Who Owns Being Found by AI: SEO, Ecommerce or the Catalog?
SEO says it's content, ecommerce says it's data, merchandising forwards PDFs. How to split being found by AI into three jobs with owners, and the role most teams lack.

Somebody runs the test from ask an AI assistant for a sofa, the brand isn't in the answer, and the question lands in a meeting: whose job is this? The SEO lead says the pages need better content. The ecommerce manager says the product data is thin and that's not their team. Merchandising says the supplier sent what they sent. Everyone's right, which is why nothing happens.
The job doesn't fit any existing box because it's three jobs that have to agree.
The three jobs
Knowing the facts. Seat depth, fabric composition, frame construction, packed dimensions, lead time. Buyers know some, suppliers' spec sheets have some, customer service has learned the rest. This is product knowledge.
Holding the facts as data. Attributes on the product record, per variant, consistent across the range. This is catalog work. In many retailers nobody owns it as a role; it's done by whoever uploaded the products, from whatever the supplier sent.
Publishing the facts where machines read them. The page template, the structured data, the feeds. This is web and SEO work.
Being found by AI needs all three, in sequence. The facts have to be known, then held, then published. Most arguments about ownership are two teams each pointing at the step that isn't theirs.
Why SEO can't own it alone
SEO owns publishing. They can make the template show a specifications table and add attributes to structured data. They can't fill the seat depth attribute on 2,000 products, because they don't know the seat depth and it isn't their data. An SEO team asked to own AI visibility will do what they can at step three, find step two empty, and report that the content is thin. Which it is.
Why ecommerce can't own it alone
Ecommerce owns the storefront and the trading, and usually the product upload. They can see the data is thin. They often don't control the attribute and family definitions, they don't own the spec sheets (merchandising does), and they don't own the template (web does). An ecommerce team asked to own it may fix priority pages by hand without solving the catalog-wide problem.
Why merchandising can't own it alone
They know the products and hold the spec sheets. They don't work in the catalog day to day, and they don't think of a spec sheet as a data source; it's a document that justified a buying decision. Asked to own AI visibility, they'll forward the PDFs.
The split that works
Give each step to the people who already have what it needs, and make the record the handoff point.
Merchandising supplies the facts. The spec sheet, the fabric data, the answers to the questions on the questions shoppers ask AI assistants about sofas. Not as PDFs forwarded, but as inputs to the record: either they fill attributes at onboarding, or the supplier file is read into attributes by the catalog and they review it.
The catalog owner holds them. Somebody owns the attributes and families: what fields exist, which family uses them, what units apply, which are required and how variants inherit. They run the bulk filling across the range, using whatever tooling reads supplier files and flags gaps, and they review the sample before it runs. In smaller retailers this is the ecommerce manager wearing a second hat. In larger ones it's a product data role. Either way it has to be a named person, because the record is the thing everything else reads.
Web and SEO publish them. Template reads attributes into a table and an FAQ, structured data and feeds carry them, content is delivered in the HTML, crawlers allowed. Once that's set up, new attributes the catalog adds can reach the page with little further web work, as long as the template reads attributes generically rather than from a fixed list.
Customer service feeds back. The questions they're asked that the page doesn't answer become new attributes. This is the loop that keeps the record shopper-shaped.
The one new responsibility
Everything above exists in some form already except the middle: the record as the source of truth, owned, with attributes defined for the questions shoppers ask and filled across the range. That's the role many furniture retailers are missing, and it's why the SEO team and the ecommerce team keep handing the problem back and forth. Neither of them owns the thing in between.
It's also the role that pays off beyond AI assistants. The same record feeds the marketplace listings, the comparison tools, the in-store app and the website. The existing piece on who should own furniture product content makes the broader case; AI visibility is the newest reason to settle it.
How to tell it's working
Not by a dashboard, at first. By a trace: pick a fact, follow it from merchandising's spec sheet to the record to the page to the assistant's answer, as in from field to page. When the trace completes without anyone emailing anyone, the ownership is right. When it stops, the step it stops at tells you whose job is undefined.
For the whole argument, see being found by AI.
The meeting question
Not "who owns AI visibility" but "who owns the product record". Answer that and the other two teams know what they're supplying to and publishing from.




