Agentic commerce visibility: how to get picked when AI shops for your customers

In classic e-commerce search, a shopper types a query, scans results and clicks through to a store. In agentic commerce, the shopper describes a need to an AI assistant, which researches options, narrows them to a shortlist and, increasingly, completes the purchase without the shopper visiting the store's site.

The traffic is already measurable. Adobe reported that visits to US retail sites from generative AI tools rose 693.4% year on year over the 2025 holiday season, and that those visits converted 31% better than other traffic.^[1] AI traffic is still a small share of the total, but it is growing fast and is concentrated in categories such as video games, toys, appliances, electronics and personal care.^[1]

Two announcements turned AI shopping from research into transactions.

• ChatGPT Instant Checkout (September 2025). OpenAI and Stripe launched in-chat purchasing for US ChatGPT users, starting with US Etsy sellers and with Shopify merchants announced as coming. It runs on the Agentic Commerce Protocol (ACP), an open standard the two companies developed, which merchants can adopt with other payment providers too.^[2]
• Google's Universal Commerce Protocol (January 2026). Google announced UCP at the National Retail Federation show as an open standard for agents to handle discovery, carts, payment and post-purchase steps.^[3] It was developed with Shopify, Etsy, Wayfair, Target and Walmart, and Google said it would power checkout on eligible product listings in AI Mode and the Gemini app.^[3] Google's developer documentation describes integration through APIs, the Agent2Agent protocol and the Model Context Protocol.^[4]

Retail assistants are a separate track: retailers' own assistants answer shopping questions inside their stores and apps, and some AI visibility vendors now sell coverage of them.^[5] Details change quickly, so check each platform's current merchant documentation before planning around a specific feature.

Checkout protocols decide how a purchase happens. They don't decide which product gets recommended. That still comes down to what the engine finds and trusts when it builds its shortlist.

1. Being in the answer at all. If an engine doesn't name your brand for "best {category} for…" questions, a checkout button won't help. This is the core of AI visibility for e-commerce.
2. Accurate, complete product data. Names, variants, prices, stock, shipping and returns stated consistently on your site and in the feeds and listings each platform uses.
3. Third-party evidence. Reviews, comparisons, retailer pages and editorial roundups that engines cite when they explain a recommendation.
4. Readable pages. Product and policy pages that AI crawlers can fetch and that state key facts in plain text. Our AI crawler checker shows what major bots can access.
5. Clear policies. Returns, warranties and delivery times, stated plainly. An agent comparing options on a shopper's behalf needs them.

Measuring shopping visibility is the same discipline as any AI visibility, with a shopping-specific question set. Because answers vary from one ask to the next, a single check is not a measurement.

• Write shopping-intent questions at each stage: discovery ("best running shoes for flat feet"), comparison ("{brand} vs {competitor}"), and purchase ("where to buy … with free returns").
• Ask in every market you sell in, in the local language, because recommendations differ by country.
• Sample repeatedly and report how often your brand and products appear, with the number of answers and a range.
• Track the sources. Note whether engines cite your own pages, retailers, marketplaces, reviewers or forums. That tells you where to act.
• Watch AI referral traffic in your analytics, separating AI assistants from other referrers, and compare its conversion rate with other channels.

Several AI visibility vendors now sell shopping-specific tracking, such as product-level visibility in ChatGPT shopping results.^[7][8] If that's what you need, ask how they collect shopping results, in which countries, and how many answers sit behind each figure.

Two habits make the plan work. First, keep the question set fixed between measurements, so a change in the numbers reflects a change in the answers rather than a change in what you asked. Second, report every result with its sample size and range. Shopping answers vary by run, by country and by season, and a holiday spike in one market can look like a lasting win if you only check once.

Expect the protocols and assistants to keep changing. Treat platform-specific features as experiments with their own measurement, and keep investing in the fundamentals that every assistant reads: accurate product pages, clear policies and credible third-party coverage.

For retail and DTC brands, our e-commerce solution page shows how Glossy Data measures product and brand visibility by market.