From SEO to Agentic Commerce: Why Structured Data Is Becoming Your Next Growth Channel 

For more than twenty years, digital marketing has largely been built around persuading people to click. Search, paid media, CRO, UX, content, even social, all ultimately existed to bring a human visitor to your site and convince them to buy.

Increasingly though, the first thing evaluating your product isn't a person at all. That matters because AI agents evaluate products differently to humans. Humans are generally driven by emotion, but this isn’t a factor for AI agents. Before an agent can recommend your product, it first has to trust that it understands it.

For the first time, the primary "visitor" evaluating your product isn't always a person, it's a large language model acting on a person's behalf, querying APIs, parsing product feeds, and deciding what to surface, compare, or buy before a human ever sees a page. 

Two standards changed the rules

Two developments in the last twelve months made this shift structural rather than theoretical.

In September 2025, OpenAI and Stripe launched the Agentic Commerce Protocol (ACP), an open standard for how AI agents, merchants, and payment providers complete transactions inside conversational AI. Shopify enabled it for more than a million merchants on day one; Etsy, Walmart, and Instacart followed within weeks.

Then Google went further. Its Universal Commerce Protocol (UCP), co-developed with Shopify, Etsy, Wayfair, Target, and Walmart, and backed by Mastercard, Visa, Stripe, and more than twenty other partners, gives any AI surface a common language to discover merchants, understand what they support, and orchestrate an entire journey. Retailers publish a machine-readable profile of what they can do. Agents read it and decide, on the spot, whether you're worth transacting with.

Every major shift in digital commerce has changed where value is created. Search rewarded content and social rewarded attention, but agentic commerce is beginning to reward data quality. Increasingly, inclusion in AI-driven commerce isn't determined by how well you rank. It's determined by whether your data is legible enough for a machine to trust it. 

Feed engineering matters as much as SEO

The skill set that wins in agentic commerce looks nothing like the skill set that won in search. The difference is that humans respond to persuasion, while agents respond to confidence.  

A shopper might be influenced by a compelling product page or an eye-catching promotion. Search optimisation evolved to reward many of those same human-facing signals: persuasive copy, backlinks and page experience. Agentic discovery rewards something much more mundane and far less forgiving: complete, accurate, structured product data. These include schema markup, GS1 product identifiers, real-time pricing and availability feeds and machine-readable return policies, shipping windows, and loyalty terms.

If any of it is missing, stale, or inconsistent across channels, you don't rank lower, you might just disappear from consideration entirely, because the agent has no basis to evaluate you at all.

This also changes what "conversion" measures. Agents aren't persuaded by creative, they're persuaded by confidence in the underlying data: price, stock, delivery promise and returns. Brand storytelling still matters for sentiment and citation-worthiness (agents weigh third-party authority and reviews before they'll recommend you), but the transaction itself is won or lost on data hygiene. 

What this looks like when it works

There are already signs of what this looks like in practice. THG Commerce's fulfilment network, backed by a £270 million investment in AI-powered automation, exists to make delivery and availability promises that can be trusted in real time, not just marketed on a landing page. That distinction matters more in agentic commerce than it ever did in traditional e-commerce: an agent that gets burned by an inaccurate stock or delivery claim won't recommend that merchant again, and there's no marketing team to smooth it over.

On the agentic layer, THG Commerce's AI Shopping Assistant, built on Google Cloud's Gemini Enterprise Agent Platform, has already moved from concept to measurable commercial results. In early pilot data from Myprotein, it delivered an 8x higher conversion rate than the site average and a 20.8% uplift in average order value; returning customers using the assistant saw a further 4.61% lift in conversion and a 9.6% AOV increase. Those numbers exist because the assistant can reason over structured product, pricing, and availability data in real time, the exact discipline UCP and ACP now require of every merchant who wants to be legible to third-party agents, not just their own.

The lesson for other brands isn't "build a chatbot." It's that agentic commerce performance is downstream of data architecture. The assistant (or the external AI agent shopping on a customer's behalf) is only as good as the feed feeding it. 

What you should actually do in 2026

None of this means SEO suddenly stops mattering. People will continue to search, browse and compare products in the ways they always have. But discovery is becoming more fragmented. Alongside search engines and marketplaces sits a growing layer of AI systems making recommendations and, increasingly, purchases on behalf of customers. Brands should now optimise for both humans and machines, rather than treating them as competing channels.

Three priorities stand out:

• Treat product data like customer infrastructure. Product feeds, schema, identifiers and inventory accuracy should be governed with the same discipline as your checkout or payments platform
• See how AI already understands your brand. Ask ChatGPT, Gemini and Google's AI Mode about products in your category. The gaps between what you know about your products and what AI understands are often revealing.
• Prepare for protocol-based commerce. ACP and UCP are still in their early stages, but understanding how they'll fit into your commerce stack today is easier than retrofitting them later.

The click isn't disappearing any time soon, but it is becoming just one stage in a much broader discovery journey. More buying decisions are beginning in conversations between customers and AI systems, long before anyone lands on a product page.