drjoydentalclinic.com), not health portals.
How to verify: open your competitor's cited page and your equivalent side by side. Sixty-second test: can you copy a complete two-sentence answer to the buyer's question from directly under a single heading? If it works on their page and fails on yours, this is your break.
Fix: restructure your key pages so every section leads with the answer. Question-form H2s, a direct first sentence under each, an FAQ block for the questions buyers actually ask, and an llms.txt file. None of this requires new content — it requires reordering the content you have.
Check 5: The web cannot agree on what you are
Engines consolidate everything the web says about a brand into one entity, and conflicting descriptions make you risky to name — this check is qualitative, but the failure is easy to spot.
Symptom: engines describe your company incorrectly, confuse you with a similarly named business, or hedge with phrases like "appears to offer," while describing your competitor crisply and confidently.
Likely cause: your homepage says one thing, your LinkedIn another, and a directory listing from 2021 a third. Rebrands, pivots, and abandoned profiles all leave residue. A model that cannot state confidently what you do will name the brand it can describe instead.
How to verify: ask "what does [your brand] do, and who is it for?" in ChatGPT, Gemini, and Perplexity. Compare the three answers with each other and with your homepage. Disagreement between the engines is the tell.
Fix: write one canonical one-sentence description and propagate it everywhere — homepage, LinkedIn, review profiles, directories. Add Schema.org Organization markup so the canonical version is machine-readable. Then fix or kill the stale listings.
Check 6: Your competitor gets talked about; you talk about yourself
Third-party mentions are the strongest measured predictor of AI visibility: Ahrefs' 75,000-brand analysis found branded web mentions correlate 0.664 with AI visibility versus 0.218 for backlinks.
Symptom: your on-site content is at least as good as theirs, all the technical checks pass, and they still win. Off-site, though, they are in podcasts, press, and community threads, and you are not.
Likely cause: engines corroborate before they recommend. Muck Rack's citation research found 84% of AI citations come from earned media, and in the Ahrefs data YouTube mentions were the strongest single signal at ~0.737. It can also be how young brands beat incumbents: in our Dubai corporate-card probes, fintech startup Pemo appeared in 83% of answers, ahead of major banks like Emirates NBD. Our reading — consistent with the Ahrefs data, though we didn't measure it directly — is that the banks have the authority while Pemo has the mentions.
How to verify: search your competitor's name excluding their own domain, then do the same for yours. Compare the volume and recency of independent mentions. The gap is usually not subtle.
Fix: shift effort from backlink acquisition to mention acquisition — industry press, podcasts, YouTube appearances, honest community participation. This is the slowest fix on the list and the biggest lever, because it also feeds the memory-only answers from Check 1. We go deeper in how to increase organic traffic from LLMs.
Check 7: Their cited pages are fresh and yours have gone stale
New pages either get cited fast or mostly never: Semrush's September 2025 longitudinal test (81 new pages, 30 days) found 42% were cited by ChatGPT within 30 days, and the rest mostly never were — Google AI Mode cited 36% on day one.
Symptom: you used to appear in answers and quietly stopped, or your competitor's cited pages all show update dates from the last few months while yours date from 2024.
Likely cause: retrieval favors current pages. The freshness effect is qualitative in our own data — we have not put a multiplier on it, so treat it as directional — but qualitatively, recently updated pages seem to be favored — treat this check as directional rather than measured.
How to verify: check the visible dates and last-modified headers on the pages the engines actually cite in your category, then on your own equivalents.
Fix: update the pages you need cited, substantively rather than cosmetically, and show the date. And take the Semrush finding seriously in the other direction: if a page has earned no citation after a month or two, waiting is unlikely to change that. Restructure it per Check 4 instead.
The 7 checks at a glance
| Check | 5-minute verification | Fix |
|---|---|---|
| 1. No retrieval | Ask the buyer question with and without a specific qualifier; look for source links | Target specific, retrieval-triggering prompts; grow footprint for the rest |
| 2. Bing gap | Search site:yourdomain.com in Bing | Bing Webmaster Tools, sitemap, IndexNow |
| 3. Missing from cited sources | Collect the cited domains under a real answer; search your brand on each | Pitch those exact listicles; claim review profiles |
| 4. Not extractable | Try to lift a two-sentence answer from under one heading on your page | Question-form H2s, answer-first paragraphs, FAQ block, llms.txt |
| 5. Entity confusion | Ask three engines what your brand does; compare their answers | One canonical description everywhere, plus Organization schema |
| 6. Weak mention footprint | Search your brand excluding your own site; repeat for the competitor | Earn mentions: press, podcasts, YouTube, communities |
| 7. Staleness | Compare update dates on their cited pages versus yours | Substantive updates on the pages you need cited |
How do I find which check is failing in my category?
Not with a single manual spot-check — answers vary too much for one prompt in one engine to diagnose anything. When we ran our 300-probe phrasing study, two engine presets looking at identical questions crowned the same winner less than a quarter of the time (22%). One anecdotal test can point you at the wrong check entirely.
A real diagnosis needs the same buyer questions run repeatedly across engines, with every cited source collected and counted. That is what AnswerMonk's free audit automates: enter your URL at answermonk.ai — no signup — and in 3-8 minutes it generates the buyer prompts in your category, runs them across ChatGPT, Gemini, and Claude, and returns your appearance rate, a competitor leaderboard, and a citation-source breakdown. The source breakdown in particular turns Checks 3, 4, and 6 from an afternoon of manual searching into a table. You can see what the output looks like in our public before/after reports at /reports.
One honest caveat to carry with you: some of the numbers above come from single-month studies, ours included. The mechanisms are stable; the exact percentages will move. Treat the checks as the durable part and the stats as this month's reading.
Run your free AI visibility audit →
Frequently asked questions
Why does ChatGPT cite my competitor instead of my website?
Because ChatGPT can retrieve, extract, and corroborate your competitor's brand and fails at one of those stages for yours. The most common breaks are absence from Bing's index, absence from the listicles and review platforms engines cite, and a weaker third-party mention footprint. Running the seven checks in order isolates which stage is failing for you.
Does Bing indexing really affect whether ChatGPT cites me?
Yes. Seer Interactive's analysis found roughly 87% of ChatGPT Search citations align with Bing's top organic results, so a site that ranks well in Google but is thin or missing in Bing can be invisible to ChatGPT Search. Verifying your site in Bing Webmaster Tools and submitting a sitemap is usually the fastest fix in the whole diagnostic.
Do G2 and Capterra profiles help me get mentioned by ChatGPT?
In our July 2026 live probes, G2 and Capterra were recommended in nearly every answer about getting mentioned by ChatGPT. Engines lean on review platforms and roundups as neutral third parties, and listicles take roughly 41% of commercial-intent AI citations in industry analyses. If you sell software and have no presence on those platforms, you are missing one of the most-cited source types in your category.
What does it mean when ChatGPT answers without citing any sources?
It means the model answered from training memory instead of searching the live web, which happened in 30.1% of the 2,994 probes we ran in July 2026. No page you publish can directly appear in those answers. The practical response is to target specific, retrieval-triggering questions first and build third-party mentions to shift the memory answers over time.
Do backlinks matter for ChatGPT visibility, or only brand mentions?
Ahrefs' 75,000-brand analysis found branded web mentions correlate 0.664 with AI visibility versus 0.218 for backlinks, and YouTube mentions were the strongest single signal at roughly 0.737. Backlinks are not worthless, but mentions are the far stronger measured predictor. Muck Rack's citation research points the same direction: 84% of AI citations come from earned media.
How does AnswerMonk's free audit diagnose why a competitor wins?
You enter your URL at answermonk.ai — no signup — and get results in 3-8 minutes. It builds your category's prompt set, runs it across all three engines, and hands back how often you appear, who appears instead of you, and which sources each engine trusted. That citation-source breakdown shows directly which of the seven checks is costing you answers.