AI Shopping Agents Expose Retail’s Messy Product Reality
AI shopping agents are software assistants that search, compare and recommend products on behalf of consumers, and their growing role in online discovery is exposing how disorganized, duplicated and inconsistent most retail product catalogs are, forcing brands and platforms to rebuild retail data management around standardized, machine-readable information instead of human-only storefront design. This pressure is no longer theoretical. On June 16, senior commerce and fashion executives met in New York for a dinner marking the launch of an agentic storefront designed to put AI at the centre of the customer journey. The timing is clear: 53 percent of people who used generative AI for search in one quarter also used it to help them shop, while global cart abandonment still hovers around 70 percent. AI is now a mainstream shopping tool, but it is colliding with decades of neglected product catalog organization—and the agents are winning.
Shopify’s Catalog: Teaching Machines to Read Products Like People
The most concrete signal that data chaos is no longer acceptable is Shopify’s new Catalog system, which uses large language models to organize merchant product data into a format AI shopping agents can understand and compare. Retailers often describe the same item in inconsistent ways—one listing protein powder as a single product with flavor variants while another splits each flavor into separate products—so agents struggle to identify when two listings refer to the same thing. Catalog tackles product deduplication by grouping related listings under a Universal Product Identifier, helping agents recognize when different pages are selling the same underlying product. This is less a clever feature than a blunt admission: if platforms do not clean up product data at the source, agentic commerce will stall. As Shopify puts it, "When we get this right, merchants’ products show up exactly where they should across every agent," turning AI-driven traffic into real orders.
Agentic Commerce Is Rewriting the Rules of Product Catalogs
The rise of agentic commerce—the idea that AI agents sit at the centre of the shopping journey—is forcing brands to rethink the very structure of their product information. Product discovery is shifting from organic search and storefront scrolling to AI agents doing that work on a user’s behalf, and Shopify’s data shows AI-driven traffic has grown eight times year over year, with orders from AI-powered searches up nearly 13-fold. That kind of growth makes messy catalogs a strategic risk, not an operational nuisance. Retailers can no longer rely on crawlers to piece together their data; they must push clean, structured product information directly to agents and protocols built for machine use. Standards like the Universal Commerce Protocol are gaining attention because they promise a common way for AI agents to interact with merchants, from cart creation to payment. In plain terms: if your products cannot be understood by an agent, they may as well not exist.
Fashion Brands Learn: Loyalty Needs Data, Not Hype
At the agentic storefront launch dinner, the conversation among executives from labels including Ganni, Versace and Tod’s Group underscored a hard truth: loyalty is now a data problem as much as a branding one. Leaders spoke about keeping customers "feeling special" and preserving the human connection even as AI mediates more of the journey, but they also acknowledged that understanding who customers are and what they want demands better technology and cleaner information. Agentic commerce does not erase brand values; it exposes whether those values are supported by coherent product data. A sustainability-focused brand that cannot clearly surface its materials, sizing and variations to AI agents will struggle to prove its claims at the point of recommendation. As one executive framed it, these immersive, agent-led shopping experiences bring the brand closer to the customer than ever before—if the underlying catalog is organized well enough to show up at that moment.
What Comes Next: From AI Noise to Clean, Agent-Ready Catalogs
The industry’s next phase will separate brands that treat agentic commerce as hype from those that treat it as an information problem. Many executives now talk less about abstract protocols and more about owning where search starts and controlling the experience once a shopper arrives. Shopify’s Catalog and Botify’s planned agentic catalog product are early attempts to make product data agent-ready at scale, and Botify is already piloting its approach with large retailers. Meanwhile, brands are tweaking their own sites—experimenting with markdown and other formats that make product information easier for AI agents to read. Swap is preparing an agentic storefront for designer Paul Smith, where the agent will speak in the designer’s own voice, signalling that brand personality and catalog cleanliness must work together. The conclusion is blunt: AI shopping agents are now powerful enough that retailers no longer control whether they matter. They control only how readable their product catalogs are when the agents come looking.
