Skip to main content
Back to List
geo·Author: RanketAI Editorial Team·Updated: 2026-10-18

How to Check If ChatGPT Mentions Your Brand: 5 Steps (and When Only Competitors Show Up)

A five-step way to check whether ChatGPT, Gemini and Perplexity mention your brand: unbranded questions, checking for sources, repeating each question three times and comparing services — plus how to read results when only competitors appear.

AI-assisted drafting · reviewed by the RanketAI Editorial Team — Editorial policy ›

Key takeaway: To check whether ChatGPT mentions your company, ask unbranded category questions in a fresh chat, note whether each answer carries sources, and see whether the result holds when you repeat the question. A question that includes your brand name ("Do you know our company?") almost always gets your brand mentioned by an AI with web search, so it tells you nothing. When RanketAI measured 12 Korean B2B SaaS companies, ChatGPT, Gemini and Perplexity all mentioned 83% of the brands, but the share citing the brand's own site as a source was 83%, 33% and 67% respectively. This guide covers five steps you can run yourself and how to read the results when only competitors appear.


Three-line summary

  • Check in a fresh chat with no personalization, using unbranded questions your customers actually ask.
  • Record whether each answer has sources, and repeat each question at least three times, judging by how often you appear.
  • Check ChatGPT, Gemini and Perplexity separately. For the same question, whether a search runs and which types of sites get cited differ by service.

Step 1: Start a fresh chat with no personalization

A signed-in ChatGPT can reflect saved memories and custom instructions in its answers, so check in an unpersonalized temporary chat or while signed out.

If you often talk about your company with ChatGPT at work, that history can be saved as memory and influence answers. Your customers ask without that history, so you should check under the same condition. According to OpenAI's help center, a temporary chat started without personalization does not use saved memories or custom instructions (OpenAI Temporary Chat FAQ).

Apply the same rule in Gemini and Perplexity: check in a new chat that does not continue an earlier conversation. If you ask several questions in one thread, the context of earlier questions affects later answers, so open a new chat for each question.

Step 2: Prepare unbranded questions your customers actually ask

Judge visibility only with questions that do not include your brand name, and prepare at least three: one recommendation, one comparison and one problem-solving question.

Type Example question What it checks
Recommendation Recommend good services for [category] Whether you make the candidate list
Comparison Compare features and pricing of [category] services Whether you are treated as a comparison option
Problem-solving How can I solve [customer problem]? Whether you are linked to the solution

Type the questions exactly as customers phrase them. Questions you hear often in sales calls, or search queries in Search Console, are good starting points. Adding conditions such as "only Korean services" in front of the question makes the answer differ from what customers see.

The reason to avoid branded questions is that their result is nearly fixed. In RanketAI's October measurement, well-known brands were mentioned by Gemini even in answers produced without a search. Confirming that "AI knows our company" says nothing about visibility. The details are in our post on the branded-query trap.

Step 3: Check whether the answer carries sources

If there are no source links under or beside the answer, it came without a web search, so first separate out answers in which your site could not have been cited at all.

Some AI services decide for each question whether to search the web. OpenAI explains that ChatGPT decides whether to search based on the question, and that users can also turn search on themselves (OpenAI, 2024-10-31). Google says the same in its Gemini Apps help.

"Not all responses include related links or sources. If you don't have the Sources button below a response, Gemini Apps didn't provide any links for that particular response." — Google, Gemini Apps Help

This changes how you read results. When RanketAI measured with web search enabled, every ChatGPT and Perplexity answer ran a search, but 62% of Gemini answers came without a search and had no sources (October measurement analysis). If your site is missing from an answer with no sources, the site is not the problem.

Record four things for each answer:

  1. Whether your brand was mentioned, and in what position
  2. Whether there is a source list
  3. Whether your site is in the source list
  4. Which competing brands were mentioned instead of you

Step 4: Repeat each question at least three times

AI answers change even for the same question, so record how many times out of several you appear, not the result of a single run.

In a study by SparkToro and Gumshoe that repeated the same prompts, the chance that ChatGPT or Google's AI search returned the same list of brands was under 1 in 100, and the same list in the same order appeared less than 0.1% of the time. The researchers concluded that appearance rate matters more than rank.

"Visibility percentage across many queries shows more consistency than ranking position." — Search Engine Journal, 2026-01-30

So record results as "mentioned 2 out of 3 times", not "recommended second." Do not treat one miss as failure, and do not relax after one appearance. When you check again on another date, keep the same questions and conditions so the results stay comparable.

Step 5: Check ChatGPT, Gemini and Perplexity separately

Even for the same question, AI services differ in whether they search and which types of sites they cite, so one service's result cannot stand for the others.

In RanketAI's October measurement, all three services mentioned 83% of brands, but the share citing the brand's own site as a source was 83% for ChatGPT, 67% for Perplexity and 33% for Gemini (October benchmark). The types of sites cited in searched answers also differed by service.

Service What to look for
ChatGPT Answers often attached each listed product's official page as a source. Check whether your official page is attached
Perplexity SaaS comparison and catalog sites, blogs and IT media were cited most. Check whether those sites carry information about you
Gemini Many answers came without a search. Judge citation only in answers with a Sources button

Reading the results when only competitors appear

If only competitors appear, first confirm the result repeats and that the answers ran a search, then look at which sources the competitors were cited from.

Result Meaning Next step
Only competitors in all 3 runs, with sources Competitors' material is being cited as the basis of the answer Check which sites cite competitors in the source list — source-gap analysis
Missing only in answers without sources What the model learned is not enough to make you a candidate First secure a way into answers that run a search
Sometimes there, sometimes not You are at the edge of the candidate list Track your mention rate with repeated checks
Mentioned but described wrongly Outdated information or confusion with another brand is likely Clean up the information on your official pages — fixing AI brand misrepresentation
Mentioned, and your site is in the sources You are in good shape Re-check regularly under the same conditions

Why only competitors get recommended, and the steps to close the gap, are covered in When AI Recommends Competitors but Omits Your Brand.

Where checking by hand stops being practical

Checking 3 questions in 3 AI services 3 times each means reading 27 answers and logging their sources, so regular checks are realistic only with automated measurement.

Running the check yourself once is worthwhile. You get a feel for which brands AI recommends in your category and on what basis. But adding questions, repeating them weekly or monthly under the same conditions and cross-checking source lists is hard to keep doing by hand.

AI visibility tools such as RanketAI automate this. RanketAI's AI brand visibility analysis takes your brand name and domain and measures ChatGPT, Gemini and Perplexity answers to a range of questions. The "Brand mention × Source citation analysis" card on the results screen shows mention and citation status for each AI service, and how many answers came without a web search when there are any. To compare side by side with competing brands, use competitor comparison. After signing in, you can try brand visibility analysis three times for free.

FAQ

Can't I just ask ChatGPT whether it knows our company?▾

That does not check visibility. When the question contains your brand name, the AI searches for that name or recalls what it learned, so it almost always mentions you. Customers ask "recommend a [category] service" without knowing your name, and whether you appear there is what visibility means. Use branded questions only to check whether AI describes your company accurately.

Do results differ if I check while signed in?▾

They can. If saved memories or custom instructions are reflected, an account that talks about your company a lot may see your brand more often. Checking in an unpersonalized temporary chat or while signed out is closer to what customers see.

If my brand showed up once, am I done?▾

No. Research shows recommendation lists change almost every time the same question is repeated. Check each question at least three times, record "how many out of 3," and check again on another date under the same conditions to judge change.

Results change every time — what should I use as the baseline?▾

Use mention rate instead of rank. Recording rate and sources together, such as "mentioned 2 out of 3 times, our site cited once," lets you compare runs. How much results move even when re-measured on the same day is covered in our August re-measurement analysis.

Can I check for free?▾

The five steps in this guide cost nothing to run yourself. For automated measurement, you can try RanketAI's brand visibility analysis three times for free after signing in, and check whether your site is structured for AI to read with the page structure check, no sign-in needed.

Does the same method work for Naver AI Briefing?▾

The principles for choosing questions are the same. Because AI Briefing appears inside Naver search results, search the same question on Naver and check whether a briefing appears and whether your content is among its sources. RanketAI measures ChatGPT, Gemini and Perplexity. How Naver's AI search works is covered in our Naver AI search overview.

Execution Summary

ItemPractical guideline
Core topicHow to Check If ChatGPT Mentions Your Brand: 5 Steps (and When Only Competitors Show Up)
Best fitPrioritize for geo workflows
Primary actionStandardize an input contract (objective, audience, sources, output format)
Risk checkValidate unsupported claims, policy violations, and format compliance
Next stepStore failures as reusable patterns to reduce repeat issues

Data Basis

  • Own measurement: report figures from RanketAI's October 2026 Korean SaaS AI Visibility Benchmark (2026-10-16, 12 Korean B2B SaaS companies, three same-day rounds, ChatGPT, Gemini and Perplexity with web search enabled) and its per-answer re-analysis (published in the 2026-10-18 post). The sample is weighted toward well-known brands, so the size of the rates is not generalized.
  • External sources: the Google Gemini Apps Help Center article and Search Engine Journal's report on the SparkToro and Gumshoe study (2026-01-30) were checked in the original on 2026-10-18. OpenAI's ChatGPT search launch post (2024-10-31) and Temporary Chat help article could not be opened directly, so their content was confirmed from public summaries in search results on 2026-10-18 and is paraphrased, not quoted.
  • Example questions use placeholders for the category and the customer's problem. The questions and the number of prompts RanketAI uses in its measurement are not disclosed.

Key Claims and Sources

This section maps key claims to their supporting sources one by one for fast verification. Review each claim together with its original reference link below.

External References

The links below are original sources directly used for the claims and numbers in this post. Checking source context reduces interpretation gaps and speeds up re-validation.

Is your site visible in AI search?

See for free how ChatGPT, Perplexity, and Gemini describe your brand.

Start Free Diagnosis →

Keep reading

Related Posts

These related posts are selected to help validate the same decision criteria in different contexts. Read them in order below to broaden comparison perspectives.

62% of Gemini Answers Came Without a Web Search — Korean SaaS Benchmark, October 2026

We measured 12 Korean B2B SaaS firms three times in one day. ChatGPT, Gemini and Perplexity all mentioned 83% of brands, yet own-domain citation was 83%, 33% and 67%. 62% of Gemini answers ran no web search; searched ones cited the own domain 52% of the time.

2026-10-18

AI Mentions the Brands It Mentioned Before — A 35K-Observation Study and Korean SaaS Data

A 2026 study of 34,960 unbranded prompts found mention rates near 3% when a brand's own domain was absent from sources and 49–58% when cited; prior mentions predicted best. Our 12-company Korean SaaS re-analysis agrees.

2026-10-06

What Does a GEO Score Measure? Prompts, Models, and Judges Change the Number (2026)

Same questions, same day, only the model changed: our brand mention rate was 43% versus 5%. A September 2026 methods paper explains why GEO scores depend on prompt sets, execution conditions, weights, and judges, and what to check in an AI visibility report.

2026-09-24

The Demand Side AI Visibility Misses — Which Sources Your Buyers Trust

AI visibility measurement checks one direction: whether your pages got cited. An August 2026 preprint releases a million-persona buyer corpus arguing the missing half is which sources buyers trust. Here is the axis you can apply today, and its limits.

2026-09-06

Perplexity Never Skips the Web — What the Answer Stream Reveals About Citations (2026)

A stream-level observation of Perplexity: all seven test queries triggered a live web search, yet only 5 to 9 of the 10 to 15 retrieved sources earned a citation. Local cited place entities, how-to cited video, comparison cited the vendor's own page.

2026-07-31