For years, websites have been built around one assumption: a human is on the other side of the screen.That assumption is starting to break.

WebMCP is an emerging web standard that lets websites expose structured tools that AI agents can use directly. Instead of an agent looking at a page, figuring out which button to click and filling out a form like a very fast human, the website can tell the agent exactly what actions are available. OpenAI is already supporting WebMCP through site tools in its desktop app, while the standard is being developed with Google and other industry players.

For a small business, the implications are surprisingly practical.

A webshop could expose tools for searching products, checking stock or adding something to a cart. A service business could expose booking or quote-request functions. A SaaS product could let an agent create a project, generate a report or update an account.

The important distinction is that this isn’t simply another way of scraping a website. The business decides which capabilities it makes available to agents.

And there is another piece falling into place on the automation side: n8n now has native MCP support that allows AI clients such as Claude, ChatGPT and coding agents to interact with n8n workflows. Its MCP server can even build and update workflows from natural-language instructions, while n8n’s MCP client connects agents to external services.

Put the two together and you get an interesting architecture:WebMCP exposes the business to the agent. n8n runs the automation behind it.An agent could receive a customer request, call a website tool, trigger an n8n workflow, check a CRM, create a quote, send an email and update the relevant records. We’re not quite at the point where every SMB needs to rebuild its website for agents.

WebMCP is still experimental, and adoption is only beginning. But the direction is clear. The next generation of websites won’t just need to be easy for humans to navigate. They’ll need to be easy for AI agents to use.

For SMBs and solo entrepreneurs, the move now is simple: identify one repetitive customer or internal workflow and ask what would happen if an AI agent could call it directly.

That’s where the next wave of automation starts.