Three separate updates landed this week, a new Shopping ad test inside AI Mode, an AI search playbook from Microsoft Advertising, and a candid podcast admission from Google about keywords, and they all point at the same underlying shift. AI is not adding a new layer on top of paid search. It is stress-testing whether the structure advertisers already have can survive being compressed into a conversation.
Google starts testing Shopping ads directly inside AI Mode
Vidhya Srinivasan, Google’s VP/GM of Ads & Commerce, confirmed in a blog post that Google is testing a Shopping ad format that recommends products based on a user’s query inside AI Mode, with similar formats also being tested outside retail, in travel. The framing was the more interesting part: Srinivasan said Google is not just “bringing ads to AI experiences in Search,” it is “reinventing what an ad is.”
That line matters because AI Mode compresses the shopping journey. Instead of scanning a results page, people ask, refine, and compare inside a single conversation. Being present is no longer about ranking or grabbing the first paid slot, it is about being one of a small number of options the AI experience is willing to surface while someone is actively comparing.
For Shopping advertisers, that raises the stakes on feed quality specifically. If AI Mode is assembling recommendations from product attributes, price, and availability, thin or inconsistent feed data becomes a visibility problem, not just a performance one. And because AI Mode likely surfaces fewer commercial slots than a traditional results page, those remaining slots get more contested; eligibility and relevance may end up mattering more than bid strength.
Reaction across PPC circles has mostly been “this was inevitable,” with the practical questions front and center. Thomas Eccel of AdSea Innovations flagged that the format will run through existing Shopping and Performance Max setups rather than requiring a new campaign type, which is the detail most advertisers wanted first. Elsewhere, commentary compared Google’s approach to how ChatGPT and Claude are monetizing their own AI products; Martin GroBe of Suchmeisterei GmbH pointed out that Google can test monetization inside AI Mode, which users already perceive as an evolution of Search, without touching the ad-free experience of “pure” Gemini, and can watch how ad strategies play out on Claude and ChatGPT before committing further. Attribution in a discovery-first, conversational flow is still an open question nobody has fully answered yet.
Microsoft’s new AI search playbook is really a paid-media brief in disguise
Microsoft Advertising published an updated AI search playbook aimed at how AI-powered search and assistants are reshaping discovery, with more focus on being understood and trusted inside generated answers than on ranking links. It also draws a clear line between traditional SEO and what it calls generative engine optimization, with guidance on writing content structured enough for AI systems to interpret confidently.
On the surface this reads like an SEO resource, but paid teams have two reasons to pay attention. First, Microsoft is treating AI-driven discovery as the current operating environment, not a future one, and is telling marketers to adjust accordingly. Second, the “structure” theme connects directly to paid performance: brands get pulled into AI answers because their information is clear, consistent, and machine-readable, not because the ad copy is clever. That is the same direction the whole industry is moving in, fewer manual levers, more dependence on clean inputs, and feed quality and landing page clarity are exactly the kind of inputs that determine outcomes. Microsoft’s playbook is effectively an instruction to start treating information architecture as performance infrastructure, a point that lines up with the broader case for why Microsoft’s AI ad strategy deserves more attention from PPC managers.
The response has been notably positive, largely because the guide explains mechanics instead of selling hype. International SEO consultant Aleyda Solis, who contributed to it, praised Microsoft for publishing something search marketers can actually use. Microsoft Ads liaison Navah Hopkins highlighted its value for paid media specifically, including budget focus, landing page insight, and communication style. The recurring reaction across the industry was some version of “finally, someone wrote this down plainly.”
Google: “keywords are a means to an end,” not the strategy itself
On the latest Ads Decoded episode, Brandon Ervin, Google’s director of Product Management for Search Ads, told host Ginny Marvin, Google Ads liaison, that in 2026 keywords are “a means to an end,” not the endpoint of strategy. His argument: advertisers should start from business goals and go-to-market approach, with keywords acting as a thematic layer supporting that, not the foundation itself. The two also covered the continued shift toward semantic matching, why exact match still earns a place for tight control, and how query matching keeps evolving through backend improvements.
This lands hardest on anyone who built their career around granularity, tight ad groups, tight keyword lists, maximum manual control. That approach earned its reputation honestly; SKAG structures and skepticism toward broad match made sense for the system that existed when they were built. But user behavior and Google’s own intent-matching have both moved on, and Google saying “keywords are a means to an end” out loud is really validation of a shift many advertisers have already been forced to make. Segmentation still has a job to do; it should just be justified by a real difference in intent, landing page, or creative approach, not preserved as a default habit left over from an earlier system.
Reaction split cleanly into two camps: one hearing “less control,” the other hearing “confirmation of how the system already behaves.” Brad Geddes, co-founder of Adalysis, thanked Ervin and Marvin for confirming something practitioners had long suspected, that Google uses conversion data from across an account for bidding and optimization. Performance marketing specialist Alexandr Stambari pushed back gently, arguing that in competitive niches like ecommerce and B2B lead generation, segmentation by intent, margin, and query type still protects against averaging out performance and missing real growth signals. Marvin responded by reiterating Ervin’s actual point: use segmentation where it’s justified, and ground both analysis and structure in business goals rather than habit. The fact that Google chose to have this conversation publicly, aimed at marketers rather than engineers, suggests it expects more advertisers to be rethinking account structure this year, a pattern that fits with how Marvin has recently been pushed to clarify how AI search is reshaping conversion measurement more broadly.
The common thread: structure is now the leverage point
None of this week’s news is really about a single new placement or a single podcast quote. It’s the same message from three directions: AI is not layering something new onto search, it is testing whether the structure already underneath your campaigns holds up once discovery happens inside a generated answer instead of a results page.
Shopping ads in AI Mode make feed quality a visibility requirement, not just a performance lever. Microsoft’s playbook makes clear structure and trustworthy inputs the deciding factor in whether a brand gets surfaced at all. And Google’s own team is telling advertisers that keyword architecture is one input among several, not the strategy itself. Teams that already run an integrated search brief across SEO, PPC, and content are better positioned here, because clean feeds, clear content, and campaigns aligned to real intent are exactly the leverage that pays off once fewer placements are carrying more of the weight. Where that foundation is missing, AI-driven search environments tend to expose the gap quickly rather than quietly.
Frequently asked questions
Will Google’s Shopping ads in AI Mode require a new campaign type?
Based on early reaction from advertisers who saw the announcement, the format appears to run through existing Shopping and Performance Max setups rather than requiring a separate campaign type, though full eligibility details are still emerging.
Does Microsoft’s AI search playbook apply to paid media teams?
Yes. While it’s framed around discovery and structured content, its core advice, clear and consistent information architecture, directly affects whether a brand gets surfaced in AI-generated answers, which is increasingly relevant to paid visibility as well.
Does “keywords are a means to an end” mean advertisers should abandon segmentation?
No. Google’s own team clarified that segmentation still matters where it’s grounded in a real difference in intent, margin, or query type. The shift is about anchoring structure to business goals rather than defaulting to granularity as the strategy itself.
How does AI Mode change what “visibility” means for Shopping advertisers?
It shifts the competition from ranking position to feed strength and relevance. With fewer visible commercial slots inside a conversational experience, clean and complete product data matters more than aggressive bidding alone.