Happy Webs

Agent readiness · Safe actions

Make your website usable by AI agents.

Discoverable services, documented tools and human-approved actions built around the smallest useful interface—not another vague AI badge.

A UK business team mapping a safe website action for an AI agent
01Discover02Describe03Approve04Verify
100 / 100Kingsland Fabrications live scan
Human approvalkept around sensitive actions
Verifieddiscovery, contracts and failures tested

Kingsland Fabrications, 100 / 100.

This is a real UK fabricator we work with. Cloudflare's independent isitagentready.com scanner rates their site 100 across every category, Discoverability, Content, Bot Access, and API/MCP/Skill Discovery. Level 5 · Agent-Native. Same playbook, applied to your site.

Cloudflare isitagentready.com scan result showing Kingsland Fabrications scoring 100/100, Level 5 Agent-Native, with perfect scores across Discoverability (3/3), Content (1/1), Bot Access Control (2/2), and API, Auth, MCP & Skill Discovery (6/6).
Live scan result · re-run the scan yourself →

Start with the real blocker.

Four signs the current setup is costing time, enquiries or confidence:

01

Agents can read a page but cannot discover the action

A contact form makes sense to a person, but an agent needs a clear description of what the action does, which fields are required and what happens next.

02

Your APIs exist without a usable contract

Internal endpoints, booking tools and quote workflows may already exist, but inconsistent schemas and missing documentation make them unreliable or unsafe for agent tool use.

03

Sensitive actions have no approval boundary

Submitting personal data, sending a message or changing a record should not happen invisibly. The workflow needs consent, validation, authentication and a human confirmation point appropriate to the risk.

04

Nobody can see when an agent call fails

Without logs, rate limits, useful errors and a fallback route, a failed tool call becomes a lost enquiry or a silent operational problem.

A clear scope, written down.

The exact mix depends on the problem, but these are the practical parts we can own:

01

Agent-readiness baseline audit

Review discoverability, machine-readable content, action endpoints, MCP or OpenAPI coverage, permissions, approvals and error handling

02

Machine-readable service summaries

Publish concise service, pricing-band and process information at stable URLs that software clients can retrieve and interpret

03

Discovery and content signals

Configure appropriate robots directives, content signals, Link headers, llms.txt files and well-known discovery routes where the platform supports them

04

Skills discovery index

Describe useful actions such as request a quote, prepare an enquiry or check availability through a consistent agent-skills index

05

MCP or tool interface

Expose the agreed read or action capabilities through a focused Model Context Protocol server when MCP is the right fit

06

OpenAPI + API catalog

Document HTTP actions with an OpenAPI contract and API catalogue so inputs, outputs and failure states are explicit

Also included
  • Consent and confirmation flow, Keep the user in control before personal data is submitted, messages are sent or consequential actions are performed
  • Authentication and safety controls, Add scoped credentials, validation, rate limits and allow-listed actions appropriate to the data and operational risk
  • Logs, errors and fallback routes, Make calls traceable, return useful errors and give the user a normal web route when an agent action cannot complete
  • Verification and handover, Test discovery and actions against the agreed contract, document the result and show your team how to monitor and maintain it

Who works on it

The Agent Readiness bench for this job.

You still deal directly with Chris and Kay. These named AI specialists support the research, build and checking behind the work, with human approval around client, commercial and release decisions.

Jon

Integrations & APIs

Designs safe API, MCP and action contracts for machine use.

Will

Frontend & accessibility

Makes important website information structured and reliably discoverable.

Ruth

Risk & compliance review

Challenges permissions, sensitive actions and the human-approval boundary.

Ethan

QA & release confidence

Tests discovery, authentication, failures and real agent journeys.

One WhatsApp number. No bot maze.

Tell us what needs sorting. The studio routes the work and you always know who is responsible.

A shape you can plan around.

Step 1

Scan.

We check discoverability, machine-readable content, available actions, permissions, interfaces and failure handling against the live site.

Step 2

Foundations.

We publish the agreed discovery files, content routes, service summaries and action descriptions, then verify that clients can retrieve them.

Step 3

Action build.

Where useful, we build the MCP or OpenAPI interface with validation, approval, authentication, logging and a safe web fallback.

Step 4

Verify.

We run discovery and action tests, document the contract and show the team how to monitor errors, permissions and future changes.

Know the cost and first priorities.

Start with a free technical scan or a £1,495 Agent Readiness Audit and 90-Day Roadmap. Foundations work starts from £1,500. A focused MCP, OpenAPI or agent-action build starts from £4,500 once the action, data and approval boundary are clear. Ongoing monitoring starts from £750 per month where the interface is operationally important.

£1,495

Agent Readiness Audit + Roadmap

  • Technical, content, action and permission review
  • One workflow mapping session around the priority action
  • Three opportunities ranked by usefulness, risk and effort
  • Written 90-day plan and implementation estimate

Free

Agent-ready technical scan

  • 2-page PDF report on your current state
  • Automated scan results and short walkthrough
  • Five quick wins you can ship in a week
  • No commitment, no email gate

From £1,500

Foundations (2 weeks)

  • Discovery files, signals and Link headers in place
  • Skills index and machine-readable service summaries
  • Structured data and stable content routes
  • Re-scan and written handover

From £4,500

Agent action build

  • One focused MCP or OpenAPI action surface
  • Validated inputs, outputs and useful error states
  • Consent, confirmation and authentication controls
  • Logging, test calls and fallback web route

From £750/month

Monitoring and improvement

  • Readiness re-scan and contract checks
  • Action logs and failure patterns reviewed
  • Security, permission and schema drift fixes
  • A written monthly improvement scope

Not every website needs MCP or callable actions. If stable pages and a normal form already serve the user well, we will say so and stop at the useful foundations.

A faster agent readiness enquiry

Let your AI prepare the useful first draft.

It can read this service page, research your business and ask only what it still needs to scope agent readiness. You approve the enquiry before anything is sent.

Get a fixed quote
  • ChatGPT
  • Claude
  • Gemini
  • Perplexity
  • Grok

Copy it, open your AI, and paste it in.

Things people ask before starting.

01What is an agent-ready website?

An agent-ready website publishes information and actions in predictable, documented formats that software agents can discover and use. Depending on the need, that can include stable machine-readable pages, a skills index, an OpenAPI contract or an MCP server, with consent and safety controls around actions.

02How is Agent Readiness different from GEO?

Generative Engine Optimisation helps a business become a useful source inside AI-generated answers. Agent Readiness focuses on discovery and safe action: can an agent understand the available capability, call the right interface and complete a task with the user in control?

03How is this different from an AI agent?

AI agents are software workers we build to carry out a business workflow such as triaging enquiries or routing documents. Agent Readiness prepares a website or service so an external or internal agent can discover and use its approved capabilities. One is the worker; the other is the interface and safety layer.

04How is this different from SEO?

SEO helps pages appear in traditional search results and earn visits. Agent Readiness does not target rankings. It makes selected information and actions accessible through machine-readable routes and documented tools.

05Do you have proof this works?

Kingsland Fabrications scores 100 out of 100, Level 5 “Agent-Native”, on the agent-readiness scan used for this service. The technical score proves the discovery layer is present; any operational action is still tested separately against its own inputs, permissions and outcomes.

06What is an MCP server and why do I need one?

Model Context Protocol is a way to expose data and tools to compatible AI clients. You may need an MCP server when an approved agent must retrieve live information or perform a defined action. You do not need one merely because the website has content; a normal page or documented HTTP API can be the simpler answer.

07Can an agent submit an enquiry on a customer's behalf?

Yes, if the workflow is deliberately built for that action. The tool should show the user the completed enquiry, require confirmation before sending, validate the data and return a clear receipt. We do not treat silent form submission as a good agent experience.

08Can you do this if my site isn't on Cloudflare?

Yes. Well-known discovery paths, API contracts, MCP servers and machine-readable content can be implemented on many platforms. Some conveniences are hosting-specific, so the audit separates portable standards from optional platform features.

09How do you verify Agent Readiness?

We verify that the discovery route resolves, the schemas validate and a compatible client can understand the declared capability. For actions, we test valid requests, invalid requests, approval points, authentication, logs, errors and the fallback experience. A high site scan alone does not prove an action is safe.

10What do you need from us?

For the audit, we need the live website and a short call about the action you want an agent to support. Implementation normally needs website or repository access, API documentation, a technical owner for connected systems and a decision on consent, permissions and who approves sensitive actions.

11Why is the audit free?

The automated scan is quick and useful for spotting technical gaps, so we make that first check available without forcing a project. The paid audit adds workflow mapping, permission and risk decisions, interface review and a practical implementation plan.

12Is this related to schema and structured data?

Yes, but structured data is only one layer. An agent may also need a stable content route, a discovery file, a tool contract and a secure endpoint. We choose the smallest stack that supports the real action instead of adding every possible protocol.

Let’s get moving

Ready to
move faster?

Bring us the bottleneck, the ambition or the half-formed idea. We will find the fastest useful first move and turn it into something working.

Plan my first move Book a 15-minute callOr let your AI brief us Direct founders · clear scope · visible progress