- Google is bringing Gemini 3.5 Flash, AI agents, multimodal inputs and generative UI deeper into Search.
- The update signals a major shift from traditional search results toward AI-powered tasks, monitoring, booking and personalized answers.
Google is turning Search into something much bigger than a box for links.
At I/O 2026, the company announced what it called “a new era for AI Search,” introducing a new AI-powered Search box, Gemini 3.5 Flash in AI Mode, background information agents, agentic booking, generative UI and expanded Personal Intelligence.
In its official announcement, Google said it is bringing “advanced model capabilities to Search” and enabling users to “use agents just by asking a question.” The company also described the new AI-powered Search box as its biggest upgrade in more than 25 years. Google announced the Search updates in a May 19 blog post by Elizabeth Reid, VP of Search.
The message is clear: Google Search is no longer just becoming more conversational. It is becoming more agentic.
Gemini 3.5 Flash becomes the default model in AI Mode
One of the biggest changes is the model upgrade behind AI Mode.
Google said it is upgrading Search with Gemini 3.5 Flash as the new default model in AI Mode globally. The company described the model as its newest Flash model, designed for sustained frontier performance for agents and coding. According to Google’s announcement, the upgrade is starting immediately for AI Mode users worldwide.
That matters because AI Mode is becoming one of the main places where Google experiments with deeper, more complex search behavior. Instead of typing a short keyword query and scanning blue links, users can ask broader questions, add follow-ups and move through a more conversational search experience.
Google also said AI Mode has surpassed one billion monthly users one year after its debut, with queries more than doubling every quarter. That gives the company a strong argument that AI Search is not just a side experiment anymore. It is becoming a major part of how people use Google.
Google is rebuilding the Search box around AI
The most visible change is the new intelligent Search box.
Google says the Search box is being reimagined with AI and will dynamically expand so users can describe what they need in more detail. It will also provide AI-powered suggestions that go beyond traditional autocomplete. Google says the new Search box is rolling out in all countries and languages where AI Mode is available.
The bigger change is input flexibility.
Google said users will be able to search across modalities using text, images, files, videos or Chrome tabs as inputs. That turns Search into a much broader interface. A query no longer has to start as a few typed words. It can start with a file, a video, an image, an open tab or a messy description of what the user is trying to do.
For publishers, marketers and SEOs, this matters because search behavior could become less keyword-shaped. Users may describe intent in longer, more specific ways or bring external context directly into the search process.
Search agents move Google from answers to monitoring
Google is also introducing Search agents.
The first version will focus on information agents. These agents can run in the background, monitor the web and send synthesized updates when something changes. Google says information agents will launch first for Google AI Pro and Ultra subscribers this summer.
That changes the role of Search.
Traditional search is reactive. The user asks a question, Google returns results and the session ends. Information agents make Search more persistent. A user can set a monitoring task and let Google keep watching the web, news sites, blogs, social posts and real-time data sources such as finance, shopping and sports.
Google gave examples such as apartment hunting or tracking sneaker collaborations from athletes. But the broader implication is bigger: Search could become a notification and monitoring system, not just an answer engine.
For media and publishers, that could create new visibility patterns. Instead of competing only for a single search result page, content may increasingly feed agentic summaries, alerts and ongoing research tasks.
Google Search can now act on local tasks
Google is also expanding agentic booking in Search.
The company said users will be able to ask for local experiences and services with detailed criteria, such as finding a private karaoke room for six people on a Friday night that serves food late. Search will then bring together pricing and availability with links to complete the booking through the provider. Google says these agentic booking capabilities are expanding to a wider range of tasks.
For selected categories such as home repair, beauty and pet care, Google said users will also be able to ask Search to call businesses on their behalf. Those capabilities are expected to roll out to everyone in the U.S. this summer.
This is one of the clearest signs that Google wants Search to move beyond finding information. Search is becoming a task layer. It can gather options, compare availability and in some cases take action.
For local businesses, that could make structured information even more important. Hours, services, availability, pricing and booking paths may become part of how AI agents decide which businesses are useful enough to surface.
Search is getting generative UI and mini apps
Another major update is generative UI inside Search.
Google said it is bringing the power of Google Antigravity and Gemini 3.5 Flash’s agentic coding capabilities into Search. This will allow Search to build custom layouts, visual tools, simulations, tables and graphs in real time based on the user’s question. Google says these generative UI capabilities will be available for everyone in Search this summer.
The company is also planning custom dashboards and trackers, which it described as mini apps for ongoing tasks. For example, a user could ask Search to build a fitness tracker using fresh sources such as reviews, live maps, local data and weather.
This could become one of the more important changes for the open web.
If Google can generate custom interfaces directly inside Search, users may spend more time interacting with Google-built tools instead of clicking through to external sites. At the same time, those tools still need data, context and supporting sources. That means publishers may still matter, but the way their content is used could continue shifting away from simple referral traffic.
Personal Intelligence expands in AI Mode
Google is also expanding Personal Intelligence in AI Mode.
The company said users in nearly 200 countries and territories across 98 languages can now connect apps such as Gmail and Google Photos, with Google Calendar support coming soon. Google said this does not require a subscription and emphasized that users can choose if and when they connect their apps. Google described Personal Intelligence as built around transparency, choice and control.
This makes Search more personalized, but also more sensitive.
If Search can understand a user’s private context, it can answer more specific questions and help with personal tasks. But it also means Google is bringing more private app data into the Search experience, even if users have to opt in.
For users, the benefit is convenience. For Google, the strategic value is clear: Search becomes more deeply connected to the user’s daily life.
What this means for SEO and publishers
The I/O 2026 Search announcements point to a bigger shift in how visibility may work.
Search is becoming more conversational, more multimodal and more action-oriented. Users can bring files, videos and tabs into Search. They can ask follow-up questions from AI Overviews. They can create agents that monitor the web. They can ask Search to build custom interfaces or help complete tasks.
For publishers and SEOs, that means the old model of optimizing only for a ranked result page looks increasingly incomplete.
Content will still need to be crawlable, useful and trustworthy. But it may also need to be easier for AI systems to interpret, summarize and use inside generated answers, monitoring agents and custom interfaces.
That puts more pressure on clear structure, strong entity signals, fresh information, original reporting, source-backed claims and content that answers specific user needs. If Search agents are scanning blogs, news sites and social posts for updates, publishers may need to think about whether their content is easy for those agents to understand and trust.
Google Search is becoming an AI operating layer
The biggest takeaway from Google’s I/O 2026 Search update is that Search is becoming less like a destination for results and more like an operating layer for tasks.
The Search box is becoming multimodal. AI Mode is getting a stronger model. Agents can monitor information in the background. Search can help book services, call businesses, generate interfaces and connect to personal data.
That is a major change in how Google wants users to interact with the web.
For users, it means Search may feel more like an assistant.
For businesses, it means visibility may depend not only on ranking but on being useful to AI systems that answer, summarize, monitor and act.
For publishers, the challenge is sharper than ever: Google still needs the web, but more of the user experience is happening inside Google.
The next era of Search may not be about getting users to type better keywords.
It may be about getting Google’s AI agents to understand why your information deserves to be used.
