Someone asks ChatGPT "what's the best tool for X" where X is your category. It names three products. You're not one of them.

You will never see this happen. No referral, no session, no line in your dashboard. The buyer never reaches you, so there is nothing to attribute. That's what makes it different from every other traffic problem — it's silent.

Here's what actually drives whether a model names you:

Models cite sources, they don't rank sites. SEO was about outranking. This is about being present in the handful of places the model pulls from when it answers your category — the comparison roundups, the "alternatives to" pages, the Reddit thread that keeps getting quoted. Being #1 on Google for your own brand name does nothing here.

Your competitors' listings are doing the work. The reason they get named is usually not that their product is better. It's that they're on eleven listicles and you're on two. That's a fixable problem, and it's mostly email.

One prompt tells you nothing. Model answers vary run to run. If you ask once, get named, and conclude you're fine, you've learned nothing. Twenty runs across four models gives you a mention rate. That number is stable enough to act on.

Generic prompts hide the truth. "Best CRM" is not what your buyer types. "CRM for a solo consultant that isn't Salesforce" is. You can be invisible on the first and dominant on the second — and only the specific ones map to revenue.

If you want to check yours, the free scorecard runs your category across all four models and shows your mention rate plus who's showing up instead of you. Takes a few minutes and costs nothing.