Google Ads

AI Search Is Nearly Invisible in Google Ads Search Terms Reports

A study of more than 400 Google Ads accounts found almost no sign of AI Mode's longer queries in search terms reports. The analysis, by Rémi Kerhoas of Eskimoz, looked for a shift toward longer queries in the US, UK, and France, and found no clear change in any of the three markets. That is a puzzle, because Google says AI Mode queries are three times longer than traditional searches. If those long queries were flowing through Google Ads, the search terms reports should show it, and they do not.

What the study found

The study examined search terms reports across more than 400 Google Ads accounts, a sample large enough to reveal a real trend if one existed. The test was simple: if AI Mode is sending advertisers longer, more conversational queries, then the average query length in search terms reports should be rising, and long queries should be appearing more often. The analysis found no clear shift toward longer queries, not in the US, not in the UK, and not in France.

That result is striking because it cuts against the most repeated claim about AI search, that people now ask much longer questions. Google says AI Mode queries are three times longer than traditional searches, so a visible change in query length would be the natural fingerprint of AI Mode traffic. The absence of that fingerprint suggests either that AI Mode traffic is still a small share of what advertisers see, or that the reporting is not showing it in the form advertisers expect.

Why the terms may not show up

One explanation lies in how Google reports AI traffic. For AI Mode and AI Overviews, Google says the reported term is its best approximation of the user's intent, not a literal transcript of what the user typed. That means a long, conversational question may arrive in the report as a short, familiar-looking keyword, stripped of the wording that would reveal its origin. That suggests the longer queries are being summarised away before advertisers ever see them.

There is also a volume explanation. AI Mode may still account for a small share of the clicks in most accounts, too small to move the average query length across hundreds of accounts. The practical takeaway is that absence of evidence is not proof of absence: the traffic may be growing without yet being large enough to bend the averages, and the study captures a moment in time rather than a final verdict.

What advertisers should do about it

The first step is to stop expecting the search terms report to reveal AI Mode on its own. If Google reports an approximation of intent rather than the real query, then query length is the wrong signal to watch. The practical takeaway is to look for other signs instead: changes in conversion rate, shifts in impression share, and new audience behaviour that cannot be explained by classic search patterns.

The second step is to segment and compare. Advertisers should track performance by campaign type and by placement where the reporting allows it, and watch for segments whose behaviour has changed without any change in their own settings. The practical takeaway is to treat unexplained performance shifts as possible AI traffic, and to test that theory with controlled experiments rather than waiting for the reporting to confirm it.

The third step is to keep the keyword strategy flexible. If AI Mode queries are being reported as approximations, then exact-match thinking may matter less than broad coverage of intent. The practical takeaway is to build campaigns around themes and outcomes, not around a list of literal phrases, so that the campaign still works when the reported query is a summary rather than a transcript.

Why it matters for SEOs and link builders

For advertisers, the study is a warning about trusting a single report. The search terms report has always been the window into what users actually typed, and if AI search blurs that window, advertisers need new ways to understand demand. The brands that adapt their measurement now will be ahead when AI traffic becomes too large to ignore.

For SEOs and link builders, the same invisibility applies to organic reporting. If AI Mode queries are summarised before they reach advertisers, they are likely summarised before they reach site owners too, which makes it harder to see which conversational queries drive visits. The practical takeaway is that content should be built around the questions people ask in full sentences, even when the reporting can only show the shortened version of those questions.

Sources

Sources: Search Engine Land, November 10, 2026. Analysis of 400+ Google Ads accounts by Rémi Kerhoas, Eskimoz.

Related reading on this site: Bing Tests Scan-to-Call QR Codes in Local Packs, Perplexity Cuts Agent API Costs and Adds an Ultra Effort Level, Google Will Switch Search Off by Default for K-12 Student Accounts, Amazon v. Perplexity Returns to Court on November 20, Google's November 2026 Core Update Is Rolling Out, and Google Gives Performance Max Page Feeds a Makeover.

More from Linkbaits

Turn this playbook into links

Linkbaits helps you create content that earns citations and backlinks. Start with the toolkit and build your first linkable asset this week.

Start free trial