How to check whether ChatGPT recommends your brand (a 30-minute method)

Why does it matter whether ChatGPT recommends you?

It matters because ChatGPT sits in the buying path of a very large audience, and its answers name brands. ChatGPT has 900 million weekly active users and 50 million paying subscribers, up from 800 million weekly users in October 2025 (TechCrunch, 27 February 2026). When someone asks it for the best hair oil under Rs 500, the answer names brands. If yours is not one of them, that buyer never sees you.

Adobe Digital Insights found AI-referred traffic to US retail sites grew 138% year over year in May 2026, and those visitors converted 54% better (Digital Commerce 360, 17 June 2026). That is US retail data, so treat it as direction, not a forecast for your store.

Google's AI summaries already cut clicks to traditional results. Pew found users clicked a traditional result in 8% of visits when an AI summary appeared, versus 15% without (Pew Research Center, 22 July 2025). Fewer clicks from blue links means an answer engine mention carries more weight than it did two years ago.

Which prompts should you test?

Test the prompts your buyers actually type, not your brand name. A brand-name query only tells you whether ChatGPT knows you exist. Category, comparison and problem prompts tell you whether it would send anyone your way.

Build a set of 20 to 30 prompts across those three groups. Here are two starter sets.

Indian D2C (skincare and haircare example)

1. Best cold-pressed hair oil brands in India under Rs 500
2. Which Indian sunscreen is good for oily skin and does not leave a white cast
3. Is [brand A] better than [brand B] for sensitive skin
4. Best paraben-free shampoo available on Amazon India
5. Recommend a night cream for dry skin in Delhi winters under Rs 800
6. Is [brand A] hair oil worth the price, and what do reviews say

B2B SaaS (marketing tools example)

1. Best HubSpot alternatives for a 30-person SaaS selling to the US
2. Which tools do agencies use for AI visibility tracking
3. Compare [tool A] and [tool B] for a startup with no sales team
4. Cheapest CRM with a native WhatsApp integration for Indian SMBs
5. Which SEO agencies in India also do AI search optimization
6. Is [tool A] good for a small team, and what are the complaints

Swap in your own category. Keep the phrasing messy and specific, the way people actually type.

How do you run the check manually in 30 minutes?

Run each prompt three times in a clean session, log what ChatGPT says, and score it in a table. Answers can vary between runs, so one pass is not a result. A first pass of ten prompts fits in 30 minutes. The full set takes a longer sitting.

1. Open ChatGPT in a temporary chat or a logged-out window. You want the answer a stranger gets, not one shaped by your own account.
2. Turn web search on. You are testing ChatGPT's search answers, the ones that can cite live pages.
3. Paste prompt one. Note every brand named, the order they appear in, and every URL cited.
4. Run the same prompt twice more, in a fresh temporary chat each time.
5. Record the result in the table below. Count a brand as "mentioned" if it appears in two of three runs.
6. Repeat for all 20 to 30 prompts, then total the columns.

The brands and sources in the example rows are placeholders, not results.

The last column is the most useful one. It tells you which brands ChatGPT favours in your category, and where its sources come from.

Citations are getting more common but are still rare. Similarweb reports the share of ChatGPT answers carrying a citation rose from 1.6% in June 2025 to 6.8% in May 2026 (Similarweb, 29 July 2026). A "mentioned, not cited" result is normal.

Can ChatGPT even reach your site?

If OpenAI's search crawler cannot fetch your pages, ChatGPT search will not cite them. OpenAI's crawler documentation states that OAI-SearchBot is used to surface websites in ChatGPT's search features, and that sites which disallow it will not appear in ChatGPT search answers (OpenAI crawler documentation, undated).

Three places to look:

• robots.txt. Open yoursite.com/robots.txt. Look for User-agent: OAI-SearchBot followed by Disallow: /. Also look for a blanket User-agent: * with Disallow: /. Either one blocks you. GPTBot is a separate crawler for training foundation models, and blocking it does not affect search visibility; ChatGPT-User fetches pages on a user's request, and robots.txt rules may not apply to it (OpenAI crawler documentation, undated). Our AI crawlers glossary entry covers the other engines.
• CDN and firewall rules. Bot-blocking settings are easy to switch on and forget. Ask whoever manages your CDN or firewall whether any rule matches OpenAI's user agents.
• Server logs. Search your access logs for "OAI-SearchBot". Requests returning 200 mean you are reachable. A 403, or nothing at all, means something upstream is blocking or the crawler has not found you yet.

What is the difference between being mentioned, cited and recommended?

Mentioned is your name in the answer, cited is a link to your page as a source, and recommended is ChatGPT placing you first or telling the user to pick you. Score all three separately.

• Mentioned. ChatGPT associates your name with the category. You get awareness, not a visit.
• Cited. ChatGPT's search feature found your page and used it as a source. This is an AI citation, and the only one of the three that can send traffic. Similarweb found that after ChatGPT's May 2026 link update, homepage referrals became roughly 62-63% of referred traffic, up from about 26-29% (Similarweb, 29 July 2026). So a citation now often lands people on your homepage, which needs to close the sale.
• Recommended. You are the answer, or first in a short list. This is the outcome that moves revenue. In our view it needs both of the above plus consistent third-party proof.

How do you track this over time, and where do you see Google's AI impressions?

Track it with the same prompts, the same method, on the same day each week, in a spreadsheet. If you change prompts every week, you cannot tell whether a shift came from your work or your method.

Keep one tab per week with three summary numbers: prompts mentioned, prompts cited, prompts recommended first. Plot them over eight to twelve weeks before you judge anything. A one-week swing means little.

For Google's side, use Search Console. Its Generative AI performance report shows impressions for AI Overviews and AI Mode, split by page, country, date and device. It rolled out to all sites worldwide by 31 August 2026 (Search Engine Journal, 31 August 2026). It shows impressions only, not clicks, and a property-level "search generative AI control" lets a site opt out of those surfaces (Google Search Console Help, 31 August 2026). For ChatGPT, your manual log is the report.

Our SNIFF audit is a one-time audit, priced at Rs 15,000 or USD 300, if you would rather we run this check for you. If you want a smaller taste first, there is a free visibility check.

What should you fix first if you are invisible?

Fix reachability first, then give each page one clear answer, then add real proof, then earn third-party mentions. In that order.

1. Indexability. A page that cannot be crawled or indexed cannot be cited. Check for noindex tags, blocked crawlers and broken canonicals. Our audit checklist is a starting point.
2. One clear answer per page. ChatGPT needs to lift a specific claim from your page. A product page that says "premium quality, loved by thousands" gives it nothing to quote. A page that says "cold-pressed coconut oil, 200 ml, Rs 349, no mineral oil, tested for heavy metals" does.
3. Real proof. Prices, specs, ingredient lists, test reports, named authors, dated reviews. Answer engines pull facts, not adjectives.
4. Third-party mentions. Look at the "who was cited instead" column. If the same comparison sites and forums keep appearing, you need to be on them. That is PR work, not on-page work.

This is the practical side of Generative Engine Optimization. Our AI search optimization service covers this work on a monthly retainer, from Rs 25,000 or USD 750 a month.

What does WoofRank's own score look like today?

You can see it at /client-zero/, where we publish our own numbers. We founded WoofRank in August 2026, it is founder-run, and we have no clients yet. So we are running the method in this post on woofrank.com first, and the results go on that page. If the method works, you will see the numbers move there. If it does not, you will see that too.

What else do people ask about this check?

Does blocking GPTBot stop ChatGPT from recommending me?

No. OpenAI says GPTBot is for training foundation models and blocking it does not affect search visibility (OpenAI crawler documentation, undated). OAI-SearchBot is the one that matters for ChatGPT search. Check both lines in your robots.txt, and your CDN rules too.

Does an llms.txt file help ChatGPT find my brand?

No evidence says so. Google's John Mueller called llms.txt "purely speculative for now" and said none of the AI systems use it (Search Engine Journal, 2 June 2026). We keep one at /llms.txt because it costs nothing, and we do not sell it as a tactic. Our post on llms.txt has the detail.

How long before I see my brand appear?

We do not know yet, and nobody can promise a timeline. Track weekly for eight to twelve weeks before you judge. In our view, consistent third-party mentions and clearly stated facts on your pages are the things to work on.

Sources

• TechCrunch, "ChatGPT reaches 900M weekly active users": https://techcrunch.com/2026/02/27/chatgpt-reaches-900m-weekly-active-users (27 February 2026)
• Digital Commerce 360, reporting Adobe Digital Insights: https://www.digitalcommerce360.com/2026/06/17/adobe-ai-referred-traffic-to-retail-sites-doubles-in-a-year/ (17 June 2026)
• Pew Research Center: https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/ (22 July 2025)
• Similarweb, generative AI stats: https://aisearch.similarweb.com/blog/gen-ai-stats/ (29 July 2026)
• OpenAI, crawler documentation: https://developers.openai.com/api/docs/bots (undated)
• Search Engine Journal, Search Console AI reports rolled out worldwide: https://www.searchenginejournal.com/google-search-console-ai-reports-rolled-out-worldwide/587836/ (31 August 2026)
• Google Search Console Help, Generative AI performance report: https://support.google.com/webmasters/answer/16984139 (31 August 2026)
• Search Engine Journal, "Google says llms.txt is purely speculative for now": https://www.searchenginejournal.com/google-says-llms-txt-is-purely-speculative-for-now/577576/ (2 June 2026)

Sources and accuracy

Researched and written by the WoofRank founders. Every statistic was checked against the primary source on 5 September 2026 and linked inline. If you find an error, write to woofrank26@gmail.com and we will correct it.