Google shopping became generative in stages, and the last stage was the fastest. What launched at the end of 2023 as an experiment called Search Generative Experience (SGE) is now AI Mode, a conversational layer that has passed 75 million daily active users and, since 2026, ships a checkout of its own. This piece traces how shopping on Google changed, what the click data says, and what retailers should do about it.
AI Mode is Google's Gemini-powered search experience, integrated into Google Search and the Gemini app, that treats a commercial query as a conversation: it interprets intent, generates relevant visuals, and shows shoppable products in a single flow.
Search stopped being a list of links
For twenty years Google Shopping worked the same way. You typed a query, you got ten blue links and a few sponsored product cards. That model is now a leftover.
In AI Mode the user describes what they want in natural language, and products appear inside the answer. No click-outs, no reformulated queries. In Google research with 4,773 participants run between October and December, users called it a more helpful way to compare brands and stores.
The rollout tells the same story. Google took AI Mode to more than 200 countries and territories, and the experience passed 75 million daily active users. EMARKETER estimates that 31.3% of the U.S. population will use generative search in 2026, while ChatGPT processes around 72 billion messages a month against Google's 417 billion searches. Google is still the giant. The center of gravity is moving anyway.
Behind the answers sits the Shopping Graph, which grew from 35 to more than 50 billion products, with 2 billion listings refreshed every hour. That scale makes conversational shopping answers possible.
Which features redefined product discovery?
AI-generated product images
Ask for a "metallic cherry-colored puffer jacket with wide pockets" and AI Mode generates photorealistic images matching the description, with shoppable real products underneath that visually match the render.
Apparel shows the benefit most clearly. Google's internal data says 20% of apparel queries run five words or longer, exactly the descriptive territory where keyword search struggled.
Shop Similar
Finding a product similar to an image used to mean a reverse image search. Now cards of matching real products appear automatically under each image, with price, seller, reviews, and availability. The match is visual and semantic, which makes image quality a de facto ranking factor.
The interface keeps adding layers, too. Analyst Glenn Gabe documented new "Sponsored stores" and "Quick results from the web" units inside AI Mode in March 2026.
Virtual Try-On, now with your own photo
Virtual Try-On started with women's tops, extended to men's with brands like Abercrombie, Banana Republic, and J.Crew, then made the decisive jump: you can upload your own photo and see a garment picked from billions of listings on yourself. Google reported noticeably higher engagement on Try-On enabled products, with more time on the listing and higher purchase likelihood.
What changed in 2026: agentic commerce
Two years ago Google took you to the product. Now it can buy the product for you. Vidhya Srinivasan, VP and GM of Google Ads and Commerce, used her third annual letter to the industry to call 2026 the year "agentic commerce moves from concept to operational reality."
Google is not alone in this race. OpenAI shipped Instant Checkout inside ChatGPT, letting users buy without leaving the chat; our practical guide to enabling Instant Checkout in ChatGPT covers the merchant side. Every major AI interface is building its own conversion channel.
Google's 2026 additions come in four pieces:
- Agentic checkout. You tap "track price" on a listing, set size, color, and budget; when the price drops, Google adds the item to the merchant's cart and completes checkout with Google Pay on your behalf. In January 2026 native checkout landed inside AI Mode and the Gemini app, with Target, Etsy, Wayfair, and other major retailers among the first partners. Walmart and Shopify are next.
- Universal Commerce Protocol (UCP). Announced at NRF 2026, an open standard developed by Google with Shopify, Etsy, Wayfair, Target, and Walmart, backed by more than 20 other industry players including American Express, Mastercard, Stripe, and Visa. It lets AI agents, retailers, and payment processors speak one language, and it is compatible with A2A, AP2, and MCP.
- Direct Offers. A pilot from January 2026: retailers place exclusive discounts directly inside AI Mode results, labeled "Sponsored deal," with Petco, e.l.f. Cosmetics, Samsonite, and Rugs USA among the first partners. In February Google extended the format with Shopping Ads inside AI Mode, shown at high-intent moments.
- Business Agent. A chat that connects the shopper with the brand inside search, like a virtual sales associate trained on the retailer's own data. Lowe's, Michael's, Poshmark, and Reebok are live at launch.
Protocols deserve one more line. Alongside UCP, WebMCP is making websites usable by AI agents directly; our explainer on how WebMCP is building the agentic web shows why e-commerce stacks should turn agent-first. For Shopify merchants, UCP onboarding is automatic: with the catalog synced to Merchant Center, products become eligible for AI Mode, Virtual Try-On, and agentic checkout without extra code. Google's official merchant guide lists the minimum feed attributes.
What does the click data say?
Look at the numbers without filters. A Seer Interactive study from September 2025 found a 61% drop in organic CTR on queries with AI Overviews, with paid CTR down 68%. A January 2026 analysis recorded that 93% of AI Mode queries produce no click to external sites.
A second data point shifts the picture. Growth Memo research from April 2026 shows that 88% of AI Mode users accept the AI's shortlist without checking elsewhere, against 56% in classic search. The AI's pick becomes the user's pick 74% of the time. Traffic goes down; the traffic that arrives converts better. Being cited by the AI is the new first position.
The commercial weight is already measurable. Salesforce attributed 20% of retail sales in the 2025 holiday season to AI, and Adobe Analytics measured AI referral traffic to retail sites growing over 1,200% year on year. Small absolute volumes, steep curve.
What should retailers do now?
The common mistake is filing all this under "one more channel." It is closer to new infrastructure for a growing share of transactions. Five priorities, in order:
- Keep the Merchant Center feed clean and complete. Descriptive titles, structured attributes, live prices and stock. A product missing from the Shopping Graph does not exist for the AI.
- Invest in image quality. High resolution, consistent backgrounds, faithful color and texture. The matching engine is visual; mediocre photos translate into invisibility.
- Write copy the way people talk. "Gift for someone who loves making fresh pasta" beats "premium kitchen utensils." Long, natural queries are exactly where generative search performs best.
- Build content with a GEO mindset. A Growth Memo analysis found that 4.8% of the URLs ChatGPT cites most are in-depth pages answering "what is it," "how to choose," and "how much does it cost" in one place. Citation rate is the new visibility KPI.
- Track AI referrals separately, and track citations too. Segment AI traffic in analytics and watch which SKUs get surfaced. Citations are less stable than rankings: 40% to 60% of cited sources change month over month in Google AI Mode and ChatGPT. That volatility is why we built RankWit to monitor AI visibility continuously; the plans cover ChatGPT, Gemini, Perplexity, and AI Overviews.
Brainlabs' Dan Connor compressed the stakes into one line: "if your product data isn't agent-readable, you're invisible". We agree, with one addition: readable comes first, chosen comes second, and both are measurable today.
FAQ
What happened to SGE?
Google folded the Search Generative Experience experiment into AI Mode, the Gemini-powered layer now integrated into Search and the Gemini app. The name changed, the direction did not: generative answers first, with shoppable products and, since 2026, agentic checkout built in.
What is the Universal Commerce Protocol?
UCP is an open standard announced at NRF 2026 by Google with Shopify, Etsy, Wayfair, Target, and Walmart, backed by more than 20 other companies. It gives AI agents, retailers, and payment processors a shared language for agentic transactions, compatible with protocols like A2A, AP2, and MCP.
Is it still worth optimizing for clicks?
Clicks still matter, but they are no longer the whole scoreboard. With 93% of AI Mode queries ending without an external click, citations inside the answer increasingly decide who gets seen. The users who do click out of AI Mode convert better, so fewer visits can still mean more revenue.
How does a small retailer get into AI Mode results?
Mostly through data. A complete Merchant Center feed, high-quality images, and natural-language product copy are the entry requirements. Shopify merchants get eligibility automatically once the catalog syncs with Merchant Center. From there, monitoring which queries surface your products tells you where to push.