Why is AI search activity nearly invisible in Google Ads search terms reports?

AI search activity is nearly invisible in Google Ads search terms reports. An analysis of more than 400 Google Ads accounts, published by Search Engine Land on November 9, 2026, found no clear shift toward the longer queries Google says AI Mode generates, in impressions or in conversions, across the US, UK and France.

The most likely reason is not that AI search ads are absent but that the report cannot show them. Google's own help pages say that for searches in AI Mode and AI Overviews, "the search term shown in your reporting represents the best approximation of the user's intent," and that terms without enough query activity are omitted. This guide explains what the study found, what you can and cannot see, and how to adjust bidding and reporting.

What did the study find?

Rémi Kerhoas, paid media and data director at the agency Eskimoz, analyzed impression and conversion data from more than 400 Google Ads accounts with an MCC-level script. The dataset covered Search and Performance Max campaigns from the first quarter of 2024 to the third quarter of 2026. He grouped search terms by word count (one word, two to three, four to seven, eight or more) and estimated unreported volume by subtracting reported search term totals from ad group or keyword totals. The author discloses his role at the agency, and the data comes from its client base.

The three markets were chosen because AI search arrived at different times:

Market	AI Overviews launch	AI Mode launch	Ads in AI Overviews	Ads in AI Mode
US	May 2024	May 2025	Mobile from late 2024, desktop from May 2025	Test from May 2025
UK	August 2024	July 28, 2025	Not available yet	Planned for Q4 2026
France	July 22, 2026	July 22, 2026	Not available yet	Not available yet

Timelines as published in the study, November 9, 2026.

If AI search ads were leaving a trace, the US accounts should show it most clearly. They did not. The main findings:

• No shift toward longer queries in US Search or Performance Max impressions, despite Google's statement that AI Mode queries are three times longer than traditional searches.
• No shift in conversions either, which rules out the idea that a small volume of high-converting AI queries is hiding in the totals.
• Broad match long tail is thin. Very few reported broad match terms are long, while a large share of terms is unreported, and that share "has remained fairly flat for close to three years."
• AI Max terms are short. Most contain one to three words, overlapping heavily with what regular keywords already capture.
• UK and France look the same as the US, even though their AI search rollouts were far more recent.

Why don't AI searches show up in the report?

The study offers three explanations and rates its own confidence in each:

Explanation	Author's confidence	Reasoning
Commercial prompts are shorter than the AI Mode average	Low	Google's "three times longer" is an average across all use; the agency's accounts skew toward performance goals.
AI ads are a small share of ad delivery	Moderate	Google has said AI Mode monetizes less than traditional search, and AI placements need broad match, AI Max or Performance Max.
The report does not show prompts as typed	High	Google reports an approximation of intent for AI Mode and AI Overviews and omits low-volume terms.

The third explanation is backed by Google's own documentation. On its page about ad group prioritization, Google says searches on Lens, AI Mode, AI Overviews or autocomplete can involve "more complex search journeys," and in those cases the reported term "represents the best approximation of the user's intent." On the search terms report page, Google says some terms "that don't have enough query activity are omitted" for privacy reasons, and that search terms insights group them into subthemes or "other queries."

Long, conversational prompts are exactly the kind of one-off queries least likely to clear a volume threshold, and an approximated term will often look like an ordinary short keyword. So an AI Mode conversation about "the best CRM for a five person agency that also handles invoicing" may surface, if at all, as something close to "crm for agencies."

Context matters too. In Alphabet's third quarter 2026 earnings call, Chief Business Officer Philipp Schindler said lower monetization in AI Mode had weighed on Search revenue growth, according to Search Engine Land's coverage. And on November 9 Search Engine Roundtable reported that ads on follow-up queries in AI Mode are now live, which adds another layer of conversation context that a single search term cannot capture.

What can advertisers see, and what can't they?

You can see	You cannot see
Reported search terms, which for AI Mode and AI Overviews are approximations of intent	The prompt as the user typed or spoke it
Search terms insights grouped into themes, plus an "other queries" bucket	The individual low-volume queries inside those buckets
Campaign, ad group and keyword performance, including conversions and value	A standard breakdown of performance by AI Overviews, AI Mode and classic results, according to the study's author
Totals that let you estimate the unreported share	The earlier turns of an AI Mode conversation that led to a follow-up ad

The practical consequence is that the search terms report has moved from a near-complete record of demand to a sample. As the author puts it, paid search is starting to look more like paid social, where you steer with signals and outcomes rather than with exact queries.

How should you adjust bidding?

Optimize on outcomes, not on query text

If the query you see is an approximation, bidding on what you see is bidding on a summary. Google's match type documentation says "it's critical to use Smart Bidding with broad match" because each query has different auction-time signals. Whether or not you trust broad match, the reliable lever is the conversion data you feed the system.

Fix your conversion data first

• Count only conversions that matter to the business. A newsletter signup and a booked demo should not carry the same weight.
• Pass conversion values, and import offline outcomes such as qualified leads or closed deals, so value-based bidding optimizes toward revenue.
• Use your margins and customer lifetime value to set targets. Our guide to calculating CAC and LTV shows how.

Control spend with structure and exclusions

• Keep exact or phrase match campaigns for your proven, high-intent terms so you retain a clean, readable baseline.
• Test broad match or AI Max in a separate campaign with its own budget, so you can compare results instead of blending them.
• Add negative keywords at the theme level when search terms insights show irrelevant subthemes, since you will not see every individual query.

Match landing pages to longer intents

AI-assisted searches often carry more context, such as team size, budget or a specific use case. A landing page that only repeats your headline keyword wastes that intent. Our guide to landing pages for paid ads covers message match.

How should you change your reporting?

1. Track the unreported share. Subtract reported search term impressions, clicks and conversions from campaign or ad group totals each month. A rising unreported share tells you how much of your spend you can no longer inspect.
2. Report on outcomes, with query data as color. Lead client and leadership reports with conversions, value, cost per acquisition and return on ad spend. Use search terms and insights to explain, not to prove.
3. Annotate rollout dates. Mark AI Overviews, AI Mode and ads launches for each market in your reporting tool, so changes in performance can be read against them.
4. Use controlled tests. Campaign experiments, geo splits or time-boxed holdouts tell you more about broad match or AI Max than reading terms ever will. That matters more now that multiturn ads in AI Mode can serve on follow-up questions, which only broad match, AI Max, Shopping and Performance Max can reach contextually.
5. Watch organic and AI visibility alongside paid. If AI answers cite competitors for your category, your paid search data will not explain it. We cover the organic side in our Google I/O 2026 AI Mode guide.

What are the limits of the study?

It is one agency's data, and the author notes that most clients focus on performance goals, so upper-funnel queries, where much AI Overviews and AI Mode inventory may sit, are underrepresented. Performance Max search terms only became available in the second quarter of 2025, which makes earlier PMax data less reliable, and multi-country campaigns were excluded. The finding is best read as "you cannot see AI search in this report," not as "AI search ads are not happening."

For a broader view of where AI-driven demand is going, compare this with our analysis of ChatGPT ads, and see how we approach measurement on our expertise page and in our case studies.

Sources

• Search Engine Land, "AI search activity is nearly invisible in Google Ads search terms reports: Study", Rémi Kerhoas, November 9, 2026, accessed November 10, 2026. searchengineland.com
• Google Ads Help, "About ad group and asset group prioritization within a Google Ads account", accessed November 10, 2026. support.google.com
• Google Ads Help, "About the search terms report", accessed November 10, 2026. support.google.com
• Google Ads Help, "About keyword matching options", accessed November 10, 2026. support.google.com
• Search Engine Roundtable, "Google AI Mode Multiturn Ads Are Live", November 9, 2026, accessed November 10, 2026. seroundtable.com

Frequently asked questions