Most businesses have no idea whether AI recommends them. They assume they’re visible because they have a website and a Google listing — and then they’re surprised when an assistant names three competitors and leaves them out. The only way to know is to measure it, and measuring it properly is more involved than typing your own name into ChatGPT once. Here’s how to do it right.

Start with the questions a customer would actually ask

Measurement begins with the right questions. Not “tell me about [your company]” — a customer who already knows your name isn’t the one you’re worried about. You want the questions of a customer who doesn’t know you yet: “who are the best [your service] in Bahrain,” “I need a [service] near [area], any recommendations,” “which [industry] firm should I use for [specific need].” These open, intent-driven questions are where recommendations are won or lost.

Ask across every platform, more than once

A single check is almost worthless, for two reasons. First, the platforms disagree — you might be named by one and absent from another, so testing only your favourite tells you little. Second, answers vary between attempts and shift over time as content changes. Reliable measurement means asking the same battery of questions across ChatGPT, Gemini, Claude and Perplexity, and repeating it, so you’re reading a pattern rather than a one-off.

Know what you’re actually measuring

“Am I mentioned?” is only the first of several things worth tracking. A complete picture looks at how often you appear across all those questions (your presence or mention rate); where you land when you do appear, since being named first is not the same as being named fifth (your position); how you’re described, because being mentioned with a weak or inaccurate caveat is different from a confident recommendation (your sentiment); and, on the platforms that cite sources, which pages the AI is drawing on (your citations). Together these turn a vague feeling of “are we visible?” into numbers you can actually move.

Turn the numbers into action

Once you can see where you’re weak, the improvements follow logically. Low presence usually points to thin or unclear content about your field — the model doesn’t have enough to go on. Poor position often reflects weaker authority than the competitors being named ahead of you. Inaccurate descriptions usually trace back to inconsistent information scattered across the web. And the pages showing up in citations tell you exactly what’s working, so you can build on it. You then re-measure the next month and watch whether the changes landed — because visibility work is a loop, not a one-off project.

The limits of doing it yourself

You can do a basic version of this on your own, and it’s worth doing — even ten honest questions across the four platforms will tell you more than you know today. The limits show up at scale: covering enough questions to be statistically meaningful, staying consistent month after month, and reading the results without wishful thinking are hard to sustain alongside running a business. That’s the gap a proper audit fills.

How Visiblify measures it

Our audit runs hundreds of real customer questions across the four major AI platforms, across the categories your customers actually search, and reports exactly where you’re named, where you’re missing, how you’re described, and who’s being recommended instead. Every month we show you the before-and-after, with the raw results and the sources, so progress is something you can see rather than something you’re asked to trust.

Want to know your real AI visibility score today? We’ll run your first audit free