Counting the fish that don't rise: measuring B2B visibility in AI search

Google is putting brands into AI answers and declining to say whether anyone clicks. The practical move for B2B teams is to stop waiting for a dashboard Google may never build, and to measure AI visibility further down the funnel where buyers actually surface. Three developments in the past week make the case, and together they point the same direction.

What changed this week

• October 29: SEO analyst Brodie Clark reported that Google appears to be testing AI Overviews without the citation cards that list the source sites behind an answer.
• October 30: Cyrus Shepard ran a one-site test. When he opted that site out of AI Overviews and AI Mode, its regular organic clicks fell too. He called the sample small, and it is, but the direction is a warning.
• November 2: Google's Gary Illyes said the team wants to share more data but has "nothing to announce yet," which lowers any expectation of clean AI click data arriving in Search Console soon.

Read together: the citations that would tell you that you were cited may be disappearing from view, opting out appears to cost ordinary traffic, and the official measurement may never come.

Opting out is the wrong protest

The instinct to pull out of AI Overviews in protest is understandable and, on the evidence so far, self-defeating. Shepard's test is a single site and he says so plainly, but it suggests the opt-out takes regular organic clicks down with it rather than protecting them. Withholding your content from the surface where buyers increasingly start does not punish Google. It removes you from the shortlist.

Measure downstream instead

If the click number is never going to be clean, the answer is to measure where AI influence actually shows up for a B2B buyer.

• Add a free-text "How did you hear about us?" field to every lead form, and actually read the answers. Buyers will tell you when an assistant named you.
• Run a fixed panel of 20 to 30 real buyer questions through ChatGPT, Gemini, Perplexity, and Google's AI Mode every month, and log who gets named in the shortlist.
• Watch branded search and direct traffic, the lagging signals where an AI mention tends to surface days or weeks later.

Measuring AI visibility by clicks alone is like counting the fish by the ones that happen to rise. Most of the activity never breaks the surface, and the number you can see is not the number that matters.

Build the content that gets cited

Downstream measurement only pays off if there is something worth citing. The pages that earn a mention tend to share three things: specific, checkable claims rather than superlatives; honest comparison pages that an assistant can quote; and third-party review coverage the engine already trusts. This is the same discipline behind building a site that is easy to cite and easy to verify, now applied to how you measure it.

We help B2B teams build and measure for the agent-then-human buyer, so AI visibility becomes a signal you track rather than a dashboard you wait on. The question to sit with: what signal are you trusting right now?

Sources: Search Engine Roundtable, Google may be testing removal of AI Overviews citation cards; Search Engine Roundtable, AI Overviews and AI Mode opt-out CTR study; Search Engine Roundtable, click data and Search Console AI reporting.