Agentic commerce: what it means for your store when AI buys

40% of Spaniards would already let an AI agent choose their clothing or tech purchases for them, according to Sopra Steria’s From Click to Agent 2026 survey, run by Cluster17 across 8,400 consumers in eight countries in February 2026. Accenture confirms it from another angle in its Consumer Pulse 2026, surveying more than 25,000 people across 16 countries: 42% of brand-loyal shoppers would let an AI agent swap their usual brand for a better option if the assistant suggests one. Agentic commerce — buying through an assistant that researches, compares suppliers and executes the purchase on the customer’s behalf — is no longer a future hypothesis for Spanish retail. It’s a behaviour a meaningful share of your customers is already willing to adopt. The question for any store, physical or online, is direct: if that agent is comparing suppliers, does your catalogue stand any chance of being picked?

What agentic commerce actually is

This isn’t a chatbot answering questions about your catalogue — it’s an assistant that acts: it searches across several stores, compares price and availability, and in many cases can go on to complete the payment. The two technical protocols building this infrastructure — the Agentic Commerce Protocol backed by OpenAI and Stripe, and Google’s Agent Payments Protocol — define how an agent checks a store’s stock, price and terms, and how it authorizes a payment on the person’s behalf. It’s the same logic we covered with AI-generated search, taken one step further: the agent doesn’t just cite your site, it uses it to buy.

Spaniards aren’t rejecting the idea, they’re conditioning it

Sopra Steria’s data adds an important nuance: 93% of respondents want the agent to propose options, with the final call staying with the person. This isn’t blind delegation — it’s an upstream filter that shrinks the number of stores a customer ever consciously considers. Accenture adds another relevant figure: 51% of Spanish consumers would tell their agent which brands to consider, and 42% would accept the agent swapping their usual brand if it finds something better. In practice, the purchase decision is shifting from a customer’s manual search to an automated pre-filter. If your store doesn’t clear that filter, you never even reach the stage where the customer decides.

What happens if your catalogue isn’t machine-readable

A shopping agent doesn’t browse your site the way a person would — it needs structured data, clear pricing and verifiable availability to include you in a comparison. If your product page is a photo, an unmarked price and a description copied from the supplier — the most common problem we see in retail catalogues — the agent has nothing to compare and moves to the next store. This isn’t a brand-visibility issue; it’s whether the system can read you reliably enough to include you in a purchase decision it’s executing on someone else’s behalf.

How to prepare your store so an agent compares you and picks you

You don’t need to rebuild the entire e-commerce site to get started:

• Product and Offer structured data (Schema.org) on every page, with price, currency, availability and shipping terms explicitly marked up, not just visible as text.
• Real-time product feeds. An agent that compares stock and finds a stale price or a false “out of stock” drops the store from future comparisons.
• Product pages with your own attributes — material, sizing, compatibility, real delivery time — instead of the manufacturer’s generic description. This is exactly what our SEO-for-AI websites service does: generating unique pages from your catalogue data instead of duplicating supplier content.
• Verifiable, citable return and warranty policies, since an agent buying on someone’s behalf penalizes uncertainty as much as price.
• Correctly marked-up aggregate reviews (AggregateRating) — one of the strongest signals when an automated system is deciding between two similarly priced stores.

Where to start without losing focus

Don’t try to tackle all five at once. If your catalogue runs to more than a few dozen SKUs, prioritise structured product data first — it’s what makes or breaks whether an agent can compare you at all — then move to your best-selling product pages. The rest can follow in batches, the same approach we recommend for any applied AI project in retail: measure the impact of each change before rolling it out across the full catalogue.

The takeaway

Agentic commerce already has real adoption data in Spain, not just forecasts: four in ten consumers would let an AI agent choose part of their purchases, and half would directly tell it which brands to prioritize. For a store, this doesn’t change what you sell — it changes how you need to present it to stay a visible option once the one doing the comparing isn’t a person browsing, but an agent executing instructions. Catalogues with structured data, reliable pricing and original product pages start ahead; the ones still relying on a photo and a copied description will keep competing for fewer and fewer decisions.

Want to know if your catalogue is ready for a shopping agent to include you in the comparison? Get in touch: the first consultation is free and we reply within 24 hours.