Your Product Is Live. But Who Does AI Think It Belongs To?
Something I've been looking at more closely over the past few months is GTIN ownership. Not whether brands have GTINs on their products because most do. I’ve been looking at who the GTIN is registered to at the GS1 level. This may sound dull as dishwater but it’s in teasingly important in an AI landscape.
The answer is not always the brand.
In some cases, it's a logistics partner. In others it's a factory or a production supplier who assigned identifiers from their own GS1 prefix during manufacturing and nobody downstream ever questioned it. And in at least one case I've looked at recently, it was a completely unrelated fashion brand with no obvious commercial connection to the product being sold.
I want to explain why this matters now in a way it didn't before. And I want to give ecommerce specialists a practical way to find out where they stand before they find out the hard way.
Before I explain anything, do this
Go to the GS1 UK GTIN lookup (https://www.gs1uk.org/our-services/data-services/verified-by-gs1/search-verified-by-gs1)
Search the barcode from your hero product.
Read who it says licensed that number.
If the answer is your brand, great. Keep reading anyway, because the rest of the trust chain still needs checking.
If the answer is someone else, this article is going to be uncomfortable in a useful way.
The importance of GTIN in 2026
Most ecommerce teams think about GTINs in one of two ways.
- They're a Google Merchant Centre requirement. Something you need to avoid feed disapprovals.
- They're a supply chain identifier. Something that lives in your ERP and tells warehouse systems what a product is.
Both of those framings are too narrow for what's happening in 2026.
AI shopping agents like Perplexity, ChatGPT Shopping, Google's AI Mode don't browse your website the way a human does. They don't read your brand story or respond to your creative. What they do is resolve product identity using structured identifiers. They take a query, find matching products, and then determine which source to treat as authoritative.
The GTIN is the first anchor point in that resolution process. It's the identifier agents use to establish what a product actually is and, critically, who the originating brand is.
If multiple listings exist for the same product such as your brand website, three stockists, and a marketplace listing or two the agent needs to decide which source to trust. And the GTIN registry is one of the most upstream, most structured data sources it can reference to make that call.
If the GS1 record points to the wrong entity, you start the trust resolution process already losing ground.
What I found on a fashion brand's hero product
A few weeks ago I was working through a structured data audit on a well-known premium fashion label. Strong identity. Great brand voice. The kind of brand where product integrity is central to the whole commercial proposition.
I ran their hero product GTIN through the GS1 lookup as part of the standard checks I do before drop season.
What came back stopped me for a moment.
The GTIN was registered to a completely separate fashion company. Different brand. Different category. Different customer base entirely. No visible connection to the fashion label whose product carried this barcode.
I went back and checked a few more SKUs from the same brand. Same result on several of them.
This wasn't a data entry error on a single product. The GTINs across a meaningful portion of the catalogue were registered to an entity that had no apparent relationship to the brand selling the products.
My best guess is that these identifiers originated with a manufacturer or production partner who used their own GS1 prefix during production. The brand adopted the GTINs operationally because they worked for logistics and feed purposes and nobody ever questioned who actually held the licence.
What this looks like from an agent's perspective
Here's the situation an AI shopping agent is now navigating when it encounters one of these products.
A customer asks for a recommendation. The agent finds the product across several listings. It attempts to resolve which source is the legitimate brand origin.
It checks the GTIN. The GS1 record returns a different company name entirely.
It checks the brand's own site. The Organisation schema there says one thing. The GTIN registry says another.
It checks the stockist listings. Some of them carry the same GTIN. None of them surface any information that resolves the conflict.
The agent doesn't have a way to call someone and ask. It follows the data. And the most upstream, most structured piece of data it has found says this product belongs to an entity the brand has never heard of.
If anybody who has been following my posts know - then AI shopping agents hate uncertainty.
I'm still working out exactly how much weight different agents place on GS1 registry data versus on-site schema markup in their resolution logic. But I'd argue the direction of that uncertainty doesn't matter much. Any conflict between upstream registry data and brand-level schema is a problem. The agent shouldn't have to choose between them. When it does, you've already lost control of how your product is being represented.
Why fashion brands end up here more than other categories
The GTIN ownership problem is not unique to fashion. But fashion has a set of structural dynamics that make it particularly common.
Wholesale distribution runs deep in this category. A brand might have 30 or 40 stockists, each maintaining their own product listings, many of them carrying the brand's GTINs without any data governance from the brand itself. The brand doesn't control what those listings say. In some cases the brand doesn't even know exactly which identifiers are circulating out there and under what attribution.
The manufacturing route I described above is probably the most common root cause. Factories that produce for multiple brands often use their own GS1 prefix for operational convenience. The identifiers work. Products ship. GTINs get absorbed into the brand's feed without anyone registering them to the brand's own GS1 account. This can run for years before anyone notices.
And there's a third dynamic that I think is under appreciated. Third-party stockists frequently have more complete, more structured product data than the brand's own site. Better attribute coverage, richer availability signals, sometimes more review data. In an agent evaluation, a stockist listing can score higher on data completeness than the brand's DTC listing even though the brand built the product and should be the authoritative source.
That's the scenario where the brand loses the agentic sale not through negligence but through a data completeness disadvantage. They did everything else right. Their GTIN ownership problem just tipped the balance somewhere they couldn't see.
The liabilities that go beyond a lost sale
I want to be direct about the broader consequences here because I think most ecommerce teams haven't fully mapped them.
Product card misattribution - AI shopping surfaces generate product cards using structured data. Google AI Mode, ChatGPT Shopping, Perplexity Shopping. When the GTIN resolves to the wrong entity, the product card may surface with incorrect or conflicting brand information. The brand loses control of its own representation at the moment a purchase decision is being made. And because most brands aren't monitoring how agents are representing their products, this can run for months without anyone noticing.
Authenticity signal confusion - A GTIN registered to a third party on a product sold under a different brand name is structurally indistinguishable from a counterfeit or grey market listing. From a data fidelity perspective, the signals look the same. More sophisticated AI shopping agents will increasingly perform authenticity resolution as the stakes of agentic commerce grow. A brand with misattributed GTINs fails that check not because the product is inauthentic but because the identity infrastructure suggests it might be.
Legal exposure most brands haven't considered - If a GTIN licensed to another entity is used commercially on your product, questions arise about how that arrangement came to exist and what liability attaches to it. If a product is misrepresented in a commercial channel as a result of that misattribution, who carries the risk? Most brands haven't had this conversation with their legal team. Most don't know the conversation is necessary.
The three layers of the trust chain and where this fits
I've written before about the trust layer that AI shopping agents use to determine which source to treat as authoritative for a given product. For those who haven't read those pieces, the short version is this.
There are three structured signals agents weight when resolving product identity and brand ownership.
The first is GTIN registry ownership at GS1. This is as upstream as it gets. It's one of the highest-trust external references an agent can consult on product identity, and it sits outside your control once a product is in market unless you actively register the correct ownership.
The second is Organisation schema on your domain. This is how your brand establishes itself as a legal entity, not just a retailer. When your Organisation markup is coherent and the entity details match your GS1 registration, the agent has a consistent signal pointing in the same direction. When they contradict each other, you have a trust conflict baked into your infrastructure.
The third is product card coherence across surfaces. The degree to which your product's identifier, brand name, offer data, and imagery are consistent across your site, your Merchant Centre feed, your agentic storefront, and your stockist listings. Inconsistencies don't just reduce trust. They create what I'd describe as ghost products. These are listings that agents find but can't confidently assign to an authoritative source, and sometimes drop from the consideration set entirely.
The GTIN ownership problem undermines the first layer. And when the first layer is unstable, the other two can't compensate for it.
What to check before the next drop
This isn't a comprehensive audit. It's the minimum a head of ecommerce should be able to answer before a significant product launch.
Pull your hero product GTIN. Run it through the Verified by GS1 tool. The entity name that comes back should match the brand entity in your Organisation schema. If it doesn't, you have a mismatch that needs resolving before you have meaningful exposure in agentic channels.
Check your Organisation schema on the site. Look at the legal name, the URL, the sameAs links. These should point to active, brand-owned profiles. I've audited brands where sameAs links were pointing to Instagram handles that were changed during a rebrand two years ago and nobody had updated the schema. That's a broken trust signal nobody on the team knew about.
Pull the top two or three stockist listings for your hero product and compare them to your own. Look at attribute completeness, not design. If their structured data is richer than yours, you have a data completeness disadvantage in agent evaluation.
And query your own product in ChatGPT, Perplexity, and Google AI Mode. Search by product name and category. Look at which domain the agent cites when it surfaces the product. If it's a stockist and not your site, the trust resolution is already going the wrong way.
None of this takes long. The issue is that it lives in the gap between technical SEO and ecommerce operations, so it tends not to be owned clearly by anyone. Which is usually why it doesn't get done.
The longer argument
GTINs are product-level identifiers. Organisation schema is brand-level. But what both are building toward is something more durable than either on its own: entity trust. This is the game changer for 2026 onwards.
Entity trust is the degree to which an AI shopping agent has learned to treat your brand as the authoritative source for your category. It isn't established quickly. It compounds across drops, across content, across structured signals that all point consistently to the same entity. And it's much harder to damage than most brands assume, provided you build it properly from the start.
The brands I've seen get this right are not necessarily the ones with the biggest budgets or the most sophisticated technical teams. They're the ones where someone decided to treat product identity infrastructure as a brand asset rather than a logistics function.
The GTIN ownership problem is one of the more invisible failure modes in that infrastructure. It doesn't show up in GA4. It doesn't trigger a feed disapproval. It just quietly shapes how agents understand your brand, and by extension how much of the agentic consideration set you actually occupy.
Start with the GS1 lookup. It takes about thirty seconds. And if the answer is wrong, you'll want to know now rather than after the next drop.