What happened on May 19, 2026
Some industry events are covered with serious headlines and forgotten within a week. Others, when you analyze them closely, make you realize they will change the way you work for years to come. Google I/O 2026 belongs to the second category.
On May 19, 2026, Google announced the biggest redesign of Search in 25 years. That is not public relations hyperbole: it is how the company itself described it. And the numbers attached to the announcement support that claim.
AI Mode reached 1 billion monthly users. Queries in AI Mode have doubled every quarter since launch. AI Overviews now serves 2.5 billion users per month.
TechCrunch framed its coverage bluntly: “Google Search as you know it is over.“. We can debate whether the headline is exaggerated. What is not up for debate is that the search ecosystem we have known since the early 2000s is being replaced by something structurally different.
And those of us working in digital visibility need to understand exactly what changed, what it means, and what needs to be done.
The five announcements that matter for SEO
Google announced more than one hundred updates at I/O 2026, from augmented reality glasses to new Gemini video capabilities. Here, the focus is on the five that directly change how search works and what it means to rank in 2026.
Gemini 3.5 Flash: the model that now defines what appears in Search
The most immediate and least visible change for the end user is also the most structural: Google replaced the default AI Mode model with Gemini 3.5 Flash starting on May 19, across every country where AI Mode is available.
Why does this matter for SEO? Because the model processing search queries is literally the referee deciding what content deserves to be cited in an answer. According to Sundar Pichai in the keynote, Gemini 3.5 Flash is roughly four times faster than most frontier models and, within Google’s Antigravity infrastructure, twelve times faster.
Speed is not just a technical performance metric: it is what makes autonomous search agents viable at scale. In addition, the benchmark shown at I/O indicated that Gemini 3.5 Flash outperforms Gemini 3.1 Pro across most evaluation categories.
A Flash model outperforming the previous Pro model is the kind of qualitative leap that changes the quality threshold content must clear to be cited.
What this means in practical terms is simple: a stronger and faster model will be more demanding, not less. Content that ranked through keyword saturation or formulaic structure has fewer chances against a model with greater semantic and contextual evaluation capacity.
The redesigned search box: the end of the short keyword
The most noticeable visual change for users is the redesign of the search box, which Google describes as the biggest update to that interface in more than 25 years. The box expands dynamically to accommodate longer queries and accepts text, images, files, videos, and Chrome tabs as inputs from a single entry point.
This evolution reflects where query behavior is heading: users are sending fuller context, not shorter keywords. Someone searching for a precision machining supplier no longer types “cnc machining spain”; they describe the problem, attach a PDF drawing, and indicate the expected production volume. Search in 2026 is designed to process that enriched query.
The implication for content strategy is direct: a strategy built exclusively around two-to-four-word queries will lose a growing share of searches where users want depth. Pages that answer specific, conversational questions in clear and direct language are better positioned for this new input format.
AI Mode and AI Overviews: two surfaces, one objective
It is important to understand the distinction between these two features because they shape strategy in different ways.
- AI Overviews appear above traditional organic results and keep the blue links underneath.
- AI Mode fully replaces the traditional SERP with a conversational interface.
In AI Overviews, traffic can still come from the organic positions below the summary. In AI Mode, you are either cited or you receive no visibility.
Follow-up questions can now be asked directly from within an AI Overview, flowing into a conversational exchange in AI Mode without the user needing to navigate away. Context accumulates across turns, and the sources that appear become more relevant as the conversation progresses.
What this creates is an inverted funnel architecture for traffic: the user enters through a search, gets an answer in AI Overviews, asks a follow-up question that takes them into AI Mode, and may get everything they need without visiting any website.
Analyst Marie Haynes summarized it precisely: “Many SEOs feared Google would announce at I/O that AI Mode becomes the default mode. That is not the case, but we are very close to it.”
Search agents: when the user is no longer searching
This is, without question, the most transformative announcement of I/O 2026 from a strategic standpoint and, paradoxically, the one that received the least headline coverage. Google announced that in summer 2026 users will be able to deploy AI agents that continuously monitor the web, 24 hours a day, on their behalf.
The workflow is straightforward: the user describes what they want to search for or monitor. The agent crawls blogs, news sites, social posts, and real-time data on finance, shopping, and sports. The user receives a notification with a synthesized answer, including the ability to take action, without needing to return to Google.
Haynes identified the structural consequence clearly: “Right now, SEO is optimized for a specific moment in time: a user typing a query right now. With background agents, you are optimizing for continuous monitoring.”
The paradigm shift is deep. For 25 years we have optimized to capture a user’s intent at the moment they open a search engine. Autonomous agents remove that moment: the user has already described the need once, and the agent searches on their behalf indefinitely. Website content is no longer competing only to appear in front of an actively searching human; it is competing to become the source the agent returns to systematically.
Agents will initially be available to Google AI Pro and Ultra subscribers, launching in summer in the United States.
Gemini Spark and personal intelligence: the search engine knows you
Spark is a personalized assistant running on Gemini 3.5 Flash and integrating Gmail, Docs, Calendar, Slides, and more Google Workspace tools. According to keynote materials, Spark keeps running even when the device is closed, using Antigravity infrastructure.
Alongside this, Personal Intelligence expanded to nearly 200 countries and 98 languages on May 19. It connects AI Mode with the user’s Gmail, Google Photos, and soon Google Calendar.
From now on, two users searching the same phrase in the same city can receive significantly different results based on their email history, saved images, and behavioral context.
What this means for B2B visibility strategy is significant: a procurement manager at an industrial company who has already received supplier proposals, has tender emails in the inbox, and has saved catalogs in Drive will see personalized search results based on that context.
Being present in the ecosystems where that decision-maker already operates —corporate email, editorial mentions, authoritative LinkedIn posts— will influence how Google weights your company within its personal intelligence layer.
The May 19 Core Update: the other story that went unnoticed
At the same time as the I/O announcements, Google launched a Core Update. Marie Haynes connected the timing directly to I/O: “It makes sense given that Gemini 3.5 Flash now powers the AI features in Search.” Harpreet Singh Chatha, an SEO and AI Search consultant, noted that this update may be targeting sites that are over-optimizing for AI citations.
The gap between Google’s public guidance and its product roadmap keeps widening. Infrastructure is changing faster than documentation. This is an observation that stands up well: Google updates its engine before it updates its documentation, which means professionals who wait for the manual to tell them what to do will always arrive late.
What Google confirmed about the right strategy
In May 2026, Google published its first consolidated guide to optimizing for generative AI features in Search, announced by John Mueller through the Google Search Central Blog. The core message is direct: there is no separate strategy for AI search. AEO and GEO are foundational SEO applied to an AI surface.
That confirmation has a clear impact on how digital visibility investment should be structured. It is not about abandoning technical SEO to chase “GEO” as a separate discipline.
It is about doing technical SEO well, adding content with original perspective and authority, and structuring it so both humans and language models can extract value efficiently.
The guide identifies five areas that underpin visibility in AI answers:
- unique, non-generic content (generic summaries that AI can produce on its own add no citation value; only content reflecting genuine expertise, original research, or first-hand perspective deserves inclusion).
- Technical accessibility of the site
- Authority signals and E-E-A-T
- Structured markup with schema.org
- And page speed and crawlability.
What this means for B2B companies in Spain
There is an understandable temptation to read these announcements and assume they are mainly relevant to large editorial publishers or mass-consumer ecommerce sites. They are not.
Industrial companies, professional services firms, and B2B technology businesses are exactly the types of organizations with the most to gain —and lose— in this new environment. Let’s look at why.
The B2B buyer is already using AI Mode. An engineering manager looking for industrial automation suppliers, a finance director evaluating production management software, a procurement lead comparing refrigerated logistics options: they are all using the same AI-powered search tools as individual consumers.
The difference is that the B2B decision cycle is longer and the evaluation criteria are more technical, which makes content quality even more decisive.
Monitoring agents will crawl your competitors. When a buyer configures an agent to “notify me when new CNC machining suppliers appear in Spain with IATF certification”, that agent will systematically review the same sources. If your site is not crawlable, if your content is not structurally legible for an AI model, if you lack proper schema markup and clear semantic structure, the agent will not include you in its results. Not because you do not exist, but because it cannot read you accurately.
Personal intelligence creates cumulative advantages.If a company has been present in industry publications, specialized media, and editorial mentions that reached decision-makers’ inboxes, that presence becomes a positive signal within Google’s personal intelligence system. It is the kind of advantage that cannot be built in 30 days, but once built, compounds exponentially.
The new metric map for 2026
The metric system most B2B marketing teams still use to evaluate SEO is outdated relative to the ecosystem just described by I/O 2026. These are the metrics that should start appearing in dashboards over the coming months:
Share of Voice in AI Overviews. Brands cited inside AI Overviews receive 35% more organic clicks than competitors that are not cited. The goal is no longer just to rank: it is to be named. Tools such as Ahrefs Brand Radar, LLM Pulse, or GEO Metrics can help begin tracking this metric.
Rising impressions with falling CTR as a warning sign. If your Google Search Console shows impressions going up while clicks and CTR decline, you are already seeing the effect. AI Overviews and AI Mode are absorbing traffic that used to reach your pages. This is the symptom indicating that the site has algorithmic visibility but not AI citability.
Brand mention frequency in authority publications. AI models weigh how often a brand appears in quality editorial content, with or without an active link. Tracking how many industry publications mention your company, in what context, and how often becomes a leading indicator of AI visibility.
What to do now: the practical roadmap
Translating the I/O 2026 announcements into concrete decisions, these are the priorities for the next 120 days:
First, audit site crawlability for AI agents. Google’s autonomous agents do not execute JavaScript the same way a human browser does. If key site content depends on client-side rendering to be visible, agents may not be able to read it properly. The priority is to ensure the most important content is available in clean HTML. See our GEO Audit service:
Second, implement complete schema markup. Article, FAQPage, and BreadcrumbList on content pages. LocalBusiness or Organization on the homepage and contact pages. Product or Service on offer pages. This structured layer is what allows Gemini 3.5 Flash to process content precisely and cite it confidently.
Third, restructure content for direct extraction. AI models extract from the first two or three paragraphs under each H2. If the opening paragraph of each section does not directly answer the implicit question in the subheading, the model moves to the next result. This is not a length problem: it is an information structure problem.
Fourth, start producing content with original perspective. Generic content will keep losing ground. First-party data, first-hand experience, and technical content that only your team can produce are what create value differentiation in AI Mode. An article that summarizes what any LLM can generate on its own will not be cited by that same LLM.
Fifth, diversify your presence across the AI ecosystem. Beyond the website, LinkedIn profiles of team specialists, mentions in industry publications, participation in technical forums, and videos with well-structured transcripts become surfaces Google agents crawl and may cite from.
Visibility strategy can no longer be limited to your own website.
The landscape ahead
I/O 2026 was not an announcement that Google is experimenting with AI in search. It was confirmation that AI search architecture is already the main architecture, with more than 1 billion monthly users in AI Mode and 2.5 billion in AI Overviews. Summer announcements —information agents, Gemini Spark, expansion of Universal Cart— will deepen that transition.
What does not change: SEO remains the foundation. Google confirmed that AEO and GEO are not separate disciplines, but layers built on a solid technical SEO base. Companies that built domain authority, coherent content architecture, and sustained editorial presence over recent years are in a stronger position to benefit from this shift than those that tried shortcuts.
What does change is this: the goal of visibility strategy is no longer exclusively to get a human to click a link. It is to get an AI model, or an autonomous agent acting on behalf of a user, to identify your company as the most relevant and trustworthy source for a category of queries.
That is the new definition of ranking in the 2026 search ecosystem.
If you want to assess how your site is positioned against these new demands, Smart Team can help with an audit that includes crawlability for AI agents, AI Overviews citability analysis, and a technical and editorial implementation roadmap.
What type of content is most likely to be cited by AI Mode? Content with original perspective and information that cannot be generated automatically: first-party data, documented use cases, technical analysis signed by experts with verifiable track records, and comparisons backed by original empirical evidence. Content that merely summarizes what others have already published has less and less differentiating value against a model that can generate that summary itself.
Consultor SEO con más de 15 años de experiencia en Marketing, optimización web y estrategias digitales. Ayudo a negocios locales, pymes y grandes empresas a mejorar su posicionamiento online, alcanzar sus objetivos de crecimiento y adaptarse a un mundo digital cada día más competitivo.
Schedule a 30-minute meeting
Want to know how we can generate more leads for your business in Barcelona?
Leave us your email and phone number and we’ll schedule a no-obligation call to give you a personalized assessment of your current marketing strategy.
