Google I/O 2026 just redefined what search actually is. The announcements around information agents, Universal Cart, and the redesigned search box aren’t incremental updates. They represent a fundamental shift in how your customers find and buy from you. If you’re responsible for marketing or revenue at a mid-to-large company, your Google agentic search strategy 2026 needs to account for a world where users get answers without ever visiting your website.

I’ve spent the last 48 hours mapping these announcements against real business models. Here’s what I found, and here’s what you should actually do about it.

The Search Box Changed for the First Time in 25 Years. That’s Not a Cosmetic Update.

Google launched what it calls the “intelligent search box” at I/O. The new interface accepts longer prompts, surfaces AI options directly, and routes queries to AI agents by default. When Google changes the entry point, user behavior follows.

Here’s how it works: instead of typing “best CRM for small business” and clicking through ten results, users now type “help me choose a CRM for my 15-person marketing agency with a $500/month budget” and get a synthesized answer. The query is longer. The intent is clearer. And the user may never see your comparison article.

This matters because your top-performing content was probably built for short keyword queries. That content still exists, but the traffic it generates may erode as users shift toward conversational, multi-step requests.

Read the full story at Search Engine Land →

Information Agents Now Scan the Web Continuously. Your Content Feeds a System, Not Just Users.

Google Search now runs on Gemini 3.5 Flash globally. But the bigger announcement was “information agents” that continuously scan the web to update answers. These agents don’t just crawl and index. They synthesize, summarize, and deliver real-time answers.

The business model implications are significant. If your revenue depends on being the destination for informational queries, that traffic is at risk. A study from Victorious found that 90% of brands have zero AI search mentions. That’s not a visibility problem. That’s a revenue problem waiting to show up in your quarterly reports.

But here’s the flip side: if your brand or product gets cited by these agents, you gain visibility without paying for ads. The question is whether you’re in the 10% getting cited or the 90% getting ignored.

Read the full story at Search Engine Roundtable →

Building Your Google Agentic Search Strategy 2026: Who Wins and Who Loses

Here’s the framework for understanding where your business falls. I’ve mapped this against the specific announcements from I/O and the early data we’re seeing from AI citation tracking.

Winners in agentic search:

Brands with strong entity recognition. AI agents need to name sources. If Google’s knowledge systems clearly associate your brand with a topic, you get cited. This isn’t about domain authority alone. It’s about whether Google understands what your brand actually is and what expertise it represents. A B2B software company that’s published original benchmarking data for five years has stronger entity recognition than one that’s published generic blog content for the same period.

Transactional businesses that integrate early. Universal Cart creates a direct path from AI answer to purchase. If you’re in the cart, you’re in the game. With 60 billion products now in Google’s Shopping Graph, the infrastructure exists. The question is whether you’ve connected to it.

Companies with proprietary data or research. An analysis of 200 GPT-5.2 responses found that high-reasoning queries run deeper research and cite more sources. When the AI “thinks harder,” it rewards content that provides information it can’t synthesize from generic sources.

Losers in agentic search:

Affiliate and informational sites. When an agent answers the query directly, there’s no click. The business model breaks. If your revenue depends on capturing traffic from “best X for Y” queries and monetizing through affiliate links, that traffic is evaporating.

Ad-dependent publishers. Fewer clicks means fewer impressions. The display ad revenue model doesn’t work when users stay in Google’s interface. Some publishers will pivot to subscription models. Others won’t survive the transition.

SEO strategies built around keyword volume. Query intent is shifting. Traffic from 100,000 informational searches may convert worse than citations in 1,000 agentic responses. The volume-based playbook is losing its edge.

“Preferred Sources” Labels: A New Competitive Moat Is Forming

Google is testing a “Preferred Sources” label on citations within AI Mode responses. We don’t know yet whether this means Google is prioritizing these sources in results or just labeling sources it would have shown anyway. But here’s the problem: by the time we know for certain, early movers will already have a significant head start.

Think of it like featured snippets in 2016. Brands that optimized for snippets early captured position zero visibility for years. If “Preferred Source” status becomes a ranking factor for AI citations, the same dynamic will play out. The competitive moat compounds over time.

Read the full story at Search Engine Roundtable →

Universal Cart: Google Is Building the Transaction Layer, Not Just Discovery

Google introduced Universal Cart at I/O. It’s a centralized shopping cart that works across Google Search, with expanded Universal Checkout Protocol (UCP) and AP2 integrations. With 60 billion products indexed in the Shopping Graph, this isn’t an experiment. It’s infrastructure.

Here’s what this means practically: a user can ask an AI agent “find me running shoes under $150 with good arch support” and complete the purchase without ever visiting a retailer’s website. The retailer gets the sale. Google owns the experience. And retailers who haven’t integrated become invisible to these agentic shopping queries.

But here’s the thing: the architecture Google is building for shopping will likely extend to other verticals. Service businesses, B2B companies, and professional services should watch this closely. The patterns that work for e-commerce will inform how Google handles complex transactions in other categories.

Read the full story at Search Engine Land →

The Infrastructure Investments You Need to Make Now

Based on everything announced at I/O and the early patterns we’re seeing in AI citation data, here are the specific infrastructure investments that will determine whether you’re visible in agentic search.

1. Structured data at scale. This isn’t just about having Schema markup on your pages. It’s about having complete, accurate, and comprehensive structured data across your entire site. Product variants, merchant listings, FAQ content, author information, organization details. The Shopping Graph has 60 billion products because retailers provided the structured data to populate it. Your content needs the same treatment.

2. AI citation tracking. You can’t improve what you don’t measure. Microsoft Clarity’s new AI citations report is a start. But you should also be manually monitoring how AI systems respond to queries about your brand, your products, and your key topics. Build a weekly monitoring process.

3. Agent-readable content. There’s a debate happening right now about llms.txt files. Google Search says you don’t need it. Lighthouse now audits for it. My take: plan for eventual adoption. The direction is clear even if the timeline isn’t. Machines need to parse your content reliably, and the sites that make it easy will have an advantage.