What Is Agentic Commerce? A Plain-English Guide for Shoppers and Brands

Ask ten people what “agentic commerce” means and you'll get ten answers - which is a polite way of saying it's become a buzzword. So whenever someone uses it, our first question is simple: how do you actually define it?

It's best understood not as one thing but as a spectrum of “flavours” - a lens we first picked up from Kelly Goetsch and have built on here. Most explainers describe those flavours from the brand's side, all protocols and plumbing. We'd rather look from where you're actually standing: what does this stuff feel like when it's your money and your time on the line?

A quick word on that third flavour, because it really has two speeds. The first is reactive: you hand the agent a job and it does it - “book my usual airport parking for next week.” The second is proactive: the agent notices your flight and hotel are already booked, spots that you've not sorted parking, and books your preferred spot before the price climbs on short notice - then tells you what it did and why. Same capability, very different feeling. One waits to be asked; the other earns the right to act.

Now my honest, personal take: I'm sceptical that a little “buy” button works for big-ticket, high-consideration purchases. When you're spending serious money, you want to research, compare and feel sure. I don't think many people will let an assistant one-click a €2,000 decision any time soon. Where AI genuinely shines for those purchases isn't buying - it's helping you decide: asking the questions you didn't know to ask, narrowing the field, building your confidence. Assistance, not autopilot.

So why does any of this matter if you run a brand rather than shop at one? Because every effortless moment above rests on some distinctly un-effortless groundwork. For an agent to recommend your product, add it to a cart and check out - inside ChatGPT, Gemini or anywhere else - it first has to find you, read your catalogue and transact against your systems. That means clean, structured product data, an API for everything, and a storefront built to be machine-readable.

This is exactly where composable and headless architecture (the MACH approach) earns its keep, and where the emerging agent standards - ACP, UCP and MCP - come in. Put bluntly: if an agent can't understand your shop, you're invisible on the fastest-growing shelf in retail. Getting brands genuinely agent-ready - on Shopify Plus, commercetools or Medusa - is precisely the kind of unglamorous, high-impact engineering we rather enjoy.