While most brands were still debating whether AI would change ecommerce, it quietly started buying things. ChatGPT's instant checkout has been live since late 2025, serving hundreds of millions of weekly users, and in early 2026 Google switched on agentic checkout across Search and Gemini, with a "Buy for Me" button now live at selected US retailers and a protocol backed by Shopify, Walmart, Target and more than twenty other partners. Agents completing purchases on a customer's behalf is no longer a forecast. It is a feature shipping today.

That changes the question in front of you. It is no longer whether agentic commerce will matter, but whether your brand is present and chosen in the channels where agents now operate. And here is the part worth getting excited about rather than anxious: this is a land-grab moment, the kind that rewards the brands that move while everyone else is still forming a committee. The opportunity and the risk are the same thing, and both are live right now.

This stopped being theoretical

The signals are concrete enough to plan around. OpenAI built shopping research into ChatGPT to produce comparison guides and act across retail sites. Google's agentic checkout lets an AI complete a purchase directly on a merchant's site. Morgan Stanley projects that close to half of online shoppers will use AI shopping agents by 2030, accounting for around a quarter of their spending. You do not have to believe every number to see the direction, and the direction is one way.

For a leadership team, the useful posture is neither hype nor denial. It is to treat agentic commerce as a channel that has just opened, the way mobile or marketplaces once did, and to ask the early-mover's question: what do we do now, while being good at this is still a differentiator rather than table stakes? The brands that asked that question early about mobile spent the next decade ahead. The same window is open here.

Why this is an opportunity, not just a threat

The instinctive read is defensive: agents will commoditise us, strip our brand to a row in a comparison, and squeeze margin. That risk is real if you do nothing. But the same shift hands an advantage to the brands that prepare, because being the product an agent confidently selects is a winnable position, and most of your competitors are not yet trying to win it.

Early movers get two things. They get share, because when agents start routing real spend, the brands that are legible, trusted and easy for an agent to choose capture it while others are invisible. And they get learning, because the brands experimenting now will understand agentic behaviour long before the laggards start, and that understanding compounds. This is the rare moment where moving early is not reckless, it is the conservative choice, because standing still is now its own bet.

Being chosen by an agent is a different game

Winning a human and winning an agent are not the same contest, and that is the heart of what to prepare for. A human is moved by your photography, your story, your design, the feeling of the brand. An agent, at least today, is matching a customer's stated criteria against what it can read: price, availability, specifications, delivery, and the structured trust signals it can parse such as ratings and returns terms. The brand with the clearest, most complete, most machine-readable information has the edge, almost regardless of who has the prettiest page.

So the work shifts toward making your products and your brand legible to machines, the same discipline that wins visibility in AI answer engines. Structured product data, accurate and complete specifications, clean availability and pricing, and trust signals an agent can actually evaluate. This is where answer engine optimisation and agentic readiness converge: both are about being understood and chosen by software, not just by people. The brands investing in that legibility now are building the foundation the whole shift rests on.

The protocol decision is already on your desk

There is also a concrete, near-term call to make. The agentic checkout protocols now have serious backing, and platforms like Shopify are already part of that ecosystem, which means participating may be far closer to a configuration and readiness decision than a moonshot build. The question of whether to enable agentic purchase paths, and on what terms, is a live commercial decision, not a someday one.

That decision deserves a real answer rather than drift. It touches how you present price and inventory to agents, how the purchase is fulfilled and attributed, and what it means for the margin and the customer relationship on an agent-mediated sale. Getting your store and data genuinely ready to participate is increasingly an integration and infrastructure question, the kind of systems work that determines whether you can move when you choose to, or are stuck watching.

First-party data is your hedge and your edge

If agents increasingly sit between you and your customer, the direct relationships you own become more valuable, not less, because they are the one thing an agent cannot intermediate away. First-party data lets you understand demand, reach customers outside the agentic channel, and keep a relationship that does not depend on winning every machine comparison. In an agentic world it is both the hedge against disintermediation and the fuel for competing well within it.

This is why the brands that will thrive are not choosing between courting humans and being legible to machines. They are doing both, and using the direct relationship to stay in control of their own demand. It is the kind of foundational capability we have helped brands like Artist AI put at the centre of how they grow, and it pays whether the agentic shift runs fast or slow.

The brand still matters, just in a new place

It is tempting to conclude that if agents buy on specifications, brand becomes irrelevant. That is the wrong lesson, and acting on it would be a mistake. Brand still does the work it always did, it just does more of it before and around the agentic moment rather than at the point of selection. It is what makes a customer instruct their agent to prefer you, what earns the reviews and reputation the agent reads, and what brings people back to buy from you directly rather than through an intermediary that owns the relationship.

So the brands that win agentic commerce are not the ones that abandon brand for machine-readability. They are the ones that keep building a brand strong enough to be requested by name, while also being legible enough to be chosen on the merits. The two reinforce each other: reputation feeds the signals agents trust, and agentic visibility puts the brand in front of demand it would otherwise miss. Treating it as brand versus machine is the trap; the answer is brand expressed through machine-readable strength.

What to do now, while it is still early

Concretely: audit how legible your products and brand are to a machine, and fix the gaps in structured data, specifications, pricing and trust signals. Make a deliberate decision on the agentic checkout protocols rather than letting it happen to you. Invest in the first-party data that keeps you in control of the relationship. And start experimenting now, in a contained way, so your team builds real understanding of agentic behaviour before it becomes mainstream.

None of this requires betting the business on a single prediction. It requires recognising that the shift has already started shipping and choosing to be early rather than late, because early is where the share and the learning are. If you want help working out where your brand actually stands and what to do first, our ecommerce consultation is built to turn a fast-moving shift into a concrete, prioritised plan.