# How Do B2B Companies Get Named When Buyers Ask ChatGPT?

> B2B companies get named by ChatGPT when crawlers can reach specific pages answering buyers' niche questions. Then measure weekly, since nobody can guarantee it.

Published: 2026-10-22 · Updated: 2026-10-23
Source: https://www.itgrows.ai/business/blog/ai-visibility-for-b2b-companies
Publisher: ItGrows (https://www.itgrows.ai/business) — AI visibility for businesses

---

A B2B company gets named by ChatGPT when the assistant can find a specific page on that company's own site that answers the narrow question a buyer is asking, and when crawlers are actually allowed to reach that page. After that you check every week, with the same questions asked several times, whether the assistant names you, because nobody can guarantee that it will.

The rest of this article covers why niche questions are where smaller B2B companies have a real chance, what one client's reported result looks like, which pages tend to work, the technical mistakes that quietly block you, and how to measure all of it.

## Why do niche B2B questions give smaller companies a chance?

Broad questions are often answered without any search at all. In a test on 23 September 2026 with 20 German business questions, the model answered 10 without searching and cited nobody. Where it did search, it cited primary sources such as laws, ministries, the OECD and large accounting firms on 9 of 10. Firms were named only on "who should I hire" or "how do I choose" questions, and two runs of the same question named two different sets of firms.

That tells you where B2B visibility actually happens. A buyer asking "what is freight dimensioning" gets a textbook answer. A buyer asking which tool measures parcel and pallet dimensions for a warehouse, integrates with their existing software and fits a mid-sized operation is asking a selection question. The assistant has to search, and very few pages on the web answer that exact combination of constraints.

The narrower the question, the fewer pages compete for it. A small vendor with a precise page can be one of the few credible answers, where on a broad question it would be buried under encyclopedias and large publishers. The research supports the idea that content choices matter: the [Princeton paper that introduced Generative Engine Optimization](https://collaborate.princeton.edu/en/publications/geo-generative-engine-optimization/) reported that specific content optimization methods can raise a source's visibility in generative engine responses by up to 40% on its benchmark of queries.

## What happened with FreightSizer's two customers from ChatGPT?

FreightSizer, a B2B SaaS for freight dimensioning and one of our clients, reported in October 2026 that it got 2 new B2B customers in one month, both from ChatGPT. That figure comes from the client, not from our logs, and it is not a promise of what any other company will see.

Two customers may sound small. Put it next to what we see on other sites. In about two weeks of logs from three of our course websites, roughly 1.2 million requests in the two busiest logs, human visitors arriving from AI assistants totalled one from Copilot and zero from ChatGPT, Perplexity, Gemini, Claude, Meta AI and others. Assistants mostly answer in place rather than send a click. We go into this in [Does AI Search Send Traffic to Websites?](https://www.itgrows.ai/business/blog/does-ai-search-send-traffic-to-websites).

For a consumer site, near-zero clicks is a problem. For B2B, the arithmetic is different: when one customer is worth a contract rather than a single purchase, a handful of buyers who were told your name by an assistant can matter more than a large volume of anonymous visits. Other clients report similar patterns at their own scale: appslift, an app store optimisation service, reports 2 to 5 leads a month from AI assistants, and MagiScan, an iOS 3D-scanning app, reports hundreds of installs a month from ChatGPT, Claude and Gemini.

The practical lesson: judge B2B AI visibility by named answers and by leads that say "ChatGPT recommended you", not by referral traffic in analytics.

## Which pages actually get a B2B company named?

Specific technical pages, and a lot of them. Two observations from our own work point the same way.

First, surface beats markup. A site with complete markup (server rendering, Article, FAQPage and Organization JSON-LD, a sitemap) but only 75 URLs was still absent from answers on its own article's exact topic. A site with about 190,000 crawlable URLs was crawled constantly. Clean code does not compensate for having little to find.

Second, being indexed by Google is not the same as being findable by an assistant. On 19 October 2026 we asked assistants for the exact titles of 18 articles, each older than three weeks, published on three client sites. ChatGPT found 1 of 18; Claude found 7 of 18. Google had indexed 29 of 30 articles of one of those sites.

When pages are found, they are fetched one question at a time. Live retrieval agents such as ChatGPT-User, OAI-SearchBot and Perplexity-User fetch a page while an assistant is answering a real person. In one fortnight, a client blog of a few dozen articles received 56 such fetches spread across 24 different articles. That spread is the point: each narrow page has its own chance to be pulled into an answer.

For a B2B company, the pages that tend to answer buyer questions look like this:

| Page type | Buyer question it answers | What it needs to contain |
| --- | --- | --- |
| Use-case page | "Which tool does X for a company like mine?" | The industry, company size and workflow it fits, stated plainly |
| Integration page | "Does it work with the software we already use?" | Named systems, how the connection works, its limits |
| Specification page | "Can it handle our volumes or formats?" | Concrete capabilities, constraints and requirements |
| Comparison or how-to-choose page | "How do I choose between options?" | Honest criteria, where you fit and where you do not |
| Pricing-model page | "What does this kind of tool cost?" | How pricing works, even if exact prices depend on scope |

Structured data still has a role. The [Schema.org vocabulary](https://schema.org/docs/schemas.html) currently lists 826 types, including Organization, which B2B companies use to describe their identity. It was [founded in 2011 by Google, Microsoft, Yahoo and Yandex](https://en.wikipedia.org/wiki/Schema.org) and can be embedded as JSON-LD. Use it to make your company unambiguous, but treat it as a supplement to pages that answer questions, not a substitute.

## What technical mistakes keep B2B sites out of answers?

The most common one is blocking crawlers without knowing it. We did it ourselves. On itgrows.ai, a robots.txt rule meant to hide a single redirect page (/b) was written as a prefix. Because robots.txt rules match prefixes, 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 2026, and Google had none of the business pages indexed. We changed the rule to /b$ on 21 October 2026.

Check your robots.txt line by line, and check it per crawler. [OpenAI documents](https://developers.openai.com/api/docs/bots) that GPTBot and OAI-SearchBot are separate user agents controlled independently, so a site can allow OAI-SearchBot to appear in ChatGPT search results while disallowing GPTBot to opt out of training data collection. A blanket block on "AI bots" can remove you from search answers you wanted to be in. The details are in [How Do I Allow GPTBot and Other AI Crawlers in robots.txt?](https://www.itgrows.ai/business/blog/how-to-allow-gptbot-in-robots-txt).

Also look at who is really visiting. Across the full retained nginx logs of an edtech platform we work with (pickaclass.com), 28 July to 11 August 2026, ClaudeBot made 19,272,125 requests, Meta-ExternalAgent 3,854,076, GPTBot 2,744,985, Applebot 2,655,555, Amazonbot 1,170,268 and Googlebot 52,292. If your firewall or CDN treats that volume as an attack, it may be blocking the very crawlers you need. Submitting sitemaps to Bing and Google and pinging IndexNow when pages change helps new pages get discovered.

## How do you measure whether ChatGPT names your company?

Write down 20 or so questions your buyers actually ask, weighted towards selection questions, and put them to ChatGPT, Claude, Gemini and Perplexity every week. Ask each question three times, because answers vary: asked the same question three times, an assistant often names different businesses. We count a business as named for a question if any of the three asks names it, and we compare weeks only on the same questions and the same number of asks. More on method in [How Do You Measure AI Visibility Over Time?](https://www.itgrows.ai/business/blog/how-to-measure-ai-visibility).

Record the cited sites too, not just the names. In our 21 October 2026 measurement of questions about getting into AI answers, the most cited sources were the blogs of GEO agencies and tool vendors themselves, such as firstpagesage.com and llmpulse.ai. In other words, in a B2B category the vendors who publish detailed answers on their own domains become the sources. That is the position you want in yours.

Expect to start at zero. In that same measurement, ItGrows was named in 0 of 160 answers and none of our pages was cited. Movement shows up as single events first: on 20 October 2026 Perplexity cited an article we had published on a client's site, a German company-formation advisory. If competitors appear and you do not, [Why Doesn't ChatGPT Mention My Company but Names My Competitors?](https://www.itgrows.ai/business/blog/why-doesnt-chatgpt-mention-my-company) walks through the usual causes.

Measurement is cheap. With API prices of October 2026, 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 report of 20 questions across four assistants with three asks costs about $5.40.

## What should a B2B company do first?

Start with the free check at [https://www.itgrows.ai/](https://www.itgrows.ai/). It generates five buyer questions from your homepage and shows in 25 to 40 seconds whether ChatGPT and Perplexity name you.

Then work in this order: fix robots.txt and firewall rules so OAI-SearchBot, Claude's and Perplexity's crawlers can reach you; submit sitemaps to Bing and Google and use IndexNow; publish one specific page per buyer question on your own domain; and measure weekly on a fixed question set. If you would rather not run this yourself, ItGrows does those steps for $499 a month (plus a one-time $499 setup, free with 3 months or a year), and reports every week which answers name you and which sites they cite. No one can guarantee an assistant will name you, and we say so up front.

## Frequently Asked Questions

**Do B2B buyers really ask ChatGPT for vendor recommendations?**
Some do, and the clearest signal is reported leads. FreightSizer reported 2 new B2B customers in one month in October 2026, both from ChatGPT, and appslift reports 2 to 5 leads a month from AI assistants. Ask new leads how they found you, since assistant referrals rarely show up as clicks in analytics.

**Is good schema markup enough to get named?**
No. A site we observed with complete JSON-LD, server rendering and a sitemap but only 75 URLs was still absent from answers on its own topic. Markup helps identify your company; pages that answer specific buyer questions are what assistants retrieve.

**Why does ChatGPT name my company in one answer and not the next?**
Answers vary between runs of the same question, so a single check proves little. Ask each question three times, count yourself as named if any ask names you, and compare weeks only on the same questions and number of asks.

**How long until a B2B company shows up in AI answers?**
There is no fixed timeline, and pages indexed by Google can still be invisible to assistants: in our 19 October 2026 test ChatGPT found only 1 of 18 client articles by exact title. Fix crawler access first, publish specific pages, then watch weekly measurements for the first named answers or citations.