regularly, you've earned backlinks, and your traffic holds steady. But

lately you've noticed something unsettling: when you ask ChatGPT a

question that your site answers perfectly, it gives the answer without

mentioning you at all. Meanwhile, it cites a Forbes article from 2022

and a Reddit thread.

So does ChatGPT even know your site exists? And if it does, why isn't

it citing you?

This guide explains exactly how ChatGPT citation works, how to find out

whether your site is being cited right now, and what the difference is

between a brand mention and an actual citation. No nonsense — just the

answers site owners need.

First: ChatGPT Operates in Two Completely Different Modes

Before you check anything, you need to understand one critical

distinction that most guides skip over.

Mode 1: Training data only. When ChatGPT answers from its training

data — everything it learned before its knowledge cutoff — it does not

fetch live websites. Any "sources" it mentions in this mode may be

fabricated. Real links to real content are not being pulled. This is

where hallucinated citations come from.

Mode 2: Web browsing / Search mode. When ChatGPT searches the web to

answer a question, it sends a crawler to fetch current pages and can

cite them with real links. This is where actual citations happen. You

can tell which mode is active by looking for a "Searched the web"

indicator or a Sources panel in the response.

If you are trying to get cited, Mode 2 is the only one that counts.

Everything else is the AI paraphrasing things it learned months or

years ago.

How to Check If ChatGPT Is Citing Your Site Right Now?

There are three methods, ranging from free-and-manual to automated.

Method 1: Ask ChatGPT directly (free, takes 10 minutes)

Open ChatGPT — make sure web search is enabled, which it is by default

in most current plans — and run prompts that reflect how real users ask

about your topic.

Be specific. Don't search your brand name, search the problem your

site solves.

For example, if you run a food blog focused on batch cooking:

• "What are the best batch cooking methods for beginners?"

• "How do I meal prep a full week of lunches?"

• "Which sites have good guides on batch cooking for families?"

After each response, check two things. First, does a Sources panel

appear? If no sources are shown, ChatGPT answered from training data

and no live citation occurred. Second, if sources do appear, is your

domain listed?

Run this in incognito mode so previous chat history does not influence

the results.

Method 2: Check your server logs for AI crawler visits

This is the most reliable method, and it is the approach tools like

Unsourced are built around.

When ChatGPT cites a page, it first sends a crawler to fetch that page.

That visit appears in your server logs as a request from a user agent

containing ChatGPT-User. If you see those requests, ChatGPT is actively

reading your content — which is a strong signal that citations are

happening or are imminent.

To check manually in your server logs, look for lines containing:

ChatGPT-User

Other AI crawlers to watch for:

• GPTBot — OpenAI's training crawler (different from citation crawling)

• PerplexityBot — Perplexity AI

• ClaudeBot — Anthropic's Claude

• Google-Extended — Google's AI products

The distinction matters: GPTBot visits are about future training data.

ChatGPT-User visits are happening because someone asked ChatGPT a

question that triggered a live lookup of your page. That is a citation

event.

If you are on a paid WordPress tier or a standard hosting setup, most control panels

give you access to raw access logs. If you use a managed host, you may

need to look in the hosting dashboard under "raw logs" or "error logs."

Method 3: Use a monitoring tool

Manual log-checking does not scale. If you want ongoing visibility

rather than a one-off check, several tools now track AI citations

automatically.

The category is young but growing fast. Tools in this space approach

the problem differently — some scan AI outputs for brand mentions,

others monitor at the crawler level like Unsourced does. Both angles

give you useful information, but crawler-level monitoring tells you

what is actually happening to your content before it ever appears in

an AI answer.

The Difference Between a Brand Mention and a Citation

These sound similar but they are not the same thing, and conflating

them leads to wasted effort.

A brand mention is when ChatGPT refers to your company or site by name

in its answer — "According to [YourSite], the best approach is..." —

without necessarily linking to a specific page. This happens more often

than hard citations and is still valuable for brand awareness, but the

user may not click anywhere.

A citation is when ChatGPT links to a specific page on your site in

the Sources panel. The user can see it, click it, and you may get a

direct visit. This is harder to earn and more valuable.

A training data reference is when the model paraphrases your content

without acknowledging you at all. This happens constantly. You will

never know it occurred unless the wording is distinctive enough to

recognise.

When people ask "is ChatGPT citing my site?", they usually mean hard

citations with links. That is the hardest to achieve and the most

worth tracking.

Why Your Site Might Not Be Getting Cited Even If It Ranks on Google

Google and ChatGPT use fundamentally different signals to decide what

to surface. Ranking well in organic search does not translate

automatically into AI citations.

A few common reasons sites that rank well on Google still get skipped

by ChatGPT:

Your content does not answer the question directly enough. ChatGPT

favours pages that front-load a clear, direct answer. Pages that bury

the answer after several paragraphs of context are at a disadvantage

— the AI may move on to a page that gets to the point immediately.

You are not ranking on Bing. ChatGPT's web search relies primarily on

Bing's index, not Google's. A site can rank well on Google and still

have poor Bing visibility, which limits how often ChatGPT finds it in

the first place.

Your domain authority is low relative to cited competitors. AI citation

is heavily concentrated among high-authority domains. Independent sites

can still get cited, but they are competing against default sources like

Wikipedia, Reddit, and major media brands that AI models return to

repeatedly.

Your content is not structured for AI extraction. Dense paragraphs are

harder for AI to parse cleanly than headers, short answers, and lists.

Pages with FAQ schema see meaningfully higher citation rates than those

without.

A Practical Starting Point

If you want to know right now whether ChatGPT is paying attention to

your site, do this:

1. Open your server logs and search for ChatGPT-User. If you see

visits, your content is being read.

1. Run five topic-level prompts in ChatGPT with web search active.

Record whether sources appear and whether yours is among them.

1. Set up ongoing monitoring — either through a tool like Unsourced

or by checking your logs weekly — so you are not relying on

one-off snapshots.

The sites that win AI citation over the next 12 to 18 months will not

necessarily be the ones with the highest Google rankings. They will be

the ones that understood the difference early, tracked it properly, and

structured their content to answer questions directly.

That starts with knowing where you actually stand today.

Unsourced monitors AI bot traffic to your website in real time, so you

can see exactly which pages are being read by ChatGPT, Perplexity, and

other AI crawlers — and when. Start monitoring your site free for 14

days at unsourced.app.