Google just changed what “search” means. At I/O 2026, Google announced that AI Mode is now powered by Gemini 3.5 Flash and that Search itself is becoming agentic, with always-on information agents, agentic booking, generative UI, and a fully reimagined Search box. For marketers, SEO leads, and agencies, this is the biggest shift in how people find information in over twenty years.
If your visibility strategy still ends at “rank on page one,” it is already behind. Here is exactly what changed, what it means, and how to build for it.
Core Concepts
- AI Mode in Google Search now runs on Gemini 3.5 Flash, Google’s strongest agentic and coding model to date.
- Search has a new intelligent Search box that anticipates intent, accepts text, images, files, videos, and Chrome tabs, and dynamically expands as you describe what you need.
- Information agents work 24/7 in the background to monitor the web and surface synthesized updates tailored to your specific question.
- Agentic booking is expanding to local experiences and services, including Google calling businesses on your behalf.
- Search now generates custom UI on the fly using Google Antigravity, including interactive visuals, simulations, and reusable mini apps.
- Personal Intelligence in AI Mode is rolling out to nearly 200 countries and 98 languages, connecting Gmail, Google Photos, and soon Calendar.
Who does this apply to
Local business owners, eCommerce brands, B2B and SaaS marketers, agencies, SEO professionals, content strategists, and anyone whose pipeline depends on being found in Google. If your audience uses Search, AI Mode, or any Gemini-powered surface, this affects how they discover you, how they evaluate you, and how they convert.
What Google Actually Announced at I/O 2026
The announcements split into two connected stories: a new model, and a new Search.
On the model side, Google introduced Gemini 3.5, a new family of models built for agentic workflows. The first release is Gemini 3.5 Flash, available today in the Gemini app, AI Mode, Google Antigravity, the Gemini API in AI Studio and Android Studio, and Gemini Enterprise. Google says 3.5 Flash delivers frontier-level intelligence at roughly four times the output speed of other frontier models, and it now outperforms Gemini 3.1 Pro on coding and agent benchmarks like Terminal-Bench 2.1, GDPval-AA, and MCP Atlas.
On the Search side, AI Mode is now the default experience for everyone globally, and it is powered by 3.5 Flash. Google also reported that AI Mode has crossed one billion monthly users, with queries more than doubling every quarter since launch, and overall Search queries hitting an all-time high last quarter. The era of “AI is killing Search” is officially over. AI is expanding Search, and Google just built the operating system for it.
Agentic Search, Explained Simply
For years, Search has been a one-shot transaction: you type a query, Google returns ten links, you click. Agentic Search flips that model. Now Search can:
- Understand intent across modalities. You can type, paste a screenshot, drop in a file, share a video, or send a Chrome tab as input. The intelligent Search box expands as you describe what you need and offers AI-powered suggestions that go beyond autocomplete.
- Run jobs in the background. Information agents continuously scan the open web plus Google’s fresh data (finance, shopping, sports, real-time inventory) for changes related to your specific question, and ping you the moment something matches.
- Take action on your behalf. Agentic booking now covers local experiences and services. For categories like home repair, beauty, and pet care, Google will literally call businesses on your behalf to confirm availability and pricing.
- Build custom interfaces. Powered by Google Antigravity, Search assembles generative UI in real time, including interactive visuals, comparison tables, simulations, and reusable mini apps you can return to and update.
- Use your personal context. With Personal Intelligence, AI Mode can securely tap Gmail, Google Photos, and soon Calendar to give you answers that already know your travel dates, receipts, and history.
That is not a search engine anymore. That is a research analyst, a personal assistant, and a custom dashboard builder, all routed through the same box you used to type “weather near me” into.
Why Gemini 3.5 Flash Is the Engine That Makes This Work
The headline metric here is not raw IQ, it is the ratio of capability to speed and cost. Google positioned 3.5 Flash in the top-right quadrant of the Artificial Analysis index because it is delivering flagship-tier performance at Flash-tier latency, often at less than half the cost of competing frontier models.
For agentic workloads, that math matters more than benchmark trivia. Agents make many small model calls in sequence. A model that is 4x faster and roughly half the cost compounds across a long-horizon task into a workflow that finishes in hours instead of days. Google’s enterprise partners are already proving it: Shopify is running parallel subagents for merchant growth forecasts, Macquarie Bank is reasoning over 100+ page customer onboarding documents, Salesforce is wiring 3.5 Flash into Agentforce for multi-turn enterprise tool calls, Ramp is using it for OCR over messy invoices, and Xero is running multi-week 1099 prep workflows for small businesses.
The takeaway for marketers: the same engine handling those enterprise workloads is now the default model behind every AI Mode answer in Search. Your customers are asking better questions, getting deeper answers, and being shown fewer ten-blue-link SERPs.
How Search Visibility Changes Under Agentic Search
Three things are shifting at once.
First, the ranking unit is no longer a page. It is a citation in a synthesized answer. AI Mode pulls from blogs, news sites, social posts, and Google’s fresh structured data to assemble a response. If your content is not the source that gets quoted or linked under that answer, you do not exist in that conversation.
Second, the query is no longer a keyword. It is an intent paragraph. Users are dumping their full requirements into the new Search box (“a private karaoke room for six on a Friday night that serves food late”), and Search is matching that against structured availability, pricing, and reviews. Long-tail, problem-statement content is going to outperform short keyword-stuffed pages.
Third, the click is no longer the only conversion. With agentic booking and Google calling businesses on your behalf, the conversion can happen inside Search. That means your Google Business Profile, schema markup, real-time availability, and partner integrations matter as much as your website copy.
If you want the technical playbook for how this connects to MCP, agents, and AI-friendly content, my breakdown on what MCP is and why marketers need to understand it is the best starting point.
Agentic Coding in Search: Custom UI on Demand
The most underrated announcement was generative UI inside Search. Antigravity-powered Search (see Google’s developer highlights from I/O 2026) can now build custom layouts on the fly: interactive visuals to explain a concept, comparison tables for a buying decision, simulations to walk through a process, or full mini apps for ongoing tasks like wedding planning, home moves, or fitness routines.
Think about what this means for content marketers. The traditional advantage of writing a long, comprehensive guide was that Google could not generate that guide on the fly. Now it can. Generic explainer content is about to get out-competed by Google’s own real-time generative UI. The content that survives will be the content with information gain that Google cannot synthesize from public data: original case studies, proprietary frameworks, lived experience, named clients, and verified results.
How This Changes the Marketing Playbook
The shift from keyword Search to agentic Search rewards a very specific set of behaviors. Here is what changes, and what stays the same.
| Marketing Lever | Pre-Agentic Search (2020-2025) | Agentic Search Era (2026 and beyond) |
|---|---|---|
| Query format | Short keywords (“plumber near me”) | Intent paragraphs and multimodal inputs (text, image, file, video) |
| Goal of ranking | Earn the click to your site | Earn the citation inside an AI Mode answer |
| Content style | Long-form keyword-optimized pages | Inverted pyramid, question-header match, information gain blocks |
| Structured data | Nice to have for rich snippets | Required so agents can read pricing, availability, hours, FAQs |
| Conversion path | Click to landing page to form | Agent-assisted booking, calls, or transactions inside Search |
| Competitive moat | Better SEO and more backlinks | First-party data, named case studies, real expertise (E-E-A-T) |
| Discovery surface | Ten blue links and a map pack | AI Mode, generative UI, information agents, Personal Intelligence |
What Marketers Should Do This Quarter
You do not need to rebuild your stack overnight, but there are five moves worth making before agentic Search behavior fully normalizes:
- Audit your structured data. Service schema, FAQ schema, LocalBusiness schema, and Article schema with proper @id linking are how agents will read your business. If it is not in your markup, it is not in the answer.
- Rewrite your top ten pages for citation, not clicks. Use the inverted pyramid (direct answer in the first 1-2 sentences), question-header match on H2/H3 tags, and information-gain blocks (original data, anecdotes, comparison tables) every 300 words.
- Lock in your Google Business Profile. Real-time hours, accurate pricing, current photos, recent reviews. Agentic booking and “Google calls businesses” workflows favor profiles with clean, current, machine-readable data.
- Build for multimodal queries. Make sure your product images, before/after photos, and explainer videos are properly tagged, alt-texted, and embedded with context. Users are already dropping images and video into Search.
- Plan for agent-to-agent traffic. Your next “visitor” may not be a person. It may be an information agent monitoring on behalf of a real buyer. Make your content easy for an agent to summarize, cite, and trust.
For a deeper look at how the underlying model landscape is shifting (and why Flash-class models are now the workhorse for agentic workloads), see my breakdown of how the major AI models compare for business use in 2026 and my take on why AI agents are the new core infrastructure.
What This Means for Local Businesses
Local is where Agentic Search hits the hardest, fastest. Google’s new agentic booking, on-your-behalf calling, and 24/7 information agents are designed to compress the path from “I need a plumber” to “appointment booked” into a single conversation. Three things matter most for local brands right now:
- Profile completeness: every category, service, attribute, and photo filled in. Missing fields disqualify you from agentic suggestions.
- Live availability and pricing: if your competitor exposes real-time slots through a Google partner integration and you do not, the agent will quietly route the booking to them.
- Reputation signals at scale: review velocity, review recency, and review quality all weigh into how an information agent decides who to surface for an apartment hunt, a karaoke booking, or a service call.
If you are still treating local SEO as “set up the listing and forget it,” agentic Search is going to make that gap painful very quickly.
Frequently Asked Questions
What is Agentic Search?
Agentic Search is Google’s new mode of search in which AI agents can run jobs in the background, monitor the web 24/7, take actions like booking and calling businesses, and build custom generative UI to answer your question, instead of only returning a list of links.
What is Gemini 3.5 Flash?
Gemini 3.5 Flash is Google’s newest Flash-tier model from the Gemini 3.5 family, released at I/O 2026. It is now the default model behind AI Mode in Google Search and is built specifically for long-horizon agentic and coding workflows.
Is AI Mode replacing Google Search?
No. AI Mode is now built into Search globally and runs alongside traditional results. Google reported all-time-high overall query volume and one billion monthly AI Mode users, meaning both surfaces are growing together, not replacing each other.
How does Agentic Search affect SEO?
The ranking unit is shifting from “the page that earns the click” to “the source that earns the citation” inside an AI-generated answer. Structured data, original information gain, E-E-A-T signals, and machine-readable business data become more important than raw keyword density.
When will information agents and agentic booking be available?
Information agents launch first for Google AI Pro and Ultra subscribers in summer 2026. Expanded agentic booking, including local experiences and services with on-your-behalf calls, rolls out to everyone in the U.S. in summer 2026. Generative UI mini apps follow in the coming months.
The Bottom Line
Google just rewrote the rules of being found online. The Search box is no longer a search box, it is an agent interface. AI Mode is no longer an experiment, it is the default. And the model behind all of it, Gemini 3.5 Flash, is fast enough and cheap enough to power workflows that used to take days.
The brands that win this era will not be the ones with the most keywords or the longest blog posts. They will be the ones that publish information worth citing, expose clean structured data agents can act on, and treat every AI surface, including AI Mode, Gemini Spark, ChatGPT, Claude, and beyond, as a real distribution channel.
Agentic Search is here. The marketers who adapt now will own the next decade of discovery.
About Jason Pollak
Jason Pollak is a marketing strategist with over 10 years of experience building campaigns for entertainment brands, artists, and businesses across music, film, television, eCommerce, and B2B SaaS. As Director of Marketing at Young Money Entertainment, he grew Lil Wayne’s Facebook following from 10 million to 50 million and managed over 60 million followers across the roster. He also served as Paid Media Director at Horizon Media, launching major TV shows for History Channel, A&E, WWE, and Lifetime, and led film marketing for Utopia Distribution, generating over $10 million in revenue on a $200K media spend. Jason specializes in paid media, organic social strategy, email automation, SEO, content development, and AI-driven marketing systems. He holds a BA in English Literature from Binghamton University and a Masters in Media Studies from Brooklyn College. Learn more at jasonpollakmarketing.com or schedule a call.