March 30, 2026

Search engine optimization is becoming irrelevant for a growing share of commerce. Gap CTO Sven Gerjets told a Shoptalk 2026 audience that brands must now show up not just in search results but in AI-generated answers, introducing the term "Answer Engine Optimization" or AEO. eMarketer's coverage of the keynote captures the full scope of what Shoptalk revealed about where agentic commerce stands in practice. Agent-managed shopping journeys, from discovery through checkout, are now the central theme in retail AI. But execution is ahead of consumer adoption. eMarketer analyst Sarah Marzano delivered the sharpest caution of the conference. "There's much less clarity on how quickly, or whether, adoption, transactions, and monetization will follow. It's never been more important to anchor decisions in data and to distinguish between preparing for change and reallocating meaningful capital ahead of proof."

Gap is not just talking about AEO. It is building for it. Chief Marketer reports that Gap has consolidated roughly 200 AI models into Google's Gemini Enterprise across all four brands: Gap, Old Navy, Banana Republic, and Athleta. The company established a dedicated Office of AI in 2024 and has since deployed AI across fit and sizing, inventory allocation, and generative product design. Whether these deployments have moved metrics is not yet public, but Gerjets described fit and sizing as the wedge use case, tackling a historically high-friction point that drives returns and cart abandonment. The strategic logic is that structured, AI-ready product data solves a measurable consumer problem today while simultaneously making Gap visible to AI agents tomorrow. When Gemini or ChatGPT recommends outfits that match a user's constraints, Gap wants its catalog metadata to be rich enough and current enough to surface in those results.

Sephora is taking a different path to the same destination. The same eMarketer Shoptalk coverage reports that Sephora has launched a dedicated app inside ChatGPT where customers can access loyalty rewards and member benefits within the AI channel, with plans to enable in-app payments and checkout. This is one of the clearest early examples of a retailer embedding its own logic and data directly into a GenAI platform rather than simply being indexed by it. The distinction matters. A retailer whose loyalty program, pricing rules, and personalization engine operate inside the AI conversation retains far more control than one whose products merely appear in a comparison table.

Google is building the infrastructure to make these retailer integrations work at scale. The Digital Runway reports that Google is constructing AI commerce rails designed for exactly the kind of plug-and-play integration Gap and Sephora are pursuing. Honcho's analysis of the shift from click-to-checkout to in-AI transactions frames this as the largest structural change in retail since the rise of search-driven commerce. Google's Universal Commerce Protocol connects AI assistants to real carts, real catalogs, and real customer identities, enabling checkout directly inside AI Mode and the Gemini app while keeping retailers as the seller of record. Google Pay and Google Wallet handle stored payment and shipping details, with PayPal integration coming. The architecture preserves transactional ownership for retailers while compressing the path from discovery to purchase into a single conversational interface.

The gap between these first movers and the rest of the industry is becoming measurable, and the problem is not just infrastructure. Mirakl's partner survey reveals a stark AI commerce readiness divide that goes beyond catalog quality. Most retailers cannot even tell which queries trigger their products in LLM-generated responses or where they rank relative to competitors in AI-mediated shopping. This monitoring blindspot is the new competitive vulnerability. In traditional search, retailers had analytics dashboards showing keyword rankings, click-through rates, and conversion attribution. In AI agent commerce, equivalent visibility tooling barely exists. Mirakl identifies marketplace operators like Ulta Beauty as better positioned because marketplace operations produce normalized data by default, but even they lack comprehensive AI search monitoring. The retailers investing in measurement infrastructure for AI channels, not just data feeds into them, are the ones building durable advantage.

Lowe's VP of customer marketing Amanda Bailey introduced another dimension at Shoptalk that complicates the pure-infrastructure narrative. "Creators give us credibility," Bailey said, per the same eMarketer Shoptalk recap. "Consumers are going to creators for ideas, inspiration, tips. They're not just going there for 'What products do I buy?' But, 'How do I use them? Is it worth it? Does it fit into my life?'" eMarketer's Sky Canaves adds a critical observation from the same coverage. Even when shoppers receive AI-generated product recommendations, they still conduct their own research outside AI, looking for trust signals and validating content. The emerging pattern is that AI agents narrow the consideration set while creators and authentic human content validate the final choice. Meta announced new tools at Shoptalk for creators to tag and recommend products directly within content, with expanded affiliate partnerships across Facebook and early testing on Instagram. Retailers who invest only in agent-facing data without maintaining the creator ecosystems that drive trust are optimizing for one half of a two-part decision process.

CrewAI's enterprise survey, published via Yahoo Finance, quantifies the organizational momentum behind this shift. Eighty-one percent of organizations say agent adoption is either fully scaled or actively expanding. They report time savings of 75%, operational cost reduction of 69%, and revenue generation impact of 62%. The average share of workflows automated by agents today is 31%, and organizations expect to add another 33 percentage points in 2026 alone. These numbers explain why Gap and Sephora are moving now, not waiting for proof that others have moved first. But Marzano's Shoptalk warning remains the most important sentence of the week. Preparing for change and reallocating meaningful capital ahead of proof are not the same thing. Gap is preparing. Most retailers are still deciding whether they need to.