In sportswear and footwear, some of the fastest-moving AI in 2026 is not inside the business - it is the shopper. As AI agents begin making buying decisions on consumers’ behalf, the brand that wins is the one whose product data they can read. That is a quieter, more fundamental shift than any campaign.

Most of this series has looked at AI inside the enterprise - redesigning functions, running agents across operations and finance. This piece looks the other way, at AI on the far side of the transaction: the customer.

Because in fashion, sportswear and footwear, one of 2026’s defining trends is that a growing share of buying decisions will be made not by a person browsing, but by an AI agent shopping on their behalf. McKinsey has put the prize at up to a trillion dollars of newly orchestrated revenue in US retail alone by 2030, and several times that globally. And the categories the analysts expect agents to reach first are, notably, exactly this sector’s: shoes, sportswear, outerwear and everyday basics.

What agentic commerce means for a brand

The mechanics are simple to picture. A consumer tells an assistant what they want - “a waterproof trail-running shoe, size nine, under £120, good for wide feet” - and the agent searches, compares options across brands, and increasingly completes the purchase. The shopper may never see your product page, your hero image or your seasonal campaign. The agent reads structured facts and makes a call.

The uncomfortable part: the agent cannot see your brand

Conventional retail is won with marketing, imagery and the feeling a brand creates. An agent buying on someone’s behalf does something colder: it weighs the product data it can find - attributes, specifications, price, availability, reviews, fit. If your product data is thin, inconsistent or unstructured, your product is invisible or misjudged at the exact moment of choice, however strong the brand behind it.

An AI shopper cannot admire your campaign. It can only read your data.

For a house that has invested years in brand, that is an uncomfortable idea. It is also the one worth internalising early, because the brands adapting now will compound an advantage while others are still optimising imagery for an audience that is, increasingly, not the one making the decision.

This is a product-data problem, not a marketing one

Winning in agentic commerce is, underneath, a data-readiness exercise. It means product data that is rich, accurate, structured and machine-readable - every attribute an agent might filter on, from materials and waterproofing to weight, fit and sustainability credentials - held consistently across brands, regions and channels, and exposed in a form agents can consume.

Some call the emerging discipline “generative engine optimisation”. Strip the jargon and it is product-data quality and accessibility, aimed at a machine reader rather than a search engine. And the early signal is that this traffic is real and high-intent: analysis of shoppers arriving via generative-AI tools shows them spending markedly longer on-site and bouncing less than the average visitor. The agents are not a threat to route around; they are a channel to be legible to.

The same data does double duty: compliance

Here is the part that turns a cost into an investment. This work is not only about discoverability. The European Union’s digital product passport, becoming mandatory for clothing, footwear and accessories, will require exactly this kind of structured, traceable product data - materials, origin, sustainability and repairability information attached to each item.

In other words, the product-data foundation that makes you legible to an AI shopper is largely the same one that makes you compliant with incoming regulation. One investment, two returns - which is a far easier business case than either on its own.

Where the data lives - and why that is the hard part

The catch, for a multi-brand business, is that this data rarely sits in one place or one shape. It is scattered across the ERP, a product-information system, brand-by-brand spreadsheets and channel feeds, and it is often inconsistent between brands and regions - the same attribute described three different ways.

Making it agent-ready means unifying and standardising it, then exposing it through the interfaces agents use - increasingly the open protocols and storefront agents we covered earlier in this series. For a multi-brand house on SAP and a cloud data platform, the raw material is already there. The work is in the consistency and the plumbing, not in acquiring something new.

What a brand should do now

This does not require a moonshot. It requires starting:

• Audit how complete and consistent your product data is across brands, regions and channels - most houses are surprised.

• Identify the attributes that agents will filter on and that the digital product passport will require - they overlap heavily.

• Unify and standardise that data, and expose it in a machine-readable form.

• Start with your highest-velocity categories - the ones agents will reach first, which in this sector means footwear and sportswear.

• Treat it as a living asset with ownership, not a one-off clean-up.

The bottom line

The brands that thrive as agentic commerce grows will not be the ones with the loudest campaigns. They will be the ones an AI agent can read, trust and choose - and, not by accident, the ones ready for the regulation arriving alongside it. That is a product-data challenge before it is a marketing one. If you are weighing what agentic commerce means for your brands, and whether your product data is ready for it, that is a conversation we would welcome.