Tapestry made Coach and Kate Spade products purchasable inside Google Search, including AI Mode, and the Gemini app in the United States on Sept. 16, moving two major fashion brands from conversational discovery to native checkout. Eligible shoppers can select a product, tap an inline buy button and approve a single-item purchase through Google Pay without being redirected to a brand website.
The transaction flow uses the Universal Commerce Protocol, or UCP, an open standard designed to let retailers, AI agents and payment systems exchange product and checkout information. Tapestry said every purchase requires explicit consumer approval and that it plans to expand the capability to more markets.
The launch is a practical test of a larger question for commerce leaders: What happens when the channel that recommends a product also closes the sale?
Discovery and conversion now happen in one interface
Retailers have spent years optimizing a funnel that moves people from advertising or search into a website or app. The visit mattered because the brand controlled the product page, merchandising, analytics, cross-selling and the path into loyalty.
Google’s AI shopping model compresses that sequence. A shopper can ask a conversational question, receive a product recommendation and complete an eligible purchase in the same environment. The reduction in steps could improve conversion, particularly for customers who already use Google Pay and know which item they want.
Tapestry is positioning the launch as a response to changing consumer behavior. The company’s point is credible: product discovery has already moved from stores and magazines to social feeds, creators, marketplaces and search. Conversational AI adds another front door, one that can understand intent expressed in natural language rather than a fixed set of filters.
Google says UCP keeps the retailer as merchant of record even when checkout happens on a Google surface. That distinction preserves responsibility for the transaction and customer relationship. It does not, by itself, guarantee that a retailer receives the same depth of browsing and behavioral data it would collect on its own site.
Analysis: convenience changes the marketer’s scorecard
Native AI checkout creates an attribution problem before it creates a branding problem. A successful interaction may produce a sale with little or no traditional site traffic. Marketing teams that use sessions, product-page views and last-click web conversions as core measures will see an incomplete journey.
The better scorecard will connect AI-surface visibility, product eligibility, recommendation frequency, checkout starts, completed purchases and subsequent customer value. Brands will also need to distinguish between sales that are incremental and sales that merely moved from a website into Google’s interface.
That measurement challenge changes content strategy. Product feeds can no longer function as basic catalogs. Attributes such as material, dimensions, color, use case, care instructions, availability and compatibility become inputs to a conversational recommendation system. Missing or ambiguous data can make a product effectively invisible even when its creative campaign is strong.
Tapestry said it built a protocol-agnostic layer so it can move into additional AI environments. That is the more durable strategic choice. Retailers should avoid hard-coding commerce operations to a single assistant when discovery may fragment across Google, ChatGPT, marketplaces, retailer agents and interfaces that have not launched yet.
The direct relationship becomes the tradeoff
Every reduction in friction transfers some control to the platform removing it. When checkout stays inside Gemini or Search, the brand has less space to tell a broader story, recommend a coordinated item, invite a customer into editorial content or create the sensory cues associated with a luxury purchase.
That does not make native checkout a bad channel. It means assortment and journey design require intention. Straightforward, high-confidence purchases may fit an embedded flow. Higher-consideration products, personalized services and relationship-building experiences may still benefit from a branded destination.
Consumer trust will also depend on clarity. Shoppers need to understand the seller, final price, return rules, shipping terms and the exact moment when an assistant stops recommending and starts transacting. Tapestry’s requirement for explicit approval is an important boundary, but it is only one part of a trustworthy purchase experience.
What commerce teams should build next
Retail leaders do not need to treat this launch as proof that every transaction will move into an AI assistant. They should treat it as evidence that checkout is becoming portable.
The immediate work is operational: improve product data, establish a clean inventory and pricing feed, map consent and customer-service responsibilities, and make order, payment, return and loyalty systems accessible through standards rather than one-off integrations. Marketing, technology, legal and customer experience teams need to set those rules together.
They should also test the economics by product type. A shorter funnel can lift conversion while reducing opportunities for cross-sell, first-party data collection or membership enrollment. The winning design will balance immediate convenience with lifetime customer value.
Coach and Kate Spade are not abandoning their stores or sites. They are adding a transactional layer where discovery increasingly begins. For other brands, the lesson is less about copying one integration than preparing for a market in which the purchase button can appear almost anywhere.
Sources: Tapestry; Google; S&P Capital IQ via MarketScreener.