Something is off about the way brands are talking about agentic commerce.
A Pulse Conversation
The conversation tends to start at the top of the stack. Which AI agents matter. Which tools to plug in. Whether ChatGPT or Gemini will drive more traffic. What “AI visibility” actually means.
These are real questions. They’re just not the first ones.
The brands likely to do this well aren’t the ones with the boldest AI roadmaps. They’re the ones whose operating models can actually move when a decision is made. Most can’t.
Brand visibility vs product visibility
A panel at Pulse recently, with Ian from Axel Arigato, Chris from Glara.ai (a GEO platform tracking AI visibility for ecommerce down to the SKU) and Lauren from Pion33rs & Operator Experience (ex-Oh Polly, now in advisory), made the point clearly.
Brand-level visibility scores are mostly useless for ecommerce. It’s nice your brand gets mentioned. It doesn’t move revenue. What moves revenue is your products getting pulled into the carousels and recommendations AI agents generate inside ChatGPT, Gemini and the rest. That’s product-level visibility, and it’s a different problem.
There was a number attached. Enriching the product catalog, the descriptions, attributes, structured facts AI models can actually reason about, gets you 30 to 40 percent of the way there.
The interesting part is the other 60 to 70 percent.
That sits in the operational layer. Whether product data is consistent across every system it lives in. Whether the brand can describe its own products the same way in its PIM, its storefront, its marketplaces, and its 3PLs. Whether the operating model can prioritise this work alongside everything else competing for attention.
Unglamorous questions. Also the ones that determine whether the rest works.
Where accountability sits
Axel Arigato came up as an example of doing this well. The brand had appointed one person internally as a “shepherd”. A specialist to guide the AI work across the business.
What mattered was not the title. It was where the accountability sat.
The shepherd is not the owner of AI at Axel Arigato. AI is a standing agenda point at the exec committee every two weeks. There are open monthly forums from junior to senior. The shepherd guides. The business owns it.
Most brands have inverted this. They hire or appoint someone senior, and quietly let the rest of the business consider AI to be that person’s problem. It becomes a silo. When the silo conflicts with anything else competing for resource, the silo usually loses.
Axel Arigato had structured it the other way round. The work moves because the whole exec is accountable. The specialist is there to make sure the conversation has the right inputs.
The misalignment underneath
A simpler version of the same point. The head of ecommerce and the IT director have to share a vision and a priority list.
If they don’t, they end up in conflict over priorities. The whole thing stalls.
The brands that struggle with operational and data transformation aren’t usually struggling because the technology is wrong. They’re struggling because ecommerce wants one thing, IT wants another, finance has a third priority, and the founder is making whichever call is loudest that week. Technology doesn’t fix that. It makes the cost of it more obvious.
A high-functioning AI strategy sits downstream of a high-functioning operating model. If ecom and IT are not aligned now, the agentic conversation will not align them. It will expose that they were never aligned to begin with.
The part that does not get talked about
The agentic commerce conversation, as it currently exists in most rooms, is heavily weighted toward strategy and tooling. Which makes sense. Strategy and tooling are visible. They have vendors attached.
The operating model work has none of that. It is internal. It is slow. It does not have a clear vendor category. It is rarely the most exciting part of any quarterly plan.
It is also the work that determines whether everything else lands.
The brands that get there first are unlikely to be the ones with the loudest AI strategy.