One of Your Engines Runs on Bing

Marketers spent fifteen years learning to ignore Bing. It was the search engine your uncle used by accident. You optimized for Google, and Bing came along for the ride or it didn’t, and either way nobody cared.

Then it turned out that when a certain very popular AI assistant runs a search to answer a question, it queries Bing. Not Google. Bing. Every retrieval, on Bing’s index, with no Google fallback.

This is the detail that breaks the phrase “AI visibility.” There is no single AI visibility, the way there was more or less a single Google ranking. Each engine draws from a different well. One assistant retrieves through Bing. Google’s own AI surfaces run on Google’s index. Others lean heavily on training data and barely retrieve live at all. A brand can be prominent in one and nearly absent from another — not because it’s better or worse, but because it lives well in one index and poorly in the next.

That’s a strange and slightly liberating fact. Strange, because it means the tidy dashboard showing “your AI visibility” as one score is hiding real variance underneath. Liberating, because it means Bing — the index everyone deprioritized — is now a lever almost nobody is pulling. For one of the biggest assistants, only the top two or three Bing results per query make it into the set the model reads from. If you’re not in that handful, you’re not in the answer. And most of your competitors aren’t watching Bing at all.

None of this means abandon Google. Google still drives the open web, and the two indexes tend to reward similar things, so strong Google performance usually earns decent Bing performance too. Usually. Not guaranteed. The point is subtler: you can no longer treat “search visibility” as one number and assume it propagates everywhere. The engines have forked.

The practical move is to stop asking “am I visible in AI?” and start asking “am I visible in each engine my buyers actually use?” — and to check them separately, because they can and do disagree. A win in one is not a win in all. A blind spot in one is invisible if you only watch the others.

The uncomfortable truth is that most teams are optimizing for an average that doesn’t exist. Real buyers use specific tools, and each tool reads from a specific place.

Beket.ai tracks you across each engine separately — ChatGPT, Gemini, Perplexity and the rest — so you see the variance instead of an average that hides it.