Ecommerce's Structural Reset: What's Actually Happening

There's a word that gets used a lot in commerce commentary at moments of significant change: "transformation". It's a word I've come to distrust. It implies a transition from one stable state to another stable state, with the suggestion that once the transformation is complete, things will settle down.

Digital Commerce 360's framing of early 2026 as a "structural reckoning" is a better description of what's actually happening, because it doesn't promise a landing point. It describes forces reshaping the underlying structure of the industry, not just the surface. DC360's analysis centres on how AI-driven product discovery is shifting where shopping begins, how machine-readable data is becoming a revenue precondition, and how a divergence is opening up between different buyer journeys.

What makes this a structural reset rather than a collection of separate challenges is that those forces are happening simultaneously and they interact with each other.

Why the Simultaneous Part Matters

Take zero-click search and agentic commerce separately, and you have two interesting developments. Take them together, and you have a coherent picture of a world where the traditional funnel (search, site visit, browse, basket, checkout) is being replaced by something that looks quite different.

In the new model, discovery happens inside an AI interface. The AI answers the question or completes the discovery step. In an increasing number of cases, the transaction also happens inside the interface through agentic checkout. The merchant's website is either not visited at all, or is visited by a customer who has already largely decided what they want and is completing the purchase. The role of the site as a discovery and persuasion environment shrinks.

This is not fully realised yet. Agentic checkout is still early, zero-click commerce is still a small fraction of total transactions, and plenty of purchase journeys still follow the traditional path. But the trajectory is clear and the infrastructure investment is serious. The Universal Commerce Protocol, co-developed by Google and Shopify and endorsed by more than 20 retailers and platforms including Walmart, Target and Mastercard, makes the agent-completes-the-purchase pathway available to any retailer who implements it. The network effects of that, once a critical mass of retailers is live, are potentially significant.

Add tariff disruption to this picture and you get an additional layer of pressure. The sourcing and pricing stability that allowed for reasonably predictable demand and margin planning has been disrupted. For UK retailers, this compounds with a distinct layer of Brexit-related supply chain friction: a separate set of cross-border complications that predates the US tariff wave and does not resolve when Washington changes course. AI-powered supply chain tools help navigate this faster, but they help the retailers who have them more than the ones who don't, which widens an existing gap.

The Compounding of Advantages

Something that sits underneath the structural reset narrative and does not get discussed enough is the compounding of advantages. Retailers who had strong first-party data going into 2025 can build better AI personalisation. Better personalisation generates more engagement and more data. More data improves the AI. The flywheel runs.

Similarly, retailers who had good product data and machine-readable site architecture going into the AI traffic surge are capturing that 393% growth in AI-referred visits more effectively than the ones who didn't. The traffic comes to everyone proportionally, but converts better at the sites that were ready for it.

And retailers who had invested in supply chain AI before the tariff disruption had a response-time advantage when the disruptions arrived, which translated into better availability and margin preservation than competitors doing it manually.

None of this is new as a dynamic. Compound advantage is how market concentration happens in most industries. What's new is the speed. My reading is that the advantage gaps that previously took a decade to open up are opening in two or three years. Whether that's exactly right I can't say, but the directional case is hard to argue with.

The Questions That Don't Have Good Answers Yet

There are a few things I genuinely don't know how to model.

What happens to brand equity in a world where a significant fraction of purchases are made by AI agents optimising primarily for price and availability? The brand experience (the discovery, the browse, the considered purchase decision) is part of how brand relationships are built. If that experience is increasingly compressed or bypassed, does brand loyalty erode, or does it shift to a different kind of relationship?

What's the right level of investment in traditional ecommerce capability (site experience, SEO, CRO) versus agentic commerce infrastructure (structured data, machine readability, agent protocols), when the agentic share is still small but growing fast? The investment horizons are different and they compete for budget.

And what does this look like for smaller independent UK retailers who don't have the data, the engineering capability, or the budget to compete on the AI infrastructure dimensions? The access gap piece I wrote last year identified this as a serious concern. Nothing I've seen since has resolved it, and the structural forces that are concentrating advantage with larger, better-resourced retailers are if anything accelerating.

The Honest Position

The structural reset is real. The direction of travel is clear enough that planning on the basis of the old model — acquire via search, convert via site, retain via email — being durable is probably a mistake. But the new model isn't fully defined yet, the transition timeline is uncertain, and the degree of disruption will vary significantly by category and customer cohort.

Holding both those things simultaneously — the environment is genuinely changing in significant ways, and the pace and shape of change is uncertain enough that you shouldn't bet the whole operation on a specific scenario — is uncomfortable but accurate.

It's not a satisfying conclusion. But anyone selling a satisfying conclusion in the current environment is probably selling something.

---

Data sources: Digital Commerce 360, Retail Brew, Adobe AI Traffic Report, Google Universal Commerce Protocol.