"WebMCP is the real answer for autonomous agents." Maybe eventually. Worth asking why it's needed at all.
WebMCP is genuinely early, worth flagging as unsettled rather than treated as decided either way. Even taken on its own terms, it's answering a narrower question than it sounds: how an AI agent copes with a web built for people, not what agentic commerce looks like if nobody had to cope with that at all.
6 min read
What WebMCP is trying to solve
A developer building an autonomous shopping agent hears that WebMCP will let it call actions directly on any website it visits, no pre-configured server required, and treats that as the eventual full answer to how agents will buy things. It's a reasonable thing to be excited about. It's worth being honest that it's still early, and worth asking a more basic question before betting everything on it: why does an agent need a way to cope with a website at all.
Unlike MCP, which needs a business or developer to set up a server in advance, WebMCP's aim is letting an autonomous agent discover and call a page's actions on its own, without anyone having configured anything ahead of time. That's a genuinely different job, connection for an agent that finds a business on its own rather than one a human pointed it at. How mature the actual specification is, who's building it, how widely it'll be adopted, are all genuinely unclear right now. It's early enough that treating it as settled would be dishonest.
The more basic question underneath it
Ask why an agent needs a way to cope with a website at all, and the honest answer is: because no two websites are built the same way, and almost none of them were built with an agent in mind. WebMCP exists to translate that mess into something an agent can act on. That's a genuinely useful thing to build. By its own premise, it's a patch for a web that isn't agent-native. Agent-native commerce, built from a blank page, wouldn't need the patch at all.
What a blank-page version skips entirely
An agent that didn't have to cope with a page-shaped web wouldn't need to discover actions on a page, it would just get the data directly, from a source built to be queried that way from the start. No page to visit, no actions to discover on it, no translation step at all. WebMCP is a genuinely clever answer to "how does an agent deal with today's web." It isn't the answer to "what would commerce look like if agents talking to data, not pages, was the starting point."
Where this shows up for a business
A retailer's facts reach an agent as structured, queryable data through Selfe, skipping the page-translation problem WebMCP exists to solve, rather than waiting for an emerging standard to finish maturing.
Should businesses build for WebMCP now?
Not with any confidence yet, it's too early to call. Worth watching, not worth treating as settled.
If it's a workaround, is it still worth having?
Yes, for the world as it currently is, most of the web isn't agent-native yet. It's a genuinely useful patch. It just isn't the destination.
A lot of what you've heard about agentic commerce isn't quite right.
Every myth in this hub comes from the same mistake: judging the future by how well it fits today's tools, a page, a widget, a chat window, a markup file, instead of asking what would make sense if you were starting clean. The actual destination underneath all of it is simpler than any of the myths: AI agents talking directly to data, not reading pages built for people.
"We've set up MCP, so we're agent-ready." That covers your own team's tools. It's a different job from the one a stranger's agent needs done.
MCP is genuinely useful, just not for the case most businesses think it covers. It's strong for an operator's own tools, a chat assistant working against your booking system, less so for a stranger's AI agent trying to book or buy from you. Both are real, but only one of them is what "agent-ready" usually means.
"Add an llms.txt file and you're SEO-optimised for AI." There's no ranking signal involved. What it's for is narrower and more specific.
An llms.txt file isn't a ranking mechanism, and treating it like a new kind of SEO misreads what it's for. What it can genuinely help with is narrower and more useful: telling an AI agent what a business can do, not making its content rank better anywhere.