Nobody can make ChatGPT recommend you, but you can raise the odds. When it searches the web for a "who is a good X" question, it needs to find several readable pages that name you, state your specialism in the same words, and show proof of the work. Then you check with the same questions every week, because the answer changes from one ask to the next.
The rest of this article covers how assistants pick people, which sources they lean on, what you control, and how to measure it without fooling yourself.
What makes ChatGPT name a person for "who is a good X" questions?
Two things decide it: the type of question and whether the assistant searches.
Assistants name people and firms mostly when someone asks who to hire or how to choose. In a test we ran on 23 September 2026 with 20 German business questions, the model answered 10 without searching at all and cited nobody. Where it did search, it cited primary sources (laws, ministries, the OECD, large accounting firms) on 9 of 10. Firms were named only on "who should I hire" and "how to choose" questions. Two runs of the same question named two different sets of firms.
We saw the same pattern for a local practice. On 20 October 2026 we asked the four main assistants 20 patient questions about an orthopaedic practice in Bonn. The practice was named in 12 of 80 answers: Claude 5 of 20, Gemini 4, ChatGPT 2, Perplexity 1. Questions about who to choose named practices. Questions about costs, risks and treatments named none.
For you, this narrows the target. On general knowledge questions, the model in our test either answered from memory or cited primary sources such as laws and ministries, not individual experts. Recommendations happen on a smaller set of "who should I hire" questions, usually after a live search. Those are the questions to aim at and to test.
What counts as a credible page? Assistants are not Google's raters, but the logic is close. Google's Search Quality Rater Guidelines overview says raters judge pages on Experience, Expertise, Authoritativeness and Trust. The guidelines contrast rewiring advice from a skilled electrician with advice from an antiques enthusiast who knows nothing about wiring. An assistant choosing between two names faces a similar question: whose pages show real, first-hand work in this field?
Which pages and mentions does an assistant rely on?
When it searches, an assistant pulls a handful of pages and builds its answer from them. For a person, those pages usually fall into four groups.
| Source | What it tells the assistant | How much you control it |
|---|---|---|
| Your personal website | Who you are, your exact specialism, your work, cases, writing | Fully, including whether crawlers may read it |
| Directories and lists | That a third party files you under a category and location | Partly: you can claim and complete profiles |
| Other people's articles, podcasts, event pages | That others treat you as an expert worth quoting | Indirectly: you earn these by being useful |
| Your LinkedIn and X posts | Your topics over time, how you phrase your specialism | Fully for content; access for crawlers is set by the platform |
Directories carry more weight than people expect. In the Bonn measurement, the directory jameda.de was cited in 12 answers, while competing practices' own websites were the most cited sources. Your own site and a trusted listing work together.
Who gets cited is worth watching too. On 21 October 2026 we asked ChatGPT, Claude, Gemini and Perplexity how to get a business into AI answers. Most of the cited sources were blogs of GEO agencies and tool vendors. That is one measurement on one topic, so treat it as an observation, not a rule: in that set, the sites quoted were ones writing about their own field. It does not prove that publishing such pages will get any particular person quoted.
A note on LinkedIn and X. Which parts of those platforms AI crawlers may read is the platforms' decision, and it can change. We are not going to guess at it here. Your posts still matter: they build the record of what you talk about, and they give journalists, hosts and peers a reason to mention and link you elsewhere. The one source whose crawler access you decide yourself is your own website.
What can you actually control?
More than you might think, and most of it is unglamorous.
One line, everywhere. Pick a single phrase for what you do ("pricing consultant for B2B SaaS", not "growth strategist") and use it on your site, your LinkedIn headline, your X bio, directory profiles and speaker bios. An assistant that meets the same pairing of name and specialism on several independent pages has something consistent to repeat.
A site crawlers can read. Check your robots.txt. We got this wrong ourselves. On our own site, a rule meant to hide one redirect page (/b) was written as a prefix. robots.txt rules match prefixes, so from 11 August to 21 October 2026 it also closed /blog, /business and every article in our business blog to all crawlers. Bing Webmaster Tools listed our business page as "discovered but not crawled" from 11 August, and Google had none of the business pages indexed. We changed the rule to /b$ on 21 October 2026.
Know which bots you are blocking. AI companies run separate crawlers for training and for live answers. OpenAI's GPTBot collects training data, while OAI-SearchBot serves ChatGPT's search. PerplexityBot indexes, while Perplexity-User fetches pages on demand, as this robots.txt guide explains. Blocking GPTBot only keeps your content out of OpenAI's training. It does not decide whether you show up in Perplexity, Gemini or Google's AI Overviews. If you want to be recommended, do not block the search and user agents.
Enough pages to be found. Markup alone does not get you into answers. One site we track had complete markup (server rendering, Article, FAQPage and Organization JSON-LD, a sitemap) but only 75 URLs. It was still absent from answers on its own article's exact topic. A site with about 190,000 crawlable URLs was crawled constantly. You do not need thousands of pages. You do need more than an About page: one page per service, problem or case you want to be known for.
A steady public record. Posting on LinkedIn and X every week builds the body of work that others find, quote and link. If keeping it up is the hard part, see building a body of work on LinkedIn and how to post consistently as a founder. Staying on one theme helps too: a consistent thought leadership narrative makes it easier for anyone, human or model, to say what you are known for.
Why isn't ranking on Google enough?
Because assistants do not find pages the way Google does. On 19 October 2026 we asked assistants for the exact titles of 18 articles published on three client sites, each older than three weeks. ChatGPT found 1 of 18 and Claude found 7 of 18. Google had indexed 29 of 30 articles on one of those sites. A page Google has indexed may still be one an assistant cannot find.
Also, do not judge success by traffic. In about two weeks of logs from three of our course websites, human visitors arriving from AI assistants totalled one, from Copilot. ChatGPT, Perplexity, Gemini, Claude, Meta AI and others sent none. Assistants mostly answer in place. The value is being named in the answer, not the click.
Your server logs do show when an assistant reads a page while answering someone. Live retrieval agents such as ChatGPT-User, OAI-SearchBot and Perplexity-User fetch pages during real conversations. In one fortnight, a client blog of a few dozen articles received 56 such fetches across 24 different articles. Look for those user agents in your own logs.
How do you check whether assistants name you?
Measure it the way a client would experience it.
- Write 10 to 20 questions a buyer would really type, such as "who is a good fractional CFO for a Series A startup in Berlin". Avoid questions with your name in them.
- Ask each one in ChatGPT, Claude, Gemini and Perplexity, three times each. Answers vary between asks, so count yourself as named for a question if any of the three asks names you.
- Note who is named instead of you and which sites are cited. That list is your to-do list: those are the directories, publications and competitors whose pages the assistant draws on.
- Repeat every week with the same questions and the same number of asks. Compare like with like.
Doing this through the APIs is cheap. At October 2026 prices, one question asked once with web search costs us about $0.069 with Claude, $0.010 with ChatGPT, $0.009 with Gemini and $0.0013 with Perplexity. A round of 20 questions across 4 assistants with 3 asks each comes to about $5.40.
Expect to start at zero. When we ran this on our own company on 21 October 2026, with 40 questions across four assistants, three times each, we were named in 0 of 160 answers and none of our pages was cited. That is a baseline, not a verdict. For a quick first look, the free check on the ItGrows homepage needs no signup: it generates five buyer questions from a homepage and gets answers from ChatGPT and Perplexity in 25 to 40 seconds.
What should you do next?
This week: fix your one-line specialism everywhere, check robots.txt for accidental blocks, and write your 10 to 20 test questions. This month: publish one page on your site for each thing you want to be recommended for, and complete the directory profiles that come up in your test answers. After that, publish every week and re-run the same questions.
If the weekly publishing is what slips, our product drafts LinkedIn and X posts in your own voice on topics you choose. You approve or edit every post before it goes out on your schedule. If you run a company and want to know whether ChatGPT, Claude, Gemini and Perplexity name it, our business service, reachable from the same product page, measures that every week. Clients report results: FreightSizer says it won 2 new B2B customers in one month, both from ChatGPT. No one can promise an assistant will name you, and we don't.
Frequently Asked Questions
Can I pay or submit my profile to get ChatGPT to recommend me?
No. There is no form for getting into recommendations. What you can influence is what an assistant finds when it searches: your own pages, directory listings and other people's mentions of you.
Do LinkedIn and X posts help me get recommended by AI assistants?
They build the public record of what you know and give others reasons to mention and link you. What AI crawlers may read on those platforms is set by the platforms themselves. So treat your own website as the source you fully control, and use your posts to feed it and your reputation.
Should I block GPTBot on my website?
Blocking GPTBot keeps your content out of OpenAI's training data. It does not stop ChatGPT search from citing you, which uses OAI-SearchBot, and it does not affect Perplexity or Gemini. If you want to be recommended, make sure the search and user agents can read your pages.
How long does it take before an assistant names me?
There is no reliable timeline, and some people never get named on broad questions. Record a baseline, even if it is zero, and re-ask the same questions every week. Watch which sites get cited, since getting listed or mentioned on those pages is often the most direct route in.