The agent has to finish on the official site.
Being cited gets you named. Agent readiness is whether the assistant that named you can then complete the job — a quote, a booking, a ticket — on your site, without guessing at buttons. WebMCP is the proposed standard for that: the page declares its actions as tools.
We test the journeys with real agents, then build the tool layer where guesswork fails. Chrome documents an origin trial. ChatGPT documents Site tools. The spec can change. The audit still shows where ordinary conversion already stalls.
Found. Cited. Able to act.
Three jobs. They stack; they do not replace each other.
Found
Tabs, Google, comparison sites. SEO still matters. It is how most traffic still arrives.
Cited
ChatGPT, Claude, Grok, Perplexity, Google's AI Overviews. The work is making the official source the one the model quotes.
Able to act
The assistant that named you now tries to finish: a quote, a booking, a ticket. WebMCP is the proposed layer that lets it call named tools instead of guessing at buttons.
MCP is how we plug agents we run for you into the tools you already use. WebMCP is how your website exposes actions to someone else's assistant. Same letters. Different direction. Search operations is the measurement layer between cited and able to act: Search Console, Bing, citation checks, on a schedule.
The official source has to be callable.
A browser agent on an ordinary site reads the page and improvises. These are the journeys where that fails — and every stall is an enquiry that never shows in analytics.
| Journey | Where it stalls | What readiness changes |
|---|---|---|
Tickets and fixtures Governing bodies, unions, clubs | Date pickers, reseller listings, last season's PDF. The official source is not what the agent can complete. | Named tools: next fixtures, official ticket purchase, membership check. Confirmation stays with the person. |
Quote request Law firms, trades, agencies | Multi-step forms, vague labels, a submit button named something clever. | One named tool with described inputs. The agent knows exactly what to send. |
Booking Clinics, consultants, venues | Slot grids built in JavaScript that an agent cannot read reliably. | A check-availability tool returns real slots. A booking tool takes a chosen one. |
Membership and accounts Associations, portals, SaaS | Actions hidden behind menus and state the agent cannot see. | Tools scoped to login, with a confirmation step on anything sensitive. |
This page registers its own tools.
Not a mock-up. In a browser where WebMCP is enabled, the tools below are registered now. Two read-only lookups, one that tells the agent how to start an audit without sending it.
Checking whether WebMCP is enabled in this browser…
get_agent_readiness_pricing
JavaScriptReturns the three service tiers with prices in GBP, so an agent can answer cost questions from source instead of summarising the page.
{
"inputSchema": {
"type": "object",
"properties": {}
}
}get_agent_readiness_details
JavaScriptReturns what the service is, who provides it, how it differs from AI agents we run for you, and how to get in touch.
{
"inputSchema": {
"type": "object",
"properties": {}
}
}request_agent_readiness_audit
JavaScriptTells the agent how to start an audit. Filling the form is allowed. Sending still needs a person to press the button.
{
"inputSchema": {
"type": "object",
"properties": {
"name": {
"type": "string",
"description": "The person's name"
},
"email": {
"type": "string",
"description": "The person's email"
},
"message": {
"type": "string",
"description": "Optional context"
}
},
"required": [
"name",
"email"
]
}
}Nothing priced in the dark.
Audit first. Build second. Monitor after deploys. The same ladder as the rest of the studio.
Agent readiness audit
The existing consultancy day, pointed at the journeys that make you money: quote, booking, ticket, variant choice. You get where real agents stalled, what they misread, which actions have no usable path, and a prioritised list. Useful even if you never add a single tool.
WebMCP implementation
The tool layer, only where the audit proves it is worth it: which actions to expose, what each one is called, what it needs, and what comes back — wired to the logic you already have, with confirmation on anything sensitive. Tested against the Chrome and ChatGPT versions that exist today.
Monitoring
Deploys silently break tools. A care plan re-runs the journeys, checks every tool still registers, and sends a short note when something moved — the platforms or your site.
The honest limits.
Four things this service is not. The work is still worth doing.
This does not bring you traffic
WebMCP is a completion layer. An agent has to reach the official site before tools matter. That reach is citation work — answer-engine optimisation — and it pairs with this.
The standard can change
WebMCP is a proposal in a Chrome origin trial, not a finished standard. ChatGPT documents Site tools separately. That is why we sell agent readiness, with WebMCP as one ingredient. The audit stays useful whatever happens to the spec.
Demand is early
Agent visits are small today and growing. The honest case is being early, plus an audit that pays for itself in ordinary conversion findings — not a promise of agent-driven revenue this quarter.
Nobody controls the platforms
We control that your actions are exposed clearly, tested against the implementations that exist, and re-tested when either side changes. We do not control what an agent does next.
You, or the assistant that brought you.
The form is for people. The tools above are for agents. Either way, a person presses send.
Default interest is the consultancy day — that is the audit. Mention agent readiness in the message if you already know the journeys that matter.
WebMCP and agent readiness.
WebMCP is a proposed web standard that lets a website expose its actions as structured tools for AI agents. Instead of the agent guessing at buttons and forms, the site declares what can be done, what each action needs, and what comes back.
Chrome runs it as an origin trial, with a JavaScript API and a form-based API. ChatGPT's browser implements it as Site tools, JavaScript only. The spec lives in a W3C community group and can change.
Two directions. AI agents are software we run for you — outreach, inbox, follow-up, citation-watch — using MCP to plug into the tools you already use.
Agent readiness is the other way: a visitor's assistant arriving on your site to finish a job. WebMCP is the proposed standard for that. Same letters as MCP. Different direction.
We run AI agents against the journeys that matter — a quote, a booking, a ticket, a product variant — and report where they stalled, what they misread, and which fixes come first.
It is the existing consultancy day: half-day £275, full day £495. You leave knowing whether a tool layer is even the job.
No. WebMCP does not get you found or cited. Answer-engine optimisation is that work. What agent readiness changes is whether the visit ends in a completed enquiry, booking or order instead of an abandoned one.
Not necessarily. Agents can still drive an ordinary website by reading the page and clicking, and the audit tests exactly that. Tools make the difference on the journeys where guesswork fails: ambiguous labels, multi-step forms, ticket flows, product variants, anything where a wrong click costs money.
This page registers its own WebMCP tools in JavaScript. In a browser where WebMCP is enabled, an agent can read pricing and service details from source instead of summarising the prose. Sending an enquiry still needs a person to press the button, on purpose.
Cited is the start. Finishing is the job.
The first conversation is unpaid. You leave knowing whether a tool layer is even the work, or whether citation and search operations come first.