Generative Engine Optimization for Retail

Generative Engine Optimization for retail is the work of getting products, brands and stores named, described correctly and made purchasable in AI answers to shopping questions. Retail is the vertical where GEO has the most measurable commercial stake and the fastest-moving plumbing: citation behaviour, product feeds and checkout protocols all changed materially between late 2025 and autumn 2026.

AI Traffic Is Small, Growing and Now Converts

Adobe's analysis of more than one trillion visits to US retail sites (reported by TechCrunch, April 2026) found traffic from AI sources up 393% year over year in the first quarter of 2026. The more important change was quality. In March 2025 AI-referred visits converted 38% worse than other traffic; by March 2026 they converted 42% better, with revenue per visit 37% higher. In Adobe's accompanying survey of more than 5,000 US consumers, 39% said they had used AI for online shopping. Similarweb data reported in June 2026 adds an indirect effect: people shown a brand in a ChatGPT recommendation were 2.5 times more likely to visit its site within seven days, and 55.9% of those visits arrived through branded search rather than a link in the answer. Much of the value of being recommended never shows up as AI referral traffic.

Two Engines, Two Shopping Philosophies

Where shopping answers get their sources depends heavily on the engine. BrightEdge (December 2025; tens of thousands of e-commerce prompts over the holiday season) found retailers made up roughly 4% of Google AI Overview citations against roughly 36% of ChatGPT's. Google's most-cited sources were YouTube, Reddit, Quora, Amazon and Facebook; ChatGPT's were Amazon, Target, Walmart, Home Depot and Best Buy. BrightEdge's reading is that Google treats the overview as research and leaves conversion to the shopping units below it, while ChatGPT has to do both jobs in one answer.

A second dataset complicates the ChatGPT half of that picture. LLM Pulse's rolling tracker of about 391,000 citations on non-branded US shopping questions (read in October 2026; vendor data) puts all major retailer and marketplace domains combined at 2.9% of citations, 6.05% on ChatGPT and 1.92% in AI Overviews, and reports that Reddit and YouTube together out-cited every big-box retailer. The two studies used different prompt sets and periods (deal and product prompts against generic shopping questions), so they are not directly comparable. What they agree on is direction: generic "which should I buy" questions are answered mostly from reviews, video and community discussion, and retailer pages gain ground as the prompt gets closer to a specific product or deal.

Ranking well in classic search is a weak proxy. BrightEdge (February 2026) found only 13.4% of e-commerce AI Overview citations also ranked in the organic top 10, up from 2.9% a year earlier, with 61.5% coming from outside the top 100. That is consistent with query fan-out, in which the engine cites pages that rank for sub-questions rather than for the shopper's original query. Seer Interactive's 2026 data shows AI Overviews on only 5% of transactional queries but 95.4% of comparison queries, which locates the contest in the "X vs Y" and "best X for Y" stage.

Feeds Decide Eligibility; Content Decides Selection

Being purchasable inside an assistant is a data-integration problem. OpenAI's product feed specification (as published in October 2026) accepts JSONL, CSV and Google-compatible feeds and requires an identifier, title, description, URL, brand, seller name, image, availability and price for discovery, with the instruction to "keep current price and availability up to date." Google's equivalent layer is Merchant Center. A feed makes a product eligible to appear; it does not make an engine prefer it. On-page structured data is a separate question, and Ahrefs' May 2026 controlled study found that adding schema markup "produced no major uplift in citations on any platform."

Agentic Checkout: What Actually Shipped

OpenAI launched Instant Checkout in ChatGPT in late September 2025, starting with Etsy and built on the Agentic Commerce Protocol it co-developed with Stripe; it supported US, single-item purchases. In March 2026 OpenAI said that "Instant checkout is moving to apps, where purchases can happen more seamlessly." Retail Systems reported that Shopify's president had told investors only about a dozen merchants were selling through AI tools at that point, and that Instacart, Target, Expedia and Booking.com had built ChatGPT apps. The feed specification still lists checkout as "a separately enabled integration."

Google went the other way. In January 2026 it introduced the Universal Commerce Protocol, co-developed with Shopify, Etsy, Wayfair and Target and endorsed by more than 20 companies including Visa, Mastercard, Stripe, Best Buy and The Home Depot, with a Buy button on eligible listings in AI Mode and the Gemini app using Google Pay (9to5Google, January 2026). The practical lesson for retailers is to treat agentic commerce protocols as unsettled, keep feeds clean for every surface, and assume most AI-influenced orders will still complete on the retailer's own site.