Microsoft Advertising published a blog post this week arguing that AI assistants, not human browsers, now decide which products get seen, and the company is selling three tools it says businesses need to keep up. Shirley Chen, writing on the Microsoft Advertising blog, frames the shift as bigger than the move to mobile. The post doubles as marketing copy for Microsoft’s own commerce stack, so its numbers and product claims deserve attribution to the company making them, not treatment as neutral findings.
The post puts AI at the top of the growth table among retail traffic sources over the last holiday season, up 693% year over year, and AI and AI agents influenced 20% of global retail sales during the holiday window, an amount Microsoft estimates at roughly $262 billion. A cited survey puts the share of shoppers who will expect retailers to offer AI-assisted buying inside a year at 72%, and Microsoft says visitors arriving from AI convert 42% better than visitors arriving from other sources. Microsoft attributes these figures to footnoted sources in the original post but does not publish the underlying methodology in the blog text itself, so the numbers should be read as company-reported rather than independently verified.
The behavioral claim at the center of the piece is that a shopper can describe what they want, such as a gift under $50 for a niece interested in art, and an AI assistant will narrow a full catalog down to a short list before the shopper ever visits a retailer’s site. That mechanic matters for search teams because it moves the competitive battle upstream: if an assistant excludes a product from its shortlist, the retailer never gets the chance to compete on-site at all.
Microsoft’s proposed remedy is a three-part playbook built entirely around its own products. First, it wants merchants feeding a complete, structured catalog into Microsoft Merchant Center, arguing that partial or outdated product feeds, the kind historically built just for ad campaigns, will cause agents to skip a brand’s products rather than merely rank them lower. Second, it is promoting Copilot Checkout, which lets a shopper complete a purchase without leaving a Copilot conversation, alongside Brand Agent, a customizable on-site assistant Microsoft describes as a personal shopper tuned to a brand’s voice and catalog. Third, it points to AI Visibility insights inside Microsoft Clarity, a feature meant to show which AI platforms cite a brand, which queries associate with it, and how it compares to competitors, explicitly tied to generative engine optimization (GEO), the practice of optimizing content for LLM-driven answers.
The post does not include any measurement of ranking or conversion impact independent of Microsoft’s own products or Microsoft’s own holiday-season data, and it does not disclose how the 20% sales-influence figure was calculated. Microsoft also does not say whether merchants using Google Merchant Center or other structured-data pipelines see comparable exclusion risk, leaving open whether the “complete catalog” requirement is specific to Microsoft’s ecosystem or a broader agentic-commerce standard. Readers should treat the playbook as Microsoft’s sales case for its own advertising and checkout products, built on top of a genuine and separately documented shift toward AI-mediated shopping.
Search and retail teams evaluating this playbook before the holidays should separate the underlying trend, which multiple independent trackers have also reported, from the specific fix Microsoft is selling, and should audit their product feed completeness across every AI shopping surface they actually use, not just Microsoft’s.
Shirley Chen, “How businesses win when AI does the shopping,” Microsoft Advertising blog, August 21, 2026.
