Artificial intelligence (AI) enables new ways for users to explore information, discover products, and interact with brands. User behavior is evolving, with people increasingly asking longer, more complex, and more conversational questions, similar to how users actually think and speak.
Search is also becoming multimodal, with users using images and audio in addition to text.
The longer and more complex your queries become, the harder it becomes to match them with relevant ads. For example, queries submitted through AI mode tend to be three times longer than traditional search queries, said Dan Taylor, Google’s vice president of global advertising.
Google introduced AI modes that offer advanced reasoning capabilities, deeper multimodal capabilities, and the ability to engage in follow-up questions, allowing users to explore topics in a more conversational and iterative way.
It builds on AI Summary, which generates a concise AI-generated summary at the top of search results, providing users with quick answers that link to deeper sources on the web.
Gemini 3, Google’s most advanced AI model to date, addresses this problem by using the same large-scale language model to generate answers that more accurately interpret a user’s underlying intent, Taylor said. This improved understanding allows Google to show you more personally relevant ads that closely match what you’re actually looking for.
“These types of long searches have traditionally been much more difficult to monetize and identify the right ads to display, but now we can better understand intent and match the right ads with the same large-scale language model in Gemini that helps give consumers the right answers,” he explained.
“With Gemini, we are able to better understand user intent for these complex queries and display personally relevant ads that meet their needs in the moment. As our models improve and our ability to employ them improves, the quality of our ads continues to improve. By integrating Gemini into our ad quality stack, we have reduced the percentage of irrelevant ads by an average of 40%, which increases value for consumers and advertisers.”
The company started with an overview of AI and integrated search with AI to respond to changing consumer behavior. Launched in 2024, AI Overview generates a concise, AI-generated summary at the top of search results, providing users with quick answers that link to deeper sources on the web. Google places ads above, below, and within these responses.
“People find ads in AI Overview helpful because they can quickly connect to relevant businesses, products, and services to take the next step when they need to. AI Overview not only matches ads to what the person is searching for, but also to the context of the generated AI Overview snapshot,” Taylor said.
He said this opens up new discovery opportunities for advertisers and allows brands to appear in contexts that were previously invisible. AI Overview is used by over 2 billion users monthly in over 200 countries and territories and supports over 40 languages.
Is there a clear distinction between advertising and organic search results? “The search team has complete autonomy over the results that are displayed and/or the results that are displayed. And the search team’s primary north star, and the company’s north star, is to focus on providing the best possible results for the user in terms of information that is accurate and reliable,” said Taylor.
“And if there’s an advertising opportunity that also serves as an answer, we’ll provide it. There’s no opportunity for the advertising business to influence organic results at all. And that’s true even with these new AI-powered experiences.”
Google is also experimenting with new ad formats that work well in this environment. One early example is the direct offer. This is a pilot within an AI mode that shows sponsored deals to high-intent shoppers at the moment they appear ready to make a purchase.
“Going forward, we will continue to experiment with both the right moments in the consumer journey in AI mode where the commercial opportunity makes the most sense for advertisers and is additive to the user experience, and different types of formats that we feel are appropriate to drive both advertiser value and user value,” Taylor said.
The next frontier in AI is agents, he said. Agent tools help businesses drive better results and get things done with less effort.
“Our goal with agent commerce is twofold: first, to take the drudgery out of shopping so consumers can focus on the fun part, and second, to work collaboratively with industry and retail on the building blocks needed to make agent commerce seamless and secure across the web,” Taylor said.
