Skip to content
Climb Google Maps: Google Business Profile optimization.

GEO and AEOGuide

How to Rank in ChatGPT: The Crawler Setting, the Business Hub Claim, and the Page Work

There is no ChatGPT ranking you can buy, but there are three things you control: whether OpenAI's search crawler can read your site, whether your business is claimed in the new Business Hub, and whether your pages give answers worth citing. We checked 53 big sites: 9 shut ChatGPT search out completely.

A hand holding a phone with the ChatGPT app open

Key takeaways

  • To rank in ChatGPT search, your robots.txt must allow OAI-SearchBot; OpenAI says sites that block it will not be shown in ChatGPT search answers, apart from navigational links.
  • Blocking GPTBot only opts a site out of model training; OpenAI treats it separately from OAI-SearchBot, so you can block training and still appear in ChatGPT search.
  • In our Nov. 11, 2026 check of 53 large US sites, 9 blocked OAI-SearchBot from their home page and 19 blocked GPTBot; 10 blocked training but left ChatGPT search open.
  • Since Nov. 10, 2026, businesses can also claim their listing in ChatGPT Business Hub, though OpenAI says claiming does not guarantee higher placement.
  • Measure ChatGPT visibility with Business Hub mentions and clicks plus chatgpt.com referral sessions in GA4, compared over the same 28 days.

Short answer: you can't pay to rank in ChatGPT's answers (ads are a separate, labeled product), and OpenAI hasn't published ranking factors. What you control is whether ChatGPT can cite you at all. That comes down to three things: let OpenAI's search crawler, OAI-SearchBot, read your site; claim your business in the new ChatGPT Business Hub; and write pages that answer questions plainly enough to quote.

The first step is the one sites get wrong. On Nov. 11, 2026 we checked the robots.txt files of 53 large US sites, and 9 of them block ChatGPT search from their home page entirely. The steps below take about 30 minutes for most sites.

Step 1: Let OAI-SearchBot crawl your site

OpenAI runs four crawlers, each with its own user agent and job, according to its crawler documentation:

User agentWhat it does
OAI-SearchBotFinds pages to show in ChatGPT search answers
ChatGPT-UserVisits a page when a user asks ChatGPT to
GPTBotCollects content that may be used to train OpenAI's models
OAI-AdsBotChecks the landing pages of ChatGPT ads

OAI-SearchBot is the one that decides search visibility. OpenAI says sites opted out of it won't be shown in ChatGPT search answers, though they can still appear as navigational links. It also says it can take "~24 hours" after a robots.txt change for its systems to adjust.

If you want to appear in ChatGPT search but keep your content out of model training, OpenAI says the settings are independent. This robots.txt does exactly that:

User-agent: OAI-SearchBot
Allow: /
 
User-agent: GPTBot
Disallow: /

Two details from OpenAI's documentation that trip people up:

  • Blocking ChatGPT-User does little. Its visits are started by a user, so OpenAI says robots.txt rules may not apply, and it isn't used to decide search eligibility. Manage search with OAI-SearchBot.
  • If you allow both OAI-SearchBot and GPTBot, OpenAI may use one crawl for both purposes.

Step 2: Check what your robots.txt actually says

Most site owners have never read their robots.txt for these names, and a plugin, a host or a security tool may have added rules. This Python script uses only the standard library. Save it as openai_bots.py and run python3 openai_bots.py https://yoursite.com/:

# Check whether a site's robots.txt lets OpenAI's crawlers reach a page.
import sys
import urllib.request
import urllib.robotparser
from urllib.parse import urljoin
 
AGENTS = {
    "OAI-SearchBot": "ChatGPT search results",
    "ChatGPT-User": "visits a user asks for",
    "GPTBot": "model training",
    "OAI-AdsBot": "ChatGPT ad landing page checks",
}
 
page = sys.argv[1]
robots_url = urljoin(page, "/robots.txt")
req = urllib.request.Request(robots_url, headers={"User-Agent": "robots-check/1.0"})
with urllib.request.urlopen(req, timeout=20) as resp:
    lines = resp.read().decode("utf-8", "replace").splitlines()
 
parser = urllib.robotparser.RobotFileParser()
parser.parse(lines)
for agent, job in AGENTS.items():
    verdict = "allowed" if parser.can_fetch(agent, page) else "BLOCKED"
    print(f"{agent:14} {verdict:8} ({job})")

On our own site it prints allowed for all four. On The New York Times' home page it prints BLOCKED for OAI-SearchBot, ChatGPT-User and GPTBot, and allowed for OAI-AdsBot. On Reuters it shows the split many publishers choose: OAI-SearchBot allowed, GPTBot blocked.

Run it on a deep page too, not only the home page: Disallow rules often cover folders such as /blog/ or /products/.

A clean robots.txt is not the whole story. A firewall, CDN bot setting or security plugin can block OpenAI's crawlers before robots.txt is ever read. OpenAI publishes OAI-SearchBot's IP ranges at openai.com/searchbot.json; if your host or CDN has a bot allowlist, add them there.

What we found on 53 large sites

On Nov. 11, 2026 we fetched the robots.txt files of 60 large US sites: 20 news publishers, 15 retailers, 15 travel, health, jobs, real estate and local sites, and 10 SEO and website platform sites. Fifty three returned a file we could read. For each of OpenAI's four crawlers we used Python's standard robots.txt parser to test whether the rules let it fetch the home page.

  • OAI-SearchBot (ChatGPT search) is blocked on 9 of 53: The New York Times, The Wall Street Journal, CNN, NBC News, USA Today, Bloomberg, the BBC, Amazon and Yelp.
  • GPTBot (training) is blocked on 19 of 53.
  • 10 sites block GPTBot but leave OAI-SearchBot open, including Reuters, AP News, eBay, Tripadvisor, Airbnb and WebMD. They opted out of training and stayed in ChatGPT search.
  • 16 sites name OAI-SearchBot in their file at all; 24 name GPTBot.

Large publishers are not typical businesses, and many of them have licensing deals or disputes that drive these choices. The useful point for everyone else is the split: blocking training and blocking search are separate decisions, and the sites that think hardest about AI crawlers mostly treat them that way.

Step 3: Claim your business in ChatGPT Business Hub

Since Nov. 10, 2026, OpenAI lets businesses claim their listing at chatgpt.com/business-hub, verify ownership (by phone, business email, a DNS TXT record or Shopify), pass an identity check with Persona, and then edit their details and see insights. Local stores and restaurants can only claim an existing US listing for now. Our news story, ChatGPT Business Hub is live, covers the steps and what each metric counts.

OpenAI says claiming doesn't guarantee higher placement or sales. Do it anyway: it is the only place to correct what ChatGPT has on file about you, and the only first-party ChatGPT data you can get.

Step 4: Write pages ChatGPT can quote

OpenAI hasn't said how ChatGPT search picks sources, so this step is our view, from how we write this site for AI answers:

  • Answer in the first two sentences. Put the direct answer under the heading that asks the question, then explain.
  • Use specifics a model can repeat safely: dates, prices, places, numbers with how they were measured.
  • Make each paragraph stand on its own. An AI answer lifts a paragraph, not a page. "It" and "this" at the start of a paragraph lose their meaning out of context.
  • Name and link your sources, and keep business facts (address, hours, services) the same on your site, your Google Business Profile and your Business Hub listing.
  • Keep the text in the HTML. If your content only appears after JavaScript runs, check that a plain fetch of the page contains it.

Step 5: Measure it

Two numbers, taken over the same 28 days, once a month:

  1. Business Hub mentions and website clicks, once your listing is verified and passes OpenAI's reporting threshold.
  2. GA4 sessions from chatgpt.com. Our guide Does Google Analytics show traffic from ChatGPT? shows how to find them and set up an AI channel.

Clicks in the hub will not equal sessions in GA4; OpenAI's own label says clicks don't confirm visits. Watch the trend in both.

What we don't know

  • How ChatGPT search ranks sources. OpenAI documents the crawlers, not the ranking.
  • Whether a claimed listing gets cited more. OpenAI says there is no guarantee, and the program is a day old.
  • What firewalls block. Our robots.txt test can't see blocks at a CDN or server, so a site that passes it can still be unreachable to OAI-SearchBot.

Sources

Photo of Kiran Dasgupta

About the author

Kiran Dasgupta, Editor

Kiran runs Kihan Marketing, a web design, SEO and ads agency, and manages dozens of websites across real estate, home services and healthcare. He writes about SEO, AI search and paid ads news, and what it means for the people who run websites.