In May I wrote this new search user was trending and that Google’s product announcements were a distraction. Google has now provided us with a year’s worth of data that confirms the behavioral change and puts hard numbers behind it. In the report “How people are using AI mode in the United States» (published on May 19, 2026 on The Keyword blog), Shivani Mohanvice president of data science and UXR at Google Search, describes a searcher who no longer exists in the persona that most SEO teams have built their 2025 strategies around.
According to the report, the average length of an AI query is now three times that of a traditional search query.
This single number invalidates a significant portion of what most SEO teams optimized for last summer. A year ago, the working assumption for most keyword strategies was that users would enter three to four words and then analyze the results. Google’s own data indicates that this assumption now describes a minority of what AI Mode users actually do.
User has moved, content has not been
The report covers the period from the launch of AI Mode in the United States in May 2025 until April 2026. A few other figures complete the picture. AI tracking queries have increased by over 40% on average per month, meaning users aren’t arriving at one answer and leaving; they stay in the conversation and go deeper. Multimodal interactions now account for more than one in six AI searches, meaning voice, image or video input rather than typed text, and image input searches have increased more than 40% month over month since launch.
The top five keywords in AI searches are 1. Information, 2. Identify, 3. Find, 4. Explain, 5. Summarize. The first five opening words are “what,” “how,” “I,” “is,” and “can.” Looking at the third entry of “I”, people tell their personal context in the search bar. No “running shoes for flat feet.” Something closer to “I have flat feet and my knees hurt, can you help me find a running shoe that won’t make it worse?” The health and wellness example given in the report is more direct than that: “I hate cardio. Give me a routine that avoids it but still works.”
It’s not a keyword. This is a person talking to someone who could actually help them.
What the content gap looks like
The report organizes AI mode behavior into five categories: Explore, Decide, Learn, Create, and Do. Brainstorming related queries grew 30% faster than the overall pace of AI queries. Scheduling queries grew 80% faster. Queries starting with “who” grew 40% faster over the past six months, suggesting that AI has become a true decision-making tool for everyday purchases, not just a discovery layer.
This is the gap that most content strategies have not filled. Content designed for a user who types in (best running shoes 2025) and lands on a list doesn’t serve a user who asks, “I’m training for my first 5K and I’ve never bought running shoes before, which pair should I start with and how do I know if they fit?” Both queries express an intention to purchase shoes. Only one of them describes what the AI mode user actually does.
The practical problem is that most teams still write for shorter queries. They optimize page titles, meta descriptions, and H2 structures for three-to-four-word keyword targets that represent a less and less important part of how people actually arrive at answers.
3 things to do differently now
Check your first 10 pages against how someone would actually ask for this information in a conversation. Take the main keyword from each page and rewrite it as a natural language prompt the way a Google AI mode user would actually type it. If your content doesn’t answer the longer version of this question, it has a gap that a competitor who does will eventually fill.
Treat follow-up questions as a content signal, not an analysis footnote. The 40% monthly growth in follow-up queries tells you that users aren’t satisfied with just one answer. If you know what the most common entry point questions are on your site, the follow-up question is now as strategically important as the entry point. For most sites, this inventory of follow-up questions does not yet exist.
Start preparing your visual assets for multimodal indexing. One in six AI queries is already non-text, and image-based search is the fastest growing query type in the system. Alt text written for accessibility and alt text written to serve a user who has photographed a product and asks AI mode what it is and where to buy it are different things. The context of the image around your product and your information resources must match that where the queries are already located.
Google now has more than 1 billion monthly active users on AI Mode globally, and the platform’s query volume has doubled every quarter since its launch. The behavioral change I talked about in May is no longer a prediction. This is a dataset. The question for practitioners is not whether they should answer it, but how quickly they can close the gap between the content they published last summer and the user currently searching.
More resources
Featured image: Anton Vierietin/Shutterstock
