Google I/O, Marketing Live and EMEA
Google I/O, Marketing Live, and an EMEA-focused digital edition ran back to back from 19–21 May 2026, amounting to the biggest restructure of Google’s advertising and search products in a decade. The message across all three was the same. Gemini now sits at the core of Search, advertising, and measurement, and the manual controls paid search has run on for two decades are being dismantled.
The headline figures from the I/O keynote frame everything that followed: AI Overviews has crossed 2.5 billion monthly active users, AI Mode has surpassed 1 billion, and the average AI Mode query runs roughly three times longer than a traditional Google search. A Google-commissioned Ipsos study fielded in December 2025 as a global average across select countries (n=13,189 online shoppers) found that 75% of people say they make faster, more confident decisions using AI Mode.
Vidhya Srinivasan, Google’s VP of Ads and Commerce, summed up the keynote ethos in the official collection post: ‘the only way to win in the age of AI, is with AI’. That’s the lens Google is asking marketers to apply to everything announced that week.
What Changed for Search
The Intelligent Search Box is Google’s redesigned Search input. The box expands dynamically as users type longer, more conversational queries. Multimodal input is accepted across text, files, images, and video. Background agents continue working on tasks after the initial query is closed.
In practice, the search box is becoming a job queue. Someone searching for ‘best wedding photographer Manchester under £2,000 available in October’ used to get ten blue links. The Intelligent Search Box can take that query, treat it as a brief, and run a background agent that compiles options, checks availability, and presents a structured comparison hours later.
That demand-side shift is what reshapes everything downstream. Queries arrive longer, contain more intent context, and come with the user expecting a synthesised answer rather than a list of links. For advertisers, the search box is the surface the new ad formats run on. For SEO and content teams, it’s the surface their organic content now competes for visibility on. The two pressures intersect.
The structural change at the model level matters for SEO and ads alike. AI Mode now runs on Gemini 3.5 Flash as its default model, with the ability to call agents for follow-up tasks. AI Overviews use the same underlying intelligence. The result is search composed from sources rather than ranked against documents, with paid placements interwoven where commercial intent is high.
The implication for the consideration funnel is significant. The user who used to visit ten product pages, a comparison site, and a forum may now have all of that synthesised by AI Mode in a single conversation. Brands cited as authoritative within that synthesis win the consideration set. Brands not cited disappear from it entirely. For both brand and performance marketers, the strategic question moves from how to be found to how to be cited.
What Changed for SEO
The SEO story from GML 2026 is largely indirect. Google did not announce new SEO tools at GML. What it announced was an acceleration of the shift from ranked links to synthesised answers, and that shift has measurable consequences for organic traffic.
SISTRIX analysed over 100 million German keywords and found position 1 organic CTR drops from 27% to 11% when an AI Overview is present, a 59% reduction. Ahrefs found a similar pattern globally on a 300,000-keyword dataset, with roughly 58% lower CTR when an AI Overview surfaces above the organic results. Both studies pre-date GML 2026, but they document the trajectory the May announcements are designed to accelerate.
The practical question for brands has shifted from ‘how do we rank?’ to ‘how do we get cited?’. AI Mode pulls answers from across the web and synthesises them, with citations to the sources used. In practical terms, citation inside synthesised answers is the new optimisation target. Three things appear to matter most for citation probability.
First, structured product data. AI Mode pulls heavily from the Shopping Graph, the product feeds, attributes, and inventory data Google has built up through Merchant Center over the past decade. Merchant Center feeds, structured data on product detail pages, and clean schema markup are not optional for retailers anymore. They’re the input layer Gemini grounds its answers on. Product titles, attributes, GTINs, and availability all flow directly into how Gemini composes shopping answers.
Second, brand authority signals. Reviews, citations across reputable publications, third-party mentions, Wikipedia presence, and trust signals such as registered companies, verifiable credentials, and named authors all increase the probability that Gemini cites a brand as authoritative when constructing an answer. Building those signals looks more like digital PR work than like keyword research.
Third, content that answers the question directly. The AI Overview pulls succinct, factually correct passages that match user intent. Long preamble before the answer hurts citation probability. Direct, structured answers help. The same logic that makes content quotable to a journalist makes it quotable to Gemini.
Google’s own framing is that brands need to ‘show up and get discovered across conversational AI surfaces’. That’s the company’s polite phrasing for what the SEO industry has started calling AEO or GEO: Answer Engine Optimisation, or Generative Engine Optimisation. The work is similar in spirit to SEO but the optimisation target is citation in synthesised answers rather than ranking position.
For UK and US brands with existing SEO programmes, the immediate priority is auditing how the brand currently appears in AI Mode and AI Overviews for its highest-intent commercial queries. The tooling has caught up over the past year, and the choice depends on which AI surfaces matter most.
Microsoft Clarity’s Citations feature went generally available on 13 May 2026 and is free, but its citation data covers Microsoft Copilot and partner platforms rather than Google. The dashboard sits inside Clarity at Dashboards > AI Visibility > Citations, and domain verification through Bing Webmaster Tools or Google Search Console is required. The Copilot-only scope is a meaningful caveat, but the structural insight is transferable: AI engines use similar retrieval frameworks, so the gaps Clarity exposes tend to correlate with gaps on Gemini-powered surfaces.
The Clarity dashboard also surfaces a useful new concept. Grounding queries are the shorter, keyword-like terms an AI system uses to retrieve content, distinct from the longer conversational queries users type into the chatbot. Content that ranks well for nuanced commercial queries can underperform on the grounding terms AI systems actually fetch with, which is the gap Citations exposes.
For Google AI Mode and AI Overviews specifically, the paid tools have caught up too. Ahrefs Brand Radar covers six AI surfaces (Google AI Overviews, Google AI Mode, ChatGPT, Perplexity, Gemini, and Microsoft Copilot) across a 150 million prompt database, with citations and share of voice broken out per platform. Semrush’s AI Visibility Toolkit covers Google AI Overviews and ChatGPT alongside its position tracking, with competitor benchmarking. SISTRIX has built AI Overview and AI Mode tracking into its platform, and in May 2026 opened its MCP server to all SISTRIX users, letting agencies query citation and ranking data directly from Claude or other AI tools without writing API code.
What Changed for Ads
Most of what GML 2026 actually announced was on the ads side. Seven items matter most.
Ask Advisor: one agent across the marketing stack
Ask Advisor was the single biggest practical announcement at GML. It’s a Gemini-powered agent that brings Google’s existing in-product advisors (Ads Advisor in Google Ads, Analytics Advisor in Google Analytics, and the equivalent agent in Google Marketing Platform) into one unified experience, with a Merchant Center agent joining the lineup soon. The GML move was to connect those agents into one cross-product surface.
You ask a question in natural language, and Ask Advisor pulls signals from across the connected products and either gives you a recommendation or executes the change. If it works as demonstrated, daily account-management work that takes an hour of clicking through tabs collapses to a paragraph of typing. The beta is live for English-language accounts now, with broader capabilities rolling out across the rest of 2026.
The strategic implication is that execution is becoming a commodity. The work earning its keep becomes strategy, measurement, and creative judgement. Routine optimisation work, which has occupied a meaningful chunk of every PPC account manager’s week for years, doesn’t.
Dynamic Search Ads retire in September
The DSA-to-AI Max migration was reconfirmed at GML, though it was originally announced on 15 April. From September 2026, every eligible Dynamic Search Ads campaign, automatically created assets configuration, and campaign-level broad match setting auto-upgrades to AI Max. No opt-out. Google Ads, Google Ads Editor, and the API stop accepting new DSA campaigns at the same point.
Google’s recommended default mix is now Performance Max, Demand Gen, and AI Max running together. For advertisers using DSA today, the September migration ends the question of whether to use AI Max. The question becomes how to configure it before it’s forced on the account with default settings.
AI Max takes four to six weeks to stabilise after migration, by Google’s own guidance via account teams. Voluntary migration tools are live now. For US Black Friday and UK peak prep alike, the arithmetic of a September forced migration before an October–November peak window is bad. The answer is to migrate in June or July, while traffic is calmer and there’s time to course-correct.
AI Brief: plain-language control over AI Max
AI Brief is the natural-language guidance layer for AI Max. Advertisers describe their brand, messaging, audience, and what to avoid in plain English. Gemini interprets those guidelines and generates ad copy and search matching accordingly. The three input types are Messaging Guidelines (what ads should and shouldn’t say), Matching Guidelines (which queries to prioritise or avoid), and Audience Guidelines (the customer profile to target).
AI Brief was announced on 30 April but received its full GML treatment as the headline AI Max upgrade. Rollout begins in English for AI Max for Search, with Performance Max and AI Max for Shopping to follow.
For the first time, advertisers can A/B test entire brand propositions inside Google Ads, rather than only creative variants. The skill that earns its keep becomes the quality of the brief itself, closer to traditional marketing thinking than to keyword-and-bid management.
Four new Gemini-powered ad formats in AI Mode
Google introduced four new ad formats at GML, all sitting inside AI Mode. From the official Google announcement:
Conversational Discovery Ads adapt to the user’s question in real time. Google’s example: someone searching for low-maintenance ways to make their home smell like a spa or a rainy forest sees creative built around that exact query, with relevant product features highlighted by Gemini.
Highlighted Answers surface relevant products or services as part of an AI Mode list recommendation. If AI Mode surfaces the best language apps for an upcoming trip, qualifying advertiser products can appear within that list as Highlighted Answers.
AI-powered Shopping Ads show Gemini-generated explainers detailing why a specific product is a good fit for the query. Google’s example covers an espresso machine search where Gemini writes a custom explainer for each surfaced product. Rolling out in the coming months.
Business Agent for Leads places an AI chat agent inside lead-gen ads. Users can ask questions and get answers drawn from the advertiser’s website rather than filling out a static form. Google’s stage demo used a university research scenario; specific verticals for the initial rollout were not detailed in the GML announcement.
Google recommends running AI Max for Search, AI Max for Shopping, and Performance Max to make full use of these formats. All four carry a ‘Sponsored’ label. And for the first time, the ad copy and the surrounding explanation are generated separately. Google describes the format as combining an advertiser’s own creative with an independent AI explainer: Gemini evaluates and synthesises product information and displays that context alongside the ad. Whether this independent-explainer architecture builds genuine trust or surfaces as obvious AI-generated copy is something marketers will watch closely across the rest of 2026.
Universal Cart, AI Max for Shopping, and the commerce layer
Universal Cart is Google’s cross-retailer cart, first introduced at I/O and detailed further at GML. It works across Search, Gemini, YouTube, and Gmail, with price tracking, deal alerts, back-in-stock notifications, and product compatibility checks built on Gemini. The Universal Commerce Protocol that underpins it is being extended into Demand Gen campaigns, YouTube Shopping ads, and AI Mode placements. Launch partners listed by Google include Nike, Sephora, Target, Ulta Beauty, Walmart, Wayfair, and Shopify merchants such as Fenty and Steve Madden. Affirm and Klarna are embedded as buy-now-pay-later options inside Google Pay.
Google’s stated rollout sequence is US first, then Canada and Australia ‘in the coming months’, and ‘later’ to the UK. The protocol is also expanding into hotel booking and local food delivery as new categories. So the UK is on the published roadmap, but at the back of the queue, and Google has not given a date. The likely reason for the gap is regulatory: the Digital Markets Act and consumer protection rules in the UK and EU create more friction for an agentic checkout that sits across multiple retailers.
AI Max for Shopping launched in closed global beta on 30 April, two weeks after AI Max for Search reached general availability. The Shopping version applies the same generative matching layer to product feeds and Shopping inventory rather than Search queries. Retailers in any market can request access through their Google account team. Broader rollout is expected before peak 2026 trading.
Direct Offers, the promotions pilot Google launched with Chewy, Gap, and L’Oréal in January 2026, gained promotion bundling. Booking and Expedia are joining the pilot with a travel-specific variant for AI-assisted trip planning.
Asset Studio and the Gemini Omni integration
Asset Studio gained two main upgrades at GML. The first is integration with Gemini Omni, Google’s new multimodal model family. Omni Flash, the first model in the family, is built around video generation from any combination of text, image, audio, and video inputs. Inside Asset Studio, that means advertisers can soon generate video assets from a brief without leaving the platform. The second is 1-Click A/B Testing for creative variants, which goes some way to addressing how poorly most accounts test creative today.
The 1-Click A/B Testing is the more immediately useful piece, because most accounts run badly designed creative variant tests and the tooling has lagged the rest of the platform for years. Both features are rolling out globally in English this summer.
Meridian, Qualified Future Conversions, and predictive measurement
Meridian, Google’s open-source marketing mix model, is being integrated into Google Analytics 360 as a unified measurement command centre. Google’s framing covers three capabilities: unifying first-party cross-channel data and signals in one place, measuring causal performance to identify what is driving business outcomes, and forecasting through predictive scenarios.
Paired with the Meridian integration is a new Gemini-powered predictive metric called Qualified Future Conversions. QFCs link upper-funnel spend to future sales through signals such as branded searches, helping advertisers attribute value that current attribution models miss. QFCs are in restricted pilot now, with broader beta later in 2026, and the predictive signal will eventually feed back into Meridian itself to refine MMM accuracy. For brands still working from last-click attribution reports, the analytical floor is rising.
Other Announcements Worth Knowing
GML 2026 covered far more than the headline items. The supporting announcements worth tracking:
Google Tag Gateway (GTG). Server-side tag routing through your own domain rather than Google’s. Solves the same data-loss problems as Meta’s Conversions API, with no code changes to existing tag setups. Google’s own primary figure is an 11% average uplift in measurement signals for advertisers who configured it during early rollout. Confidential computing is being added to the gateway by default, providing additional privacy guarantees on how first-party data is processed before encryption.
Demand Gen on Google Maps. Demand Gen ads launched directly on Google Maps as a new inventory surface, with automotive product feeds added and AI-assisted campaign creation that copies settings from an existing Performance Max campaign. Checkout links also expanded to nine new markets.
Product Value Adjustments (PVAs). Multipliers for conversion values on specific SKUs in Shopping and Performance Max, configured directly in Merchant Center or Google Ads. Lets retailers bid more aggressively on high-margin, low-return, or overstocked products without restructuring campaigns.
Data Manager API expansion. New direct connectors to Mailchimp, ActiveCampaign, and Klaviyo, plus programmatic setup for first-party data flows. Targets the mid-market adoption gap that has held back Data Manager since its 2025 launch.
Demand Gen Uplift Experiments and Campaign Type Attribution. A new incrementality testing framework and a clean Demand Gen attribution view that isolates conversions for proper comparison against paid social. Together these address the ‘where does Demand Gen fit’ question that has slowed budget allocation.
Privacy change taking effect 15 June 2026. Settings that control Google Ads cookie identifier collection from the Analytics tag become destination-specific. Data used in Google Analytics will be controlled by Google signals; data used in Google Ads will be controlled by Ads consent mode. Advertisers with ad storage set to denied may see remarketing list sizes decrease. Worth checking your consent configuration before mid-June.
Regional Rollout: What’s Available Where
For US advertisers, every announcement covered here is either live, rolling out across the summer, or in active US testing. Universal Cart starts in the US, with Canada and Australia in the coming months and the UK later.
For UK and EMEA advertisers, Ask Advisor’s beta, AI Max for Search general availability, Asset Studio’s Gemini Omni features, and the Meridian-in-Analytics-360 integration are all available now or imminently. AI Max for Shopping is request-only through Google account teams. Universal Cart and the new AI Mode ad formats are not in the UK yet; the UK is on the published Universal Cart roadmap but without a date. The agentic commerce surface is the largest single gap, driven by the regulatory friction that doesn’t apply in the US.
What to Do in the Next Four Weeks
Four actions are worth the time of any marketing team operating across Search, SEO and Ads, regardless of region, before the end of June.
First, audit every account for Dynamic Search Ads, automatically created assets, and campaign-level broad match settings. Build a migration calendar with manual upgrades scheduled for July at the latest. The September auto-migration arrives at exactly the wrong moment in any peak-season calendar, US or European.
Second, get into the Ask Advisor beta on the accounts you manage. Spend an hour a week in it through summer. The teams who know its limits and strengths by Q4 will be the ones still making good decisions when Gemini becomes the default interface across Google’s marketing products.
Third, if you sell online, run a Merchant Center audit and request access to AI Max for Shopping. Promotion feeds, product titles, structured data on product detail pages. UK and European retailers should treat AI Max for Shopping as the priority over Universal Cart, since the Shopping beta is reachable now while Universal Cart has the UK at the back of its rollout queue.
Fourth, run a citation audit for the SEO side of the house. Pick the ten highest-intent commercial queries in your category and check how AI Mode and AI Overviews answer them now. Microsoft Clarity is the free starting point for Copilot citation data; Ahrefs Brand Radar, Semrush, and SISTRIX cover Google AI surfaces in their paid tiers. The brands cited in those answers are the ones converting from organic traffic. The brands that aren’t cited will see CTR continue to compress through the rest of 2026.