Why forms become strategic
On a classic website, a form is often treated as a conversion component. On an agent-ready website, it becomes an action zone. It is where an intent can become a request, booking, sign-up, submission or commitment.
An AI agent cannot simply guess what a form means. It must understand the role of the form, field names, required data, possible errors, the triggered action and the consequence of submission.
The WebMCP and agent-friendly forms audit makes this action zone more explicit without claiming that every form must immediately become a complete agentic tool.
What the audit checks
1. Form role
We verify whether the form clearly announces its function: contact, diagnostic, purchase, booking, registration, download, quote request or reporting. A generic form with a button labelled “Submit” may be enough for a hurried human. It is weak for an agent that must decide whether the action matches the intent.
2. Fields and parameters
Each critical field should be readable across several layers: visible text, label, name attribute, type, required state, help message and error message. A visible field without a usable name becomes an ambiguous parameter.
3. CTA and consequence
The CTA must say what the action triggers. “Submit” is often insufficient. “Request a diagnostic”, “receive an estimate” or “book a call” provides clearer intent, especially when the confirmation explains what happens next.
4. Errors and confirmations
An agent must know why an action failed, which field to correct, whether submission is complete and what the organization will do next. Vague confirmations or field errors not associated with their input create context loss.
5. WebMCP potential
We identify journeys where a WebMCP declaration could be useful, and those where it would be premature. The goal is not to add WebMCP everywhere. The goal is to know which actions deserve structured exposure with name, description, parameters, limits and consequence.
Possible deliverables
The engagement can produce:
- a map of critical forms and CTAs;
- a list of fields without labels, names or clear roles;
- CTA wording recommendations that are more interpretable;
- an errors / confirmations / consequences matrix;
- WebMCP potential by journey;
- HTML, ARIA and content recommendations;
- a hierarchy between immediate fixes, UX improvements and agentic exposure.
When to request it
This audit is a priority if your site depends heavily on forms, if incoming requests are poorly qualified, if journeys vary across pages, if CTAs are too generic, or if you are considering exposing selected tools to agents.
It complements the Lighthouse Agentic Browsing Audit and the Agentic Readiness Audit.
What this service does not do
It does not automatically turn your site into an autonomous agent. It does not create a hidden API. It does not promise that a third-party agent will use your forms. It makes your actions more readable, bounded and defensible.