BriefGoogle
Alphabet says AI Mode queries run three times longer than search
Alphabet says AI Mode queries run three times longer than searches, so pages that answer a buyer's full question gain and short-keyword pages lose.
TL;DR
- Alphabet CEO Sundar Pichai said on 28 October 2026 that queries in Google AI Mode are on average three times longer than traditional searches.
- Brands whose pages answer a buyer’s full question in its own words gain in Google AI Mode, and pages built around short keywords lose.
- Marketing leads should list buyers’ long questions, give each a page that opens with the answer, and track those questions in Google AI Mode.
What changed
Google AI Mode queries are on average three times longer than traditional searches, Sundar Pichai, CEO of Google and Alphabet, said in his remarks on the Q3 2026 earnings call on 28 October 2026. He added that “one in five AI Mode searches globally use voice, images, or documents.” Pichai said Google is building one experience across AI Overviews, AI Mode and traditional Search.
Pichai said Alphabet’s AI investments are driving overall Search query growth, including commercial queries, and that the year-over-year query growth rate is the highest in the last five years. He also said Google has passed 50 billion visual searches per month and that the Gemini app is at one billion users.
Alphabet’s Q3 2026 results release reports 12% growth in Google Search & other revenue for the quarter ended 30 September 2026, to $63,350 million from $56,567 million a year earlier. Neither the release nor Pichai’s Q3 remarks gives a user count for AI Mode or AI Overviews, or a separate ad revenue figure for AI features.
Google repeated the length figure the same day in its post on ads in AI Mode, which brings those ads to 10 more countries, including Australia, Canada and India. The post says sponsored and organic product list items already get similar click-through rates in AI Mode. Its footnotes date the length figure to a US report of May 2026 and the click-through data to US internal data from 20 August to 3 September 2026.
Why it happened
Google built AI Mode for complex, multi-part questions from the day it introduced AI Mode as a Labs experiment on 5 March 2025. AI Mode uses a query fan-out technique, which runs several related searches at once across subtopics. When Google expanded AI Mode to over 200 countries and territories on 7 October 2025, it reported people asking questions nearly three times as long as traditional searches.
On 20 May 2026, Google said it was testing two new ad formats in AI Mode. A Conversational Discovery ad answers a person’s specific question, and Highlighted Answers let eligible ads appear in AI Mode lists of recommendations. On the Q2 2026 earnings call on 22 July 2026, Alphabet CEO Sundar Pichai said AI Mode had passed 1 billion monthly active users since its global expansion in October 2025. He added that AI Mode was increasing overall Search queries.
What it changes for GEO
Longer questions in Google AI Mode favor brands whose pages answer one buyer’s situation in the buyer’s own words, and they cost brands whose pages were built around short keywords. A question three times the length of a search describes what the buyer needs solved, and AI Mode answers it with a recommended solution and the business behind it. That recommendation is the most valuable moment in the purchase, so a brand missing from it loses the revenue.
Google says AI Mode turns each question into several related searches through query fan-out. Those searches are worded differently from the buyer’s question and can switch language, so a page wins by matching the searches the model runs.
The pages that get picked are shaped like the question: comparisons, “best X for Y” pages and pages whose title and opening carry what the buyer asked. A “who can do this for me” question goes to a service page written in the buyer’s words.
E-commerce brands feel the paid side first, because Google says sponsored and organic product list items in AI Mode already get similar click-through rates. A shop missing from the organic list leaves that click to an ad.
Each engine answers from its own sources, so a brand’s standing in ChatGPT says little about its standing in AI Mode. ChatGPT, Google AI Overviews and Perplexity share only 2 of their top 10 cited sources when answering the same questions about the same brand, according to Agenzy’s 2026 source study (n=109 business projects).
Visibility in AI Mode is an intermediate number and revenue is the result. Buyers who meet a brand in an answer often return later through a Google search or directly, so analytics undercount AI. Brands that ask every new customer how they found them count that revenue.
What to do
The head of marketing should first collect the full questions buyers ask before they buy, word for word, from sales calls, support emails and site search. Questions asked by voice or about a photo or a document count too.
The content team then gives each question one page whose title and first lines answer it: a comparison, a “best X for Y” page or a service page. Google is building AI Overviews, AI Mode and traditional Search into one experience, so the groundwork for getting pages cited in Google AI Overviews applies to these pages too.
To check the result, the team tracks that question set engine by engine, AI Mode included, with the methods for measuring how often AI answers mention a brand. Sales adds “How did you find us?” to every new customer’s first call, because that answer counts the revenue analytics miss.
Sources
- blog.google/company-news/…/alphabet-earnings-q3-2026
- s206.q4cdn.com/479360582/…/2026q3-alphabet-earnings-release.pdf
- blog.google/products/…/ads-in-ai-mode-global
- blog.google/products-and-platforms/…/ai-mode-search
- blog.google/products-and-platforms/…/ai-mode-expands-languages-locations
- blog.google/company-news/…/alphabet-earnings-q2-2026
- blog.google/products/…/google-marketing-live-search-ads
- agenzy.lt/blog/which-sources-ai-engines-cite-2026
The week in AI search, in one email
One email a week with what changed in ChatGPT, Google AI Mode, Perplexity and Gemini, and the week's data piece.