What to Expect at Google Marketing Live 2026

Google Marketing Live 2026 is scheduled for May 20

Google Marketing Live 2026 lands on May 20 at 8:45 AM PT, and the official framing makes the agenda clear before a single slide drops. Google is centering the event on what it calls “the Gemini advantage,” with agentic commerce, AI-powered campaigns, and YouTube performance as the headline themes. [3] For performance marketers, that framing is a signal, not just a tagline. Every major product area, from Performance Max to measurement to creative, is likely to get an AI-first reframe at this event.

The event runs fully virtual, with a 90-minute Product Innovation Keynote from 9:00 to 10:30 AM PT, followed by a new segment called “Ads Decoded Live,” a live podcast recording with product leaders including Ads Product Liaison Ginny Marvin. [1] That format change from 2025’s Q&A is worth watching. It suggests Google wants tighter editorial control over how announcements get interpreted in real time. An EMEA edition follows on May 21 at 11:00 AM BST. [2]

GML 2026 also falls the day after Google I/O 2026 closes, which means any Search or Gemini model announcements from I/O could flow directly into the Ads keynote. That overlap is not accidental.

The anticipated shift to fully agentic AI campaigns

Agentic commerce is the term Google has been building toward for months, and GML 2026 is where it will likely get a formal product definition. [3] The concept involves AI agents that can take autonomous actions on behalf of users, such as researching products, comparing prices, and completing purchases, without the user manually navigating each step. For advertisers, that changes the conversion funnel at a structural level.

If Google’s agents are handling discovery and decision-making for consumers, then the traditional click-based model of search advertising starts to look incomplete. Advertisers may need to optimize for agent-readable signals rather than just human-readable landing pages. That could mean structured data, product feed quality, and first-party data become more important than ad copy or keyword bids in certain contexts.

From what I have seen in Google’s pre-event messaging, the agentic framing is not limited to Shopping. The official keynote description promises tools to “create, capture and convert demand across the expanding Google universe,” which reads like a pitch for cross-surface agent behavior spanning Search, Maps, YouTube, and Gmail. [3] Whether Google delivers a coherent product or a collection of loosely connected demos is the real question. The ambition is clear; the execution is what GML will have to prove.

Speaker Vidhya Srinivasan, VP and GM of Ads and Commerce, is likely to anchor this section of the keynote. Her presence alongside Philipp Schindler, Google’s Chief Business Officer, signals that agentic AI is being positioned as a commercial priority, not just an engineering show. [1]

How Performance Max will likely evolve again

Performance Max has been Google’s most contested campaign type since its rollout, and each GML has brought another round of controls, reporting additions, or scope expansions. GML 2026 will almost certainly continue that pattern. The question is whether Google adds meaningful transparency or just more automation levers that advertisers cannot fully inspect.

The likely direction is deeper Gemini integration inside PMax, particularly in asset generation and audience signal interpretation. Google has been expanding Asset Studio, its AI creative tool, and PMax is the natural home for generated assets at scale. [1] Expect announcements around automated video creation, image adaptation across formats, and possibly text generation tied directly to product feeds.

Reporting has been the persistent complaint from PMax advertisers, and Google knows it. GML 2025 added search term insights and asset group reporting, but the campaign type still gives advertisers less visibility than Standard Shopping or Search. [10] A reasonable expectation for 2026 is some expansion of channel-level breakdowns, showing how budget is splitting across Search, Display, YouTube, and Discover within a single PMax campaign. That data exists inside Google’s systems; the decision to surface it is political as much as technical.

There is also the question of Demand Gen’s relationship to PMax. Google has been positioning Demand Gen as the upper-funnel complement to PMax’s conversion focus, but the boundary between them has never been crisp. GML 2026 may clarify that positioning or blur it further by giving PMax more upper-funnel reach capabilities. Either outcome has real budget allocation implications for advertisers running both campaign types.

New measurement solutions for a cookieless world

Gaurav Bhaya, VP and GM of Ads Measurement, is confirmed as a GML 2026 speaker, and his presence on the keynote stage is a strong indicator that measurement will get significant airtime. [1] Google has been building toward a post-cookie attribution model for years, and the practical tools for advertisers, improved conversions, consent mode, and data-driven attribution, are all due for updates.

Improved conversions, which use hashed first-party data to recover signal lost to consent gaps, have been one of Google’s more credible measurement answers. Expect GML 2026 to push harder on adoption, possibly with new integrations that make setup easier for advertisers who have not yet connected their CRM data to Google Ads. The pitch will likely center on signal quality: Google’s argument is that better first-party data inputs produce more accurate attribution, which improves bidding and ultimately ROAS.

Privacy Sandbox has had a rocky road, with Google abandoning its plan to deprecate third-party cookies in Chrome in favor of a user-choice model. [11] That shift does not eliminate the measurement problem; it just changes its shape. Advertisers still need durable attribution that works across consented and non-consented environments. Google’s answer is likely to involve a combination of modeled conversions, consent mode v2 signals, and possibly new integrations with Google Analytics 4 that tighten the loop between on-site behavior and campaign performance.

What I am skeptical about is whether Google will address incrementality measurement in any serious way. Data-driven attribution is still a last-touch-adjacent model in many configurations, and advertisers running large budgets need incrementality testing to understand true campaign contribution. Google has the tools to support this, but it has been slow to make them accessible outside of managed accounts.

Creative in AI-driven advertising

Asset Studio is Google’s clearest bet on AI-generated creative, and GML 2026 will almost certainly expand what it can do. The current version handles image and text generation for Display and PMax assets, but the logical next step is video, and Google has the model infrastructure to support it through Veo, its video generation model announced at I/O 2025. [1]

AI-generated video for YouTube ads would be a significant shift. Producing video has always been the barrier that kept smaller advertisers off YouTube’s premium inventory. If Google can lower that barrier through automated video creation tied to product feeds or landing page content, it opens YouTube to a much broader advertiser base. That is good for Google’s revenue and potentially useful for advertisers who have been priced out of video production.

The creative quality question is real, though. AI-generated video assets from current models are often visually coherent but tonally generic. For brand advertisers with strict identity guidelines, automated creative is a liability risk, not a time-saver. Google will need to show that Asset Studio can work within brand constraints, not just produce plausible-looking content at scale. Expect the keynote to show polished demos; the actual output quality in production environments is a different story.

There is also a brand safety dimension here. If Google is generating creative assets on behalf of advertisers, questions about accuracy, representation, and compliance become Google’s problem as much as the advertiser’s. That liability question will not be answered at GML, but it will be the subtext behind every creative AI announcement.

Expected updates for Google Shopping and retail

Google Shopping has been getting steady updates as Google tries to compete with Amazon’s closed-loop retail media model. GML 2026 is expected to bring more agentic commerce features into the Shopping surface, potentially including AI-assisted product discovery that surfaces items based on conversational queries rather than keyword matches. [3]

Product feed quality has become the new keyword strategy for Shopping advertisers, and Google has been pushing retailers to enrich their Merchant Center data with attributes like sustainability certifications, return policies, and detailed product specifications. Expect GML 2026 to introduce new feed attributes or AI-assisted feed enrichment tools that pull missing data from product pages automatically. That would reduce the manual work of feed management while giving Google’s systems more signal to work with.

Virtual try-on and augmented reality shopping features have been in Google’s pipeline for several years, and the technology has matured enough that a broader rollout is plausible. Google’s New Fronts 2026 messaging referenced the “Gemini advantage” in the context of shopping experiences, which could include generative AI product visualization. [6] For fashion, home goods, and beauty retailers, that capability would be meaningful. For most other categories, it is a feature looking for a use case.

Selin Song, Global President of Google Customer Solutions, is on the speaker list, and her focus area typically covers small and mid-market advertisers. [1] That suggests at least some Shopping announcements will be framed around accessibility for advertisers who do not have dedicated feed management teams or large creative budgets.

How these changes will impact advertiser workflows

The cumulative effect of agentic AI, automated creative, and modeled measurement is a Google Ads platform that requires less manual input and offers less manual control. That trade-off is not inherently bad, but it demands a different skill set from the people managing these accounts.

Advertisers who have built their competitive advantage around keyword-level bid management and granular ad copy testing will find that advantage eroding. The new use points are upstream: feed quality, first-party data infrastructure, audience signal richness, and creative strategy at the brief level rather than the execution level. An advertiser who feeds Google’s systems better inputs will outperform one who tries to override the automation with manual controls.

The “Ads Decoded Live” segment at GML 2026 is worth watching for exactly this reason. [1] Ginny Marvin and the product leaders joining her will likely field questions about control, transparency, and what advertisers are actually expected to manage versus what the system handles. The answers will tell you more about practical workflow implications than the keynote demos will.

One thing to watch closely is how Google handles the tension between automation and accountability. When a PMax campaign underperforms, advertisers currently have limited tools to diagnose why. If Google is expanding automation further into creative, bidding, and audience selection simultaneously, the diagnostic problem gets harder, not easier. GML announcements that add automation without adding corresponding reporting are a net negative for sophisticated advertisers, regardless of how they are packaged in the keynote.

The event’s timing, immediately after Google I/O, also means that any Gemini model improvements announced on May 19 could be referenced as the foundation for GML’s Ads announcements on May 20. Advertisers should watch both events together rather than treating them as separate tracks. The underlying model capabilities announced at I/O are the engine; GML is where Google shows you the dashboard.

Warsaw Streamer Łatwogang Breaks Guinness World Record for Charity Fundraising

How to Get Started in Digital Marketing

What Is llms.txt and Why It Matters for Your Content