Ranked #1, Recommended Never: How Products Get Into AI Answers

A good LinkedIn post, tested against the data

A version of this argument is making the rounds: there are now three ways to show up in search, SEO, AEO and GEO, and most ecommerce brands only work on the first. Rank higher and win clicks is the old model. Be the answer and get recommended is the new one.

It is a useful framing, and it is also the kind of claim that deserves a stress test before you rebuild a budget around it. We have written about the labels themselves before, including where AEO, GEO and SEO differ. This article does something narrower. It takes each claim in the popular version, checks it against current evidence, and adds what that version leaves out for online stores.

Claim one: ranking no longer guarantees a place in the answer

This one holds up, and the data is stronger than the slogan. Ahrefs, which sells SEO software and so has an interest in how visibility is measured, examined 863,000 keywords and 4 million AI Overview URLs. Search Engine Journal reported that the share of cited pages ranking in the top 10 fell sharply.

The direction is consistent even where the numbers differ: a ranking is an entry ticket, not a seat. And being cited is not the same as being clicked. Pew Research found people clicked a traditional result in 8% of visits when an AI summary appeared, versus 15% without one.

Claim two: AI recommends products by name, from a short list

Also true, and OpenAI has been unusually specific about how its version works. In ChatGPT, a product appears in the shopping carousel when ChatGPT judges it relevant to the user's intent.

Product results are selected independently by ChatGPT and are not ads, nor influenced by any OpenAI partnerships.

That line comes from OpenAI's help page on shopping in ChatGPT, which says the system considers price, reviews and ease of use, and draws on "structured metadata from first-party and third-party providers" plus other third-party content. Its shopping research feature, launched in November 2025, builds a personalized buyer's guide by reading product pages across retail sites, and OpenAI says the results are organic.

The consideration set really is small. When a shopper asks for a moisturizer for dry skin, they typically see a handful of products, not ten pages of results. If yours is not one of them, you are not ranked lower. You are simply absent.

Claim three: what the web says about you shapes the answer

The popular version says genuine mentions across reviews, Reddit, niche publications and roundups influence AI recommendations. The evidence broadly supports it, with one twist.

Ahrefs studied 75,000 brands across ChatGPT, Google AI Mode and AI Overviews, and The Next Web reported the Spearman correlations with AI brand visibility.

The twist is that video led, not forums. YouTube mentions beat every other metric, and backlinks correlated only weakly. The researchers were careful to note that correlation does not equal causation: brands people already talk about may simply be the brands that sell well.

Citation data points the same way for retail. Comscore data reported by Search Engine Land found the most cited sources in AI answers for retail searches in June 2026 were YouTube at 23%, Amazon at 14%, Reddit at 12% and Wikipedia at 11%. Reddit matters, as we explored in what Reddit's AI citations teach us about GEO, but it is one voice among several.

The popular version leaves out your product data

For an online store, the most concrete lever is not content or mentions. It is the structured data you hand to the AI platforms directly, and almost every SEO, AEO and GEO explainer skips it.

OpenAI now accepts product feeds from merchants. In March 2026, it said merchants can share product feeds and promotions through its Agentic Commerce Protocol, via third-party providers, or through Shopify Catalog, so their catalogs are fully represented in ChatGPT. Google is doing the same in Merchant Center, where it announced in January 2026 dozens of new data attributes for conversational shopping, such as compatible accessories and substitutes.

Access matters too. OpenAI's crawler documentation says sites that block OAI-SearchBot "will not be shown in ChatGPT search answers." A blanket robots.txt rule meant to keep out AI bots can block OAI-SearchBot along with the training crawlers, and quietly remove you from ChatGPT's answers.

Then there is readability. Adobe scored what AI models can actually read on retail sites and found product pages were the weakest, with only 66% of their content readable, compared with 75% for homepages. Adobe sells analytics and content software, so it has a stake in this problem. The usual culprit is simple: prices, specs and reviews that only appear after scripts run, which many crawlers never execute.

Why this is worth the effort

AI-referred traffic is growing fast, and it converts. Adobe's April 2026 report, based on more than 1 trillion visits to U.S. retail sites, found AI-referred traffic grew 393% year over year in Q1 2026. In March 2026, those visitors converted 42% better than non-AI traffic, a reversal from March 2025, when they converted 38% worse.

The purchase still happens on your site. OpenAI stepped back from in-chat checkout in March 2026 after finding that people researched products in ChatGPT but rarely bought there, repositioning ChatGPT as a discovery layer that sends shoppers to merchants. That makes the recommendation itself the prize. If you are named, the shopper comes to you ready to buy.

Where the three labels help, and where they do not

The SEO, AEO and GEO split is a good way to divide the work. It is a poor way to divide the budget, because the layers share one foundation. Google says there are no additional requirements to appear in AI Overviews or AI Mode beyond sound SEO, and pages that cannot be crawled or indexed are invisible to every layer at once.

Think of it as one system with three outputs. Our earlier piece on the three jobs of modern SEO includes a one page test for whether a page can rank, answer and be cited.

What to do this quarter

The popular post is right that ranking is no longer the whole game. It undersells how concrete the new game is. The brands that get recommended are the ones people talk about, whose products machines can read, and whose catalogs sit inside the AI platforms directly. None of that is mysterious, and all of it is work you can start this quarter.