By Parcel Perform

Here is an uncomfortable experiment for any eCommerce brand leader: open ChatGPT and ask it to recommend brands in your own category, in your own market. Then do the same in Gemini and Perplexity. Now count how many times you appear.

If your brand is like most of the household names we analysed, the answer is: roughly one time in three. The other two answers went entirely to someone else, and no dashboard you currently own recorded the loss. This matters more with every passing month.

AI-referred shoppers arrive pre-qualified, pre-convinced, and pre-filtered. The filtering happened inside the answer, before your analytics ever saw a visitor. Which raises the only question that matters: were you in the answer at all?

How often does AI actually mention your brand?

In May 2026, Parcel Perform analysed AI Commerce Visibility data covering 152 eCommerce brands across six retail categories – luxury, fashion, beauty, sportswear and specialty retail in the US, UK and Germany, measuring how often each brand appears when ChatGPT, Gemini and Perplexity answer real shopping prompts in its category.

These are not obscure brands. Every one of them is a top-ten name in its category and market. And yet:

● The median brand appears in roughly 1 in 3 relevant AI answers.
● More than a quarter appear in 1 in 5 or fewer.
● Three in four brands are absent from more than half of the AI answers where they should be a natural recommendation.
For a category leader, absence is not the exception. It is the norm.

The individual numbers are starker than the averages. On one major AI engine, UK shoppers asking for fashion recommendations hear about H&M in fewer than 7% of answers. Levi’s and Ralph Lauren both register under 8% on the same engine.

In UK luxury, Chanel, Louis Vuitton, Prada and Gucci each appear in just 1 in 8 answers,  brands that have spent a century and billions building recognition, reduced to a rounding error in the channel growing fastest.

Why doesn’t your Google ranking carry over?

Because rank and recall are different problems, and almost every brand is only measuring the first one.

Rank is the question your SEO programme answers: when a shopper searches, where do you appear on the list?

Recall is the question AI assistants force: when a shopper asks for a recommendation, does the model think of you at all? A search results page shows ten blue links; an AI answer names three or four brands and moves on. There is no page two.

This is why strong search performance creates a false sense of security. AI assistants don’t read Google’s index, they build recommendations from their own source ecosystems: editorial coverage and directories, community discussion, brand-owned content and structured data.

A brand can dominate search rankings while barely existing in the sources AI actually reads. Your SEO agency cannot see this, because it is not their job to. It has, until now, been nobody’s job.

The result is a blind spot with a specific shape. Of the brands in our dataset tracked across all three AI engines, more than half are effectively invisible – appearing in 1 in 5 answers or fewer – on at least one of them. Most have no idea which one.

Who is winning the answers you’re missing?

Every answer that skips one brand names another, and the winners are rarely who category logic predicts.

In UK fashion, on the same engine where H&M sits below 7%, Next appears in 65% of answers and Marks & Spencer in 48%. In UK luxury, Burberry appears in half of all relevant answers while the global houses cluster at 12.5%.

The pattern repeats across categories and markets: visibility on the AI shelf doesn’t track store count, media budget or brand valuation. It tracks the depth and structure of what AI can read about you – reviews, community discussion, editorial presence, and machine-readable content on your own domain.

That last point deserves emphasis, because it is the part fully within a brand’s control. Adobe’s site-level analysis found the same structural gap from the other direction: on average, around a third of retail homepage content is not readable by AI models at all, with product pages scoring worse.

Shoppers are pouring into a channel that literally cannot see large parts of what brands publish.

And this shopper doesn’t behave like a browser. Adobe’s survey found 79% of consumers using AI for shopping feel more confident in their purchase because of it.

When an AI assistant names three brands and yours isn’t among them, you haven’t lost an impression. You’ve lost a decision.

What should retail leaders do about it?

Not panic-publish content, measure first. Three moves, in order:

1. Establish your baseline. Before anything else, find out how often each AI engine mentions you, in each market you trade in, against the competitors it actually names (which are frequently not the ones on your battlecard).

The lowest-effort starting point to look at AI visibility indexes, which can show the most AI-visible brands per industry and market – helping you see whether you’re on the list at all, and who is there instead of you.

2. Locate the gap, not just the score. A visibility number without a source breakdown isn’t actionable. The useful question is why an engine skips you: thin editorial coverage, no community footprint, or product and delivery content on your own site that models can’t parse. Each cause has a different owner and a different fix.

3. Treat your own website as an AI source, not just a destination. The fastest-moving lever is machine-readable content on your own domain – structured product data, explicit delivery and returns information, content that answers the questions shoppers actually ask assistants. It’s the one source ecosystem where no third party stands between you and the model.

The shelf you can’t see

For twenty years, digital shelf strategy meant winning a ranked list a human would scroll.

The AI shelf works differently: it is short, it is spoken with confidence, and it is assembled from sources most eCommerce brands have never audited.

The traffic flowing from it is already the highest-converting in retail, and the brands being named are collecting a compounding advantage with every answer.

None of this requires a leap of faith to act on. It requires a willingness to look – the measurement now exists (AI Commerce Visibility tracks brand presence, ranking and sentiment across all three engines at product-category level), and checking takes less effort than a single SEO audit.