To see where competitors get cited by AI and you do not, ask the same buying questions your customers ask ChatGPT, Perplexity and Google AI Overviews, and record which brands the answers name. Do it across your core topics, log who appears, and note the pages the AI pulls from. That gap list, the questions where a rival is named and you are absent, becomes your AI content plan.
For the classic traffic side, start with the free tools that show a competitor's website traffic.
How to spy on a competitor's AI traffic
You cannot see a competitor's analytics, but you can estimate how much traffic they get from AI tools. Check whether they appear when you ask ChatGPT, Claude, Perplexity and Google AI Overviews the questions their pages target. Use a traffic-estimation tool such as Similarweb or Semrush to see whether AI domains turn up among their estimated referral sources (these are modelled estimates, not their real analytics), review which of their pages are structured for citation (clear answers, FAQs, comparison tables), and track branded search growth as a rough proxy for AI visibility (rough, because people search your name for plenty of reasons besides an AI answer). The step-by-step method is below.
Why competitor AI visibility is the new blind spot
You can see who ranks above you in Google. What you cannot see as easily is who the AI answers name when someone asks ChatGPT, Perplexity or Google AI Overviews for a recommendation. And the AI answer increasingly sits between your customer and the click. A Pew Research Center study of Google users, based on the browsing data of 900 US adults in March 2025, found that 18% of their Google searches produced an AI summary. When a summary appeared, people clicked a normal search result in 8% of visits, against 15% when there was no summary, and they clicked a link inside the summary itself in just 1% of visits. So ranking alone gets you less than it used to. If a competitor is the brand named in that summary and you are not, you are losing customers you never even see.
How to actually check it
The simplest method costs nothing: ask the AI engines the questions your customers ask. Take your top ten buying-intent queries, run them through ChatGPT, Perplexity, Google AI Overviews and Bing Copilot, and write down which brands get named and which sources they cite. Do it for your category, your product type and your problem statements.
Keep a log, because memory is useless here. A spreadsheet with six columns is enough:
- Date
- Engine (ChatGPT, Perplexity, Google AI Overviews, Copilot)
- Exact prompt, word for word
- Brands named, in the order they appear
- Sources cited (the pages the answer links to)
- You named? Y or N
If you were an accountant, your first five prompts might be: "best accountant for freelancers in Leeds", "how to choose an accountant", "accountant vs bookkeeper", "who offers fixed fee tax returns" and "alternatives to [your biggest competitor]". Run each prompt three times on each engine, because the same question can return different brands on different runs. Answers also shift with the wording, your location, whether you're signed in and the date, so one run proves nothing; what you're recording is the share of runs in which each brand appears.
Be clear about the limits too. You can see public answers and the pages they cite. You can't see a competitor's analytics, and you can't see what anyone asked in a private chat, so treat your log as a sample of what customers are shown, not a measure of anyone's traffic. I offer these checks as an AI visibility audit service, so you don't have to run them by hand.
Reading what you find
Two things matter. First, which competitors keep coming up, because that is your real AI share of voice. Second, and more useful, which sources the AI pulls from. Across Google as a whole, the same Pew study found Wikipedia, YouTube and Reddit were the most commonly linked sources, in AI summaries and ordinary results alike. In your own niche the list will be more specific: industry roundups, directories, review sites and trade publications often turn up, but your log is what tells you which ones. Those sources are the doors. If your competitors are inside them and you are not, that is your to-do list.
Here's what a filled-in gap list looks like (an illustrative example, not real data). You run 10 prompts. Competitor A is named in 6, competitor B in 4, and you in 1. Your AI share of voice is the number of prompts where a brand appears divided by the prompts you ran, so A has 60%, B 40% and you 10%. The sources cited were two roundup articles and one directory. The action is concrete: get listed in those two roundups, claim and complete the directory profile, and publish one comparison page that answers the questions where A is named and you aren't.
How to close the gap
You do not close it by writing more blog posts. You close it by getting your brand into the sources the AI already trusts, making your own pages clean and quotable for machines, and building the kind of consistent entity signals that teach the AI who you are. It is slower than paid ads but it compounds. This is the heart of getting cited by AI. See how my AI citation service works, or talk to me about it here.
Measure your own side at the same time. Google's Analytics Help gives the GA4 route: select Reports in the left menu, then Acquisition, then Traffic acquisition. Click + Add filter at the top of the report, leave Include selected, choose Session source as the dimension, and tick any AI tools in the values list. Session source is the site a visit came from, so an AI assistant that passes a referrer appears there by name. Some AI visits arrive with no referrer and show up as direct traffic, so treat the figure as a minimum. As for timing: run your first check this week, re-run the same prompts monthly, and compare the gap lists side by side to see whether the work is moving you.
For your own side of the numbers, my analysis of three months of Copilot citations shows what Bing Webmaster Tools reveals.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
How to spy on competitors AI traffic
The method above tells you whether a rival is named. This is the page-by-page version, for when you want to know which of their pages do the work. Open the Sources cited column in your log and filter it for the competitor's domain: what you entered was your own prompt, word for word, and what came back was the answer plus the links it cited, so every row with their domain is a page an AI engine leaned on. Count how often each URL appears. Then search the same prompts in Google and note which of their pages hold page-one spots, because pages that rank for a question are often the ones AI answers draw from too. Put the two lists side by side and you have their AI traffic map: the questions they own, and the pages that answer them. Then write the better answer to each one, with the answer in the first two sentences, a comparison table where the question invites one, and an FAQ that covers the follow-up questions.
To catch mentions and trends before they pass you by, see my guide to the best social listening tools.
Frequently asked questions
Why does ChatGPT give different answers each time I ask?
Answers vary with the exact wording, the date, your location and your chat history, so one run tells you very little. Run each prompt three times on each engine and log how often each brand appears, not whether it appeared once.
Can I see a competitor's traffic from ChatGPT?
No. Only they can see it, in their own analytics. You can observe the public answers, which brands they name and which of the competitor's pages they cite, and that is enough to build a gap list.
How do I measure AI traffic to my own site?
In GA4, open Traffic acquisition and filter the session source for AI domains such as chatgpt.com and perplexity.ai. Some AI visits show as direct traffic, so treat the number as a minimum.
How do I get my business cited by AI?
Get into the sources AI already cites for your questions, make your pages clean and quotable for machines, and build consistent entity signals so the AI learns who you are. Run your first check this week, re-run it monthly, and judge progress by your share of voice in the log.
Once you can see where rivals get cited, you can close the gap with answer engine optimization services.
About Lilach Bullock
I am a marketing consultant working in digital marketing since the dial-up days, and these days I help business owners put AI to work inside their marketing and operations. I build the automations that hand people back their week. To talk it through, work with me here.
If your customers are local, the AI answers lean on local signals, so start with A Guide to a Successful Local SEO Strategy: Boost Your Local Business; for the traditional side of the traffic picture, see Proven strategies to drive traffic to your blog.
To go beyond AI answers and map the whole picture, reverse-engineer the competitor's wider marketing strategy.
Before you change anything, run the free AI Visibility Checker so you have a baseline to measure against rather than a guess.
Run a business, a brand or a blog? I feature tools, products and expert guest posts for an audience of 15,000. See how it works