The Short Answer
If your customers research before they buy, AI is almost certainly intercepting some of them. 58% of consumers now use AI to research major purchases, up from 41% a year earlier (Invoca, B2C Buyer Experience Report, 2026). The good news: you can find out in about ten seconds, without any tools or logins.
The 10-second test
Open ChatGPT, Gemini, or Perplexity and ask the question a real buyer would ask before spending money with you:
“Who are the best companies for [your category]?”
Read the answer. If it names three or four competitors and your brand isn’t one of them, you have AI leakage. That ready-to-buy customer just got handed to whoever the AI trusts, and right now that isn’t you.
Five signs you have an AI leakage problem
- Your brand isn’t named when you ask AI for the best options in your category.
- New-customer growth is slipping with no obvious reason, even though your ads and rankings look the same.
- Competitors you used to beat are suddenly “everywhere.”
- Your cost per acquisition is creeping up while conversion quality drops.
- You’ve never checked how you appear inside AI tools, so you genuinely don’t know.
Why your analytics won’t catch it
Traditional analytics measure the traffic that reaches you. AI leakage is about the traffic that never does, the buyers who got their answer inside the AI and never clicked through to compare. There’s no line item in your dashboard called “customers AI sent to a competitor,” so the loss hides in plain sight as a vague, frustrating decline.
What to do next
First, get a number. Our estimator turns your revenue and customer profile into a directional figure for what AI leakage is costing you each year. Then, once you have the number, the path is to get into the AI ads and the AI answers before your category catches on.
For the mechanics behind why AI names some brands in that test and not others, see our complete guide to AI visibility.
Run your revenue through our estimator to find out what AI leakage is costing you, then see if you qualify to have us plug the leak.
Frequently Asked Questions
What exactly should I type when I run the test?
Ask what a real buyer would ask before spending money, not what you would ask about your own company. “Who are the best companies for [your category]” works. So does naming the specific problem you solve rather than your industry label. Avoid searching your own brand name, that tells you the AI can find you, not whether it recommends you.
Should I run the test logged in or in a fresh session?
A fresh session with no personalization. If you are logged into an account that has discussed your company before, the model may surface you because of your own history rather than because it genuinely regards you as a leading option. Use a private window or a signed-out session so you see what a stranger sees.
What if the AI mentions my brand but gets the details wrong?
Treat that as a version of the same problem. An old complaint, a discontinued product, outdated pricing, or a former executive presented as current fact will cost you the sale as surely as being left out, sometimes faster. A human researcher would contextualize stale information. A machine may hand it back flatly as though it were true today.
Do the different AI platforms give different answers?
Often, yes, which is why checking only one is misleading. ChatGPT, Gemini, and Perplexity draw on different sources and weight them differently, so you can be named consistently in one and absent from another. Run the same prompts across all three before drawing a conclusion about where you stand.
How often should I re-check?
Quarterly is enough for most brands, unless you are actively working on visibility, in which case monthly shows you whether the work is moving. What matters more than frequency is consistency: keep the same prompts each time so you are comparing like with like rather than reacting to a differently worded question.