Google's Own Data Just Ended the AI Visibility Debate

Google just published the data that settles it. The average AI Mode query is now triple the length of a traditional search. Follow-up queries are growing 40% per month. The top five words users open with are "what," "how," "I," "is," and "can." That third one is the one that should keep you up tonight. Users are not searching. They are talking. And most brands are still optimized for a user who no longer exists.

The "Fool's Gold" Argument Deserves a Fair Hearing

A MediaPost column this week asked whether AI search visibility is "fool's gold or the future." The author made honest points. Measurement is expensive. Citation consistency looks low. Most AI searches are informational, not transactional. And Rand Fishkin's SparkToro data from early 2026 showed less than 1 in 100 consistency in which brands get cited for the same prompt.

I have spent eight years building the measurement infrastructure for this exact problem. I am not going to pretend those concerns are frivolous. They are real. But they are real about the wrong question.

The question is not whether AI visibility is worth the investment. The question is whether the old thing you are investing in still has a user on the other end.

Google's Data Says the User Already Left

Google's report, "How People Are Using AI Mode in the U.S.", covers the period from AI Mode's U.S. launch in May 2025 through April 2026. The numbers are not ambiguous.

Queries are 3x longer. Not slightly longer. Triple. That means the user who typed "best CRM software" is now typing something closer to "I run a 40-person B2B SaaS company and our sales team hates our CRM, what actually works for teams that sell to enterprise." SearchEngineJournal's analysis put it plainly: that is not a keyword. That is a person talking to someone who might actually help them.

Planning queries have grown 80% faster than overall AI Mode growth. Queries starting with "which" grew 40% faster over the past six months. One in six AI Mode searches now uses voice, image, or video input instead of typed text.

The person typing three words and scanning ten blue links? That person is a minority now. And the content built for that person is decaying whether anyone calls it fool's gold or not.

The Consistency Problem Is a Measurement Problem

The SparkToro finding on citation inconsistency is real data. But it measures something specific: whether the same brand appears when the same prompt is repeated verbatim. In a system where the query is now a paragraph of personal context, repeating the same prompt verbatim is not how users behave. They go deeper. They follow up. They add context about their own situation.

Consistency measured against a static prompt tells you about the model. It does not tell you about the user. The user is doing something different every time, and the brands that show up are the ones whose content answers the shape of the question, not the exact string.

In September 2025, Seer Interactive published the largest publicly disclosed study on this shift: 3,119 queries across 42 organizations and 25.1 million organic impressions. Organic click-through rates fell. That is the old metric dying. The new metric is whether your brand is the one the AI pulls from when a user describes their actual problem in their own words.

What the Content Gap Actually Looks Like

A study of 1 million keywords found that 29% of high-volume keywords are declining while long-tail, intent-rich queries are growing. That is not a traffic collapse. It is a migration. The searches did not disappear. They changed shape.

Most content strategies have not caught up. They are still built for someone who types "running shoes flat feet" and gets ten links. The actual user now types "I have flat feet and my knees hurt after long runs, can you help me find a shoe that will not make it worse." The health and wellness example in Google's own report is even more direct: "I hate cardio. Give me a routine that avoids it but still works."

That query has no keyword. It has a person, a constraint, a goal, and an emotional charge. The brand that answers it is the one the AI cites. The brand optimized for "best no-cardio workout" might show up. Or it might not. Because the AI is not matching keywords. It is matching usefulness.

The Move: Stop Optimizing for Keywords, Start Being Citable

Here is what I tell every founder who asks me whether AI visibility is worth it.

Stop asking whether it is worth it. Ask whether anyone can find you when they describe their problem the way they actually experience it.

Three things to do this week:

1. Run your buyer's actual question through ChatGPT, Perplexity, and Google AI Mode. Not your brand name. Not your target keyword. The messy, specific, emotional version of the question your buyer would actually ask a smart friend. See who shows up. If it is not you, that is the gap.

2. Audit your top 10 pages for answer density. Does each page contain a direct, extractable answer to a specific question a real person would ask? Not a keyword cluster. A question with a name, a context, and a constraint. If the answer is buried in paragraph six behind a wall of introductory throat-clearing, the AI will skip you and cite whoever got to the point faster.

3. Measure what the AI actually cites, not what ranks. The Machine Relations Index tracks citation rates across ChatGPT, Perplexity, Claude, Gemini, Google AI Mode, and AI Overviews. The data shows which sources get cited, how often, and with what confidence. That is the metric that matters now. Not keyword position. Not impressions. Citation rate by segment.

The Debate Is Already Over

The MediaPost author said AI visibility might be a "brand investment, not a performance channel." That framing assumes the old performance channel still performs. Google's data says the user who made it perform has changed how they search. The 3x query length is not a trend to watch. It is a migration that already happened.

I built AuthorityTech because I saw this shift before the industry named it. I built the measurement infrastructure because I knew "is it worth it" would be the wrong question. The right question was always: when users stop typing keywords and start describing their lives, will the AI know your brand has the answer?

That question has a binary outcome. You are in the answer, or you are not. There is no middle ground in a conversation the AI is having on your buyer's behalf.

FAQ

How long are AI Mode queries compared to traditional search?

Google's data shows AI Mode queries are now triple the length of traditional search queries. Users are typing full sentences with personal context, constraints, and goals rather than two or three keyword fragments. Follow-up queries within a session are growing 40% per month, meaning users stay in the conversation rather than bouncing to a results page.

Is AI search visibility worth the investment for smaller brands?

The question is misdirected. Every brand's buyers are already using AI search, and those queries are getting longer, more personal, and more intent-rich every month. The cost of not being citable is not $0. It is invisibility in a channel your buyers are actively using. The Machine Relations Index measures citation rates across six AI engines so brands can see exactly where they appear and where they do not.