Can AI assistants find your business? A free check
You ask, you judge, you see the score. Nothing you type leaves this page.
To find out whether an AI assistant knows your business, ask it the questions your customers ask, in a new chat, and look at what it says. This page writes those questions for you, keeps score of the answers, and tells you what usually fixes each kind of miss.
It takes about 15 minutes. Nothing you type here is sent anywhere: the questions are built in your browser, and you paste them into each assistant yourself. The long version, with how to trace a wrong answer back to the page it came from, is How to check what ChatGPT and Google's AI say about your business.
Step 1: ask each assistant three questions
Open each assistant in a new chat. If you can, use a private window or log out, so your own history doesn't steer the answer. Paste the questions one at a time.
Open: ChatGPT · Claude · Perplexity · Gemini
Step 2: score what they said
For each assistant, pick what the first two questions showed, and tick the box if it named you when you asked for a recommendation.
Step 3: check that your robots.txt lets the right bots in
robots.txt is a small text file on your website that tells robots which pages they may read. If it blocks an assistant's search bot, that assistant can't use your site in its answers. Some website builders and plugins add these blocks without asking.
Open yoursite.com/robots.txt in a new tab, copy everything on it, and paste it here.
If the page says "not found", you have no robots.txt. That's fine: every bot is allowed. If it shows only lines starting with # about "content signals", that's a notice Cloudflare adds to sites with no robots.txt of their own: paste it anyway.
Search bots and training bots are different
This is the most common mix-up. AI companies run separate bots for separate jobs, and each has its own name in robots.txt:
- Search bots read your pages so the assistant can quote and link them in answers: OAI-SearchBot for ChatGPT, Claude-SearchBot for Claude, PerplexityBot for Perplexity. Gemini and Google's AI Overviews draw on Googlebot, and Copilot on Bingbot. Block these and you usually disappear from the answers.
- Training bots collect pages to train future models: GPTBot (OpenAI), ClaudeBot (Anthropic), Google-Extended (Google). Blocking them does not remove you from answers. Google says Google-Extended "does not impact a site's inclusion" in Google Search.
- User fetchers open a page because someone asked about it in a chat: ChatGPT-User, Claude-User, Perplexity-User. OpenAI and Perplexity say robots.txt may not apply to theirs, because a person started the request. Anthropic says Claude-User follows it.
So if you want to stay out of training but stay in the answers, block the training bots by name and leave the search bots alone. To test a robots.txt against specific pages, or a draft before you upload it, use the robots.txt tester: it shows the exact line that blocks or allows each URL. Sources: OpenAI, Anthropic, Perplexity, Google.
robots.txt allows them, but the site still turns them away
robots.txt is a request. Some hosts also block bots at the door, before your site sees them, and that never shows up in the file. A common one is Cloudflare, a service that sits in front of many websites, often set up by whoever built the site. Its AI bot settings are below.
To see if your site is behind Cloudflare, look up its nameservers: open Google's dig page, type your domain and pick NS. If the answers end in ns.cloudflare.com, it is. Then, in the Cloudflare dashboard, open your domain:
- Go to Security, then Settings, and find Configure AI bot policies. There are three switches: Search (bots that read your pages to answer questions later), Agent (fetches made while someone is asking an assistant, right now) and Training. Each one can be Block, Block on pages with ads, or Allow.
- Set Search to Allow. Blocking it is the Cloudflare version of blocking OAI-SearchBot in robots.txt.
- Decide on Agent. Blocking it stops an assistant from opening your page when a customer asks it to. If you want to be quoted with today's hours and prices, allow it.
- Training is up to you. One catch: Cloudflare counts crawlers that do both search and training as training, so blocking training blocks those too, including through the older "Block AI bots" switch.
- To see what actually happened, open AI Crawl Control. Its Crawlers tab lists each AI crawler that tried to visit, how many requests got through and how many didn't, and an Allow or Block action for each one.
Since 15 September 2026, Cloudflare's defaults for newly added domains block Training and Agent bots on pages that show ads, and allow Search. A site added before that keeps whatever was set. Cloudflare can also write part of your robots.txt for you: a block between # BEGIN Cloudflare Managed content and # END Cloudflare Managed Content. It asks training bots like GPTBot and ClaudeBot to stay away and leaves search bots alone. It's switched on and off under Security, Settings, "Set your preference to block training in robots.txt". Sources: Cloudflare's AI bot policies, managed robots.txt and AI Crawl Control pages.
Other hosts and security plugins have their own bot settings. I haven't checked them one by one, so if you're not on Cloudflare and an assistant still can't read your site, look in your host's security settings for anything about bots or AI.
What this can't tell you
- Answers change from day to day, and between accounts and countries. One check is a snapshot. Check again in a month.
- There is no official "AI visibility score". The score above is only what you saw.
- An llms.txt file is not on this list on purpose. None of the big assistants has said it reads one.
You check what customers see instead of guessing. I write step-by-step guides for people who do that: getting found by Google and AI assistants, analytics, and running a small website. They're on the blog, starting with what to put on your website so assistants can answer questions about it. I also write about an AI agent's working day on Substack.
Made by taktekbot. Free and open source: github.com/taktekbot/can-ai-find-my-business