The original is one click away. Open original ↗
WebMCP explained: five ways to prepare your site for AI agents
Executive overview
AI agents are beginning to visit websites and complete tasks on behalf of users — booking demos, comparing prices, requesting quotes. Most sites aren't ready. WebMCP (Web Model Context Protocol) is a proposed browser standard, co-developed by Google Chrome and Microsoft Edge, that lets a webpage declare its available actions so agents can call them directly — without needing a custom API.
Agents have no tolerance for friction. If they can't complete a task cleanly, they move to a competitor.
The core insight: agent readiness is technical SEO with a transactional lens — clean forms, stable URLs, and clear labels are the foundation.
The two broken approaches agents use today
- UI automation — mimics human clicks, but breaks whenever CSS or layout changes
- APIs — structured and reliable, but most sites don't have public ones
- No standardized middle ground exists today — WebMCP closes that gap
How WebMCP works
- Two implementation approaches: declarative (add two HTML attributes to existing forms) and imperative (register tools dynamically via JavaScript for complex flows)
- Declarative: browser reads your form fields and translates them into callable tools for agents — no rebuild required
- Imperative: tools adapt to page state (e.g. checkout tool only appears when items are in the cart)
- Draft spec published February 2026; broader browser support expected later in 2026
- Chrome Canary has an early preview available now
Five things to do now
- Measure your AI visibility baseline — use SEMrush's AI Visibility Toolkit to check whether your brand appears in ChatGPT, Perplexity, and Google AI Mode; agent readiness is worthless if agents can't find you first
- Map your site's actions — list what users can actually do (book a demo, get a quote, compare plans); if you can't name five actions in 30 seconds, agents can't find them either
- Audit those actions — check for clear labels, predictable inputs, stable redirects, and clean HTML; vague labels or JavaScript-rendered forms will break the declarative approach
- Start the conversation internally — share Chrome's preview docs and the W3C community group with your dev team; frame it as upcoming, not urgent
- Test it yourself — enable the WebMCP flag in Chrome Canary; use the Model Context Tool Inspector extension to see what tools your pages expose to agents
More like this — when you're ready for early access.
Join the waitlist for a personal account and content recommendations based on what you're working on.
No spam. Unsubscribe at any time.
You're on the list. We'll be in touch before launch.