Google’s ad business generates roughly $175 billion a year. That number is both its greatest strength and its most uncomfortable pressure point right now. As AI-powered search changes how people actually find things — fewer blue links, more conversational answers — the question of where ads fit has been hanging over every Google earnings call for the past two years. At Google Marketing Live 2026, the company finally showed its hand. New Gemini-native ad formats for Search are coming, and they’re built around the idea that the AI answer IS the shopping experience, not something you navigate away from.
Why Google Had to Move Now
Cast your mind back to 2023. Google’s initial response to the ChatGPT moment was Search Generative Experience (SGE), later rebranded to AI Overviews. The pitch was simple: give users a synthesized answer at the top of the page. The problem advertisers immediately spotted was equally simple: if users get what they need without scrolling, where do the ads go?
That tension has defined Google’s ad product roadmap ever since. Early AI Overviews kept ads below the fold or slotted in awkward positions that felt bolted on. Click-through rates for some commercial queries reportedly softened. Advertisers got nervous. Analysts wrote the usual breathless takes about Google’s moat eroding.
The thing is, Google never lost control of the situation — it just took time to engineer a more elegant answer. What’s being announced now is the result of roughly 18 months of testing inside Google’s ads infrastructure, combined with the multimodal reasoning capabilities that the agentic Gemini era brought to the surface at I/O 2026. The timing isn’t accidental.
What’s Actually New: Breaking Down the Formats
Google is introducing several distinct changes here, and it’s worth separating them because they serve different parts of the funnel.
AI-Organized Search Ads
The most significant structural change is that Gemini now organizes ad units contextually within AI Overview responses. Instead of a static block of sponsored links sitting above or below the AI answer, ads are embedded in a format that mirrors the conversational structure of the query. Ask something like “best running shoes for wide feet under $150” and the AI Overview might surface a structured comparison — with sponsored product cards woven into the response at points where commercial intent is highest.
This is different from what Meta does with its Advantage+ placements or what Microsoft has tried with Copilot ads in Bing. Google’s version uses Gemini’s understanding of the full query context, not just keyword matching, to decide where a commercial placement makes sense. The ad doesn’t interrupt the answer — it’s supposed to be part of it.
Direct Offers Expansion
The Direct Offers pilot, which Google first tested with a handful of retail partners, is now expanding more broadly. The mechanic is straightforward but genuinely useful for shoppers: retailers can surface exclusive deals, promo codes, or bundle offers directly inside the Search interface without requiring the user to click through to a landing page first.
Think of it as the coupon clipping experience meeting the Google Shopping graph. A user searching for a particular laptop model might see a Direct Offer badge showing a $75 instant discount from a verified retailer, redeemable at checkout. The offer data is pulled from the retailer’s feed in real time.
Creative Generation with Gemini
Google is also pushing harder on AI-generated ad creative. Advertisers using Performance Max campaigns can now prompt Gemini directly inside Google Ads to generate headlines, descriptions, and image assets that are tailored to specific audience segments. This isn’t entirely new — Google has offered asset generation for a while — but the quality jump from Gemini’s latest models is meaningful, and the workflow is faster.
Key features rolling out across these announcements include:
- Contextual ad placement inside AI Overviews — ads matched to conversational query intent, not just keywords
- Direct Offers with real-time retailer deal feeds — discount visibility without a click-through required
- Gemini-powered asset generation inside the Google Ads UI for Performance Max campaigns
- Visual shopping ads in AI-organized results — product imagery pulled from Google Merchant Center, displayed within the AI response card
- Improved brand safety controls for advertisers who want to restrict placement next to specific AI-generated content types
What About Pricing and Availability?
Google hasn’t published a separate rate card for the new formats — they’re sold through existing Performance Max and Search campaigns, which means advertisers bid into them through the same auction they’re already running. Direct Offers expansion is rolling out in the US first, with international markets to follow later in 2026. The Gemini creative tools inside Google Ads are available now in beta for advertisers with access to Performance Max.
What This Means for Advertisers, Shoppers, and Google’s Rivals
Here’s the thing about Google’s position: it controls both the search surface AND the ad auction. That’s a structural advantage no competitor currently matches at scale. Microsoft’s Bing Ads inside Copilot is interesting but Bing’s market share outside of enterprise Windows deployments remains thin. Amazon Ads is powerful within its walled garden but doesn’t own a general-purpose search surface. Meta knows more about people but less about what they’re actively searching for right now.
The new Gemini-native formats effectively close the loop Google needed to close. The worry was that AI Overviews would train users to get answers without engaging with commercial content. What Google has done instead is make commercial content part of the answer layer itself. That’s a smart defensive move.
For advertisers, the implications are mixed. On one hand, contextual placement in AI responses should improve relevance — you’re reaching someone at the exact moment they’re reasoning through a purchase decision, not just keyword-matching to a query. On the other hand, the auction dynamics are more opaque. When Gemini decides where in an AI Overview to surface a sponsored product, advertisers have less direct control over placement than they did with traditional top-of-page bidding. Expect agencies to spend considerable time in the next six months figuring out how to optimize for these new surfaces.
For shoppers, Direct Offers is genuinely the most user-friendly change here. Not having to hunt through landing pages for a promo code that may or may not work is a real quality-of-life improvement. Whether it makes Google Search feel more like a shopping app is a different question — and one some users won’t love.
The competitive pressure on comparison shopping sites and affiliate publishers is real. If Google is surfacing deals and product comparisons directly in the answer layer, a site like Wirecutter or a deals aggregator loses its position in the user journey. That’s a pattern we’ve seen play out before with featured snippets and local packs — Google gradually internalizes the value that used to flow to third parties.
It’s also worth watching how this intersects with Google’s broader agentic push. Gemini 3.5’s real-world action capabilities point toward a future where the AI doesn’t just show you an offer — it completes the purchase. Google stopped short of announcing that here, but the infrastructure being built with Direct Offers and real-time retailer feeds is exactly what you’d need for that to work.
Key Takeaways
- Google’s new ad formats are built natively into Gemini’s AI Overview layer — not appended to it
- Direct Offers lets retailers surface real-time discounts without requiring a landing page click
- Gemini generates ad creative inside Google Ads UI, now available in beta for Performance Max
- Advertisers gain relevance but lose some placement control in the new contextual auction
- Comparison sites and affiliate publishers face structural pressure as Google internalizes more of the commercial discovery journey
- All new formats run through existing Performance Max and Search campaign auctions — no separate buy required
Frequently Asked Questions
What are Google’s new Gemini search ad formats?
They’re ad units designed to appear natively within Google’s AI Overview responses in Search, using Gemini to match ads to the conversational context of a query rather than relying purely on keyword triggers. They include contextual product placements, visual shopping cards, and the expanded Direct Offers feature for real-time retail discounts.
How does Direct Offers work for shoppers?
Retailers submit deal and promo data through a feed integrated with Google Merchant Center. When a shopper searches for a relevant product, verified offers from participating retailers can appear directly in Search results with discount details visible upfront — no click-through required to see the deal. The program is expanding in the US throughout 2026.
Do advertisers need a new campaign type to access these formats?
No. The new ad placements are served through existing Performance Max and Search campaigns. Advertisers don’t need to set up a separate campaign type, though optimizing bids and creative for the new contextual surfaces will likely require updated strategies as the formats mature.
How does this compare to what Microsoft is doing with Copilot ads?
Microsoft has been experimenting with sponsored placements inside Bing’s Copilot responses, but the scale difference is significant — Google processes vastly more commercial queries daily. Google also has tighter integration between its ad auction, Merchant Center product data, and the AI response layer, giving it more signal to work with when deciding where and how to surface a commercial placement.
Google’s ad business has survived every previous search format shift by owning the transition rather than being disrupted by it. The Gemini-era formats feel like the same playbook applied to a harder problem — and if the Direct Offers expansion and contextual placements perform as designed, I wouldn’t be surprised to see transaction-layer features follow within the next product cycle. The managed agents infrastructure Google has been building makes that next step look less like speculation and more like a matter of timing.
