Google this week published the season one finale of Ads Decoded, a 41-minute special recorded backstage at Google Marketing Live 2026. The episode, uploaded to the Google Ads YouTube channel on May 28, 2026, features host Ginny Marvin in conversation with four product leaders: Vidhya Srinivasan, VP and GM of Ads and Commerce; Chris Monkman, Senior Director for Ads in AI Experiences; Christine Turner, Managing Director for Measurement, Data and Audiences; and Josh Moser, Advertising Platform Senior Director. Together, they move through the technical rationale behind the biggest announcements from the event held on May 20 - covering search intent, AI mode ad formats, the AI Brief control layer, Asset Studio, and a set of new measurement metrics that Google says are now always on.

The episode sits in a specific context. Google Marketing Live 2026 was the 13th annual edition of the event, staged at the Bay View campus in Mountain View and livestreamed to more than 100 countries. The backstage Ads Decoded recording lets the product leaders go further than the keynote allowed.

Gemini and the shift beyond keywords

Srinivasan opened by describing the current search environment as "an expansionary moment." According to the recording, users are sending Google two distinct types of queries: short questions that need fast answers, and long conversational threads where they want to engage with Gemini across multiple turns. Both types are growing simultaneously, and the gap between how consumers use search and how advertisers build campaigns has widened as a result.

The keyword, in this framing, is no longer the organizing unit of search advertising. Monkman was explicit about the technical reason. According to the recording, it is "very difficult to ever choose keywords manually that can match these queries." The query stream has, in his characterization, "materially changed" - queries are longer, carry more embedded constraints, and arrive with context that a fixed keyword list cannot anticipate. Gemini, he explained, looks at the full AI Mode response as well as the user's query and uses that combined context to decide which ads are relevant. That is a qualitatively different matching process from the traditional keyword auction.

Srinivasan added that Gemini is "incredibly good at understanding content" across all of an advertiser's assets simultaneously, in a way that previous systems could not manage. The model can aggregate across a full asset library and produce a contextually appropriate response. The world is also multimodal, she noted - people see, talk, and hear - and Gemini's multimodal capabilities mean that the system can work with images and video inputs alongside text, not only the text-based signals that older infrastructure was built for.

For the advertising community that PPC Land has tracked through the AI Max rollout, this is not an abstract point. Dynamic Search Ads will automatically upgrade to AI Max for Search starting in September 2026. That structural shift, announced April 15, 2026, is the operational consequence of what the product leaders describe.

What AI Max and PMax unlock that standard campaigns cannot

The requirement attached to the new AI Search ad formats is specific: advertisers must adopt AI Max for Search campaigns or Performance Max with text customization enabled to appear in the new experiences. Monkman explained the two parts of that requirement in detail.

The first is targeting. According to the recording, automated targeting lets Gemini match ads to complex queries in ways that manual keyword selection cannot replicate. The second is creative generation. The new formats dynamically generate ad copy at query time - the copy is not pre-written but constructed to answer the specific user query in context. According to Monkman, user research shows that users "are expecting answers" rather than results that link out to a page. They want to understand why an ad is relevant before they click.

AI Max for Search campaigns delivered 27% more conversions at a similar CPA or ROAS compared to campaigns using mostly exact and phrase match keywords, according to Google's internal data cited in the new AI search ad formats announcement. The same data indicates that AI Max and Performance Max are unlocking billions of net new queries that those advertisers were not previously reaching.

The AI Brief as a control layer

Advertiser concern about control has been a recurring theme in the AI Max adoption conversation. The episode addressed it directly. According to Monkman, the AI Brief rolls targeting and creative messaging into one input. Advertisers describe their brand, their audiences, and their guardrails in plain language; the system interprets that into operational guidelines with previews. The feature builds on the text guidelines that already existed, with those guidelines automatically rolling into the AI Brief structure.

The episode explained the AI Brief explicitly as a response to advertiser feedback. There is a tension Google has acknowledged throughout the transition: AI-mediated matching and creative generation extend reach, but advertisers in brand-sensitive categories need constraints they can inspect and adjust. The Brief is designed to let both happen without requiring the advertiser to choose between automation and control.

AI Mode ad formats and the principal-agent problem

Monkman described a specific design principle shaping the new AI Mode ad formats: what he called the "principal agent problem." According to the recording, for users to keep trusting the AI product, they must feel that it is acting in their best interests. That principle drove several format decisions.

The independent AI explainer, one of the formats released at GML 2026, provides what Monkman described as an objective, independent evaluation of an ad and why it might make sense for the user. Critically, advertisers cannot directly influence that text. The explainer is generated by fetching search results about the advertiser and summarising them - the same architecture that standard AI Mode responses use. The goal is to give users a fair evaluation that sits alongside the advertiser's own creative message, which is separately labeled.

According to Monkman, user research shows that explainers "really do improve trust overall in the product as well as consideration of ads." His view is that higher trust leads to users being more willing to look at ads in AI experiences and ultimately to convert, producing higher-value clicks.

Direct Offers follow a similar logic. According to the recording, the format is designed so that users can only receive a better deal - the system cannot be used to charge more. That constraint is built into the architecture, not left to advertiser discretion.

The ad formats released at GML 2026 build on a structure that PPC Land has documented since January 2026, when Direct Offers first appeared as a sponsored deal format tied to the Universal Commerce Protocol. The shopping ad format for AI Mode was formally announced on February 11, 2026. GML 2026 expanded the range of formats available in that environment rather than introducing the category from scratch.

Ask Advisor across four products

Srinivasan described the unified Ask Advisor as a shift from having separate conversational agents in each product to having a single agent that works across Google Ads, Google Analytics, the Google Marketing Platform, and Merchant Center. The motivation she gave was practical: marketers increasingly want to complete tasks that span more than one product, and they want to do that quickly.

According to the recording, Ask Advisor can surface insights and data from other products while a user is working inside any single product - giving a broader picture before any change is made. It also works with the specialist agents inside each tool to bring in the relevant context. Srinivasan described it as "a brainstorming partner that you can go back and forth with" regardless of which product the user happens to be in at a given moment.

Ask Advisor was launched at Google Marketing Live 2026 on May 20, absorbing Ads Advisor, Analytics Advisor, and the forthcoming Merchant Center agent under a shared memory layer that retains user goals across tasks. It is currently in beta for English-language accounts globally.

Asset Studio: video, A/B testing, and external asset bridges

Josh Moser explained the capability that, according to the recording, stands out the most in the creative space: the ability to go from a general concept to a full slate of headlines, descriptions, images in multiple aspect ratios, and now video. Asset Studio at GML 2026 added Gemini Omni multimodal video creation, a natural language interface for briefing creative work, and 1-Click Creative Testing.

The 1-Click Creative Testing feature addresses a structural problem Moser described in detail. Historical performance metrics - click-through rates and conversion rates - show how an asset performed in aggregate, but they do not show whether the asset incrementally contributed to that performance. An image with a low click-through rate might appear to underperform, but it might simply be shown to an audience segment that had a low click-through rate to begin with. The asset could still be incrementally helping. The only way to separate those effects is a controlled experiment.

According to Moser, the previous process for running creative experiments in Google Ads was "a little bit painful." The new capability reduces it to a single workflow step: select an asset to test, and 1-Click creates the A/B structure without copying the campaign or creating synchronization issues. The interface, as shown in materials released with the May 20 announcement, allows a traffic split - the example showed 50% - and a defined test window. The output is a full readout of what incremental results the new asset drove.

Moser also addressed asset imports from external platforms. Many advertisers produce creative in Adobe products, Canva, or in their own digital asset management systems. Bridging those assets into Google Ads and tailoring them for different campaign types and surfaces has historically required significant manual effort. The updated Asset Studio provides a pathway for those external assets to appear directly in the advertiser's library, where they can be customised in a small number of steps and then shared for review with legal, marketing, or other stakeholders before going live.

The global English rollout of the updated Asset Studio is planned for summer 2026, as PPC Land reported separately on May 20.

Merchant Center feeds and the multimodal shift

Srinivasan addressed merchant discovery directly. According to the recording, the first action merchants can take is to maintain and update their Merchant Center feed with the highest quality data they can supply, refreshing it as quickly as products change. The second action is to use the conversational attributes that Google is introducing into Merchant Center feeds - fields designed specifically to capture the kind of information that AI-generated conversational responses need. That includes Q&A pairs and additional review types that would otherwise be difficult to surface. According to Srinivasan, having that detailed information in the feed "goes a long way" in providing contextual answers to conversational queries.

She also highlighted asset diversity. Because search is becoming more multimodal, having videos and high-quality images alongside text assets "becomes very important." The argument she made is that AI capabilities raise the baseline for all merchants - the average of what is possible goes up - but standing out within that new baseline still comes down to what makes a specific brand unique. Understanding and amplifying that is, in her characterisation, now the core of the merchant's job.

Measurement: from report card to performance fuel

Christine Turner opened her section by framing what she described as a graduation. After more than 20 years in measurement, she said the conversation has shifted from campaign metrics - tracking clicks - to business outcomes. Measurement is no longer just a report card on past performance. According to the recording, it is "becoming the fuel for performance moving forward and becoming more predictive."

Three specific metrics were discussed. Campaign Type Attribution for Demand Gen addresses a long-standing advertiser request for a way to isolate the direct link that Demand Gen campaigns have in conversion paths. According to Turner, the default attribution in Google's systems has been to de-duplicate conversions across all campaign types within a Google mix. Campaign Type Attribution allows marketers to see the specific contribution of Demand Gen and to compare it on an apples-to-apples basis with social platforms. Only 40% of Demand Gen conversions occur within the first 30 days, according to a figure Turner referenced, which makes the attribution window question material.

Attributed Branded Searches links ad exposure to subsequent branded search activity - whether a user who saw an ad then searched for that brand or specific product. Turner described it as "highly connected to future intent." According to the recording, it is already largely rolled out and always on in the interface, meaning it does not require advertisers to configure a separate experiment. In some categories, marketers can use it to predict future market share and future sales.

Qualified Future Conversions uses attributed branded search signals and other post-exposure consumer behaviors to mathematically predict what future profitable growth will look like for a given marketer. Like attributed branded search, it is always on, not something requiring a separate test. According to Turner, it gives marketers a tool to identify conversion events that may arrive further downstream - relevant for businesses with long decision cycles. The Meridian integration into Analytics 360, also announced on May 20, extends this predictive direction to Marketing Mix Modeling, moving it from a standalone data-science workstream into a standard component of an analytics platform.

Turner emphasized the continuing importance of tagging infrastructure. According to the recording, bringing Google Tag Manager's capabilities directly into Google Ads and Google Analytics is a significant step - making tag management accessible regardless of which product a marketer works from. She noted that tagging is something teams often configure once and do not revisit, and that the Tag Gateway integration - which routes tags through first-party infrastructure rather than third-party Google domains - produces an average 14% increase in observable conversions by recovering signals that ad blockers would otherwise strip. More partnerships for Tag Gateway are planned, according to Turner.

For Lead Gen advertisers specifically, Turner's advice was to build connections between early-signal conversions - form completions, website visits - and the downstream value that eventually results from those leads. Longer sales cycles mean that the AI campaign systems need to see data from across the full path, not only the initial conversion event, to optimise accurately. She acknowledged that not every advertiser starts from a position of having that data fully assembled. The instruction was to start with what is available rather than waiting for a complete data set.

The broader context for the advertising community

The episode closes with action items from each panelist. Moser suggested trying creative tools, with the explicit point that as creative inputs multiply, the role of the marketer shifts toward feeding the system with better material - assets that match the right audiences across the right surfaces. Monkman pointed to AI Max with text customization as the practical first step to access the new AI Mode formats, and to Direct Offers as an experiment worth running and reporting back on. Turner returned to data - specifically, Google Tag Gateway and Data Manager as the foundation, with third-party data linking in Google Analytics as an additional layer for seeing the full consumer path.

The subtext running through the episode is that the advertiser-facing consequences of AI Mode are arriving at scale. AI Mode surpassed one billion monthly active users globally according to figures disclosed at Google I/O 2026 on May 19. The query patterns that Monkman described - longer, more complex, arriving with embedded constraints - are not a future condition. They are the current state of the search environment that these formats and tools are responding to.

Timeline

Summary

Who: Vidhya Srinivasan (VP/GM, Google Ads and Commerce), Chris Monkman (Senior Director, Ads in AI Experiences), Christine Turner (Managing Director, Measurement, Data and Audiences), and Josh Moser (Advertising Platform Senior Director), in conversation with host Ginny Marvin.

What: A 41-minute Ads Decoded episode recorded backstage at Google Marketing Live 2026, covering the technical rationale behind the event's major announcements - including why AI Max and Performance Max are now required for new AI Mode ad formats, how the AI Brief provides brand control within automated creative generation, how 1-Click Creative Testing works in Asset Studio, and how Qualified Future Conversions and Attributed Branded Searches function as always-on measurement signals.

When: The episode was published on May 28, 2026, three days after Google Marketing Live 2026, which took place on May 20, 2026.

Where: Recorded at Google's Bay View campus in Mountain View, California, and published on the Google Ads YouTube channel at https://www.youtube.com/watch?v=EtGd3DpI5uA

Why: The episode provides additional technical depth on announcements that the two-hour keynote covered at a higher level, addressing practitioner-level questions about why the new ad formats require specific campaign types, how measurement metrics like Attributed Branded Searches differ from existing tools, and what the shift away from keyword-centric campaign management means for advertiser workflows.

Share this article
The link has been copied!