Get Recommended, Cited, and Booked by AI
Search is splitting in two. Half your future customers will still type into Google. The other half will ask ChatGPT, Claude, Perplexity, or an AI agent working on their behalf. We make your website readable, recommendable, and bookable by AI — so when the question comes up, you are the answer. Same work we did for Kingsland Fabrications, who now scores a perfect 100 / Level 5 on Cloudflare's agent-readiness check.
The promise
Proof · live result
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.
Where it hurts
You don't need convincing. You need fixing.
Four signs you're ready for this conversation:
Your site is invisible to AI
ChatGPT, Claude, and Perplexity are recommending your competitors because their sites are readable and yours is a maze of JavaScript. Every month you wait, more buyers use AI instead of Google — and your name never comes up.
Agents can read your pages but can't do anything
An AI assistant helping a buyer find a fabricator can parse your services page. But it can't submit an enquiry, request a quote, or book a call on their behalf. That friction costs you the lead.
You have no idea if you're being cited
Unlike Google, AI search engines don't give you a rank report. You don't know whether ChatGPT mentions you when someone asks, or whether Perplexity pulls your content. Right now you're guessing.
Your website is not ready for 2027
In two years, most B2B buying research will start with an AI assistant. Sites built the old way — JS-heavy, no structured data, no agent endpoints — will be skipped entirely. Retrofitting later will cost more than building it right now.
What you get
Not just a deliverable — a working system.
Agent-readiness baseline audit
Full Cloudflare-style scan: Content Signals, Link headers, Markdown for Agents, MCP discovery, skills index, OpenAPI, API catalog. 20+ checks.
Robots.txt + Content Signals
Declare AI training, search, and grounding preferences properly — let the right bots in, keep the scrapers out.
llms.txt + llms-full.txt
Citation-optimised brand summary and full knowledge base formatted for AI grounding. Fed directly into ChatGPT, Claude, and Perplexity.
Skills discovery index
Publish your key actions (contact, quote, review, book) as agent-discoverable skills at /.well-known/agent-skills/
MCP server
A live Model Context Protocol server exposing your lead forms, services, pricing bands, and content — so Claude, ChatGPT, and SDK agents can submit leads on behalf of users.
OpenAPI + API catalog
RFC 9727 API catalog + OpenAPI 3.1 spec so any HTTP-aware agent can discover and call your endpoints.
- Markdown content feed — Every blog post and service page available as clean markdown at predictable URLs — AI engines prefer this to HTML scraping.
- Generative Engine Optimisation (GEO) — Content rewrites and schema work targeted at citation in ChatGPT, Claude, Perplexity, and Google AI Overviews. Not the same as Google SEO — different game, different tactics.
- GA4 + Search Console review — We take access to your analytics and Search Console, identify what AI referral traffic looks like for your industry, and track citation lift over time.
- Quarterly citation tracking — Test how often your brand appears in ChatGPT, Claude, and Perplexity for target queries. Report, tune, re-test.
Process
A shape you can plan around.
Scan.
Free Agent-Ready Audit against your live site — 20+ checks covering discoverability, content, bot access, MCP, and skill discovery. You get a 2-page PDF and a Loom walk-through.
Foundations.
We implement Content Signals, Link headers, llms.txt, skills index, schema fixes, and Markdown for Agents. 2 weeks, fixed fee, measurable score lift.
Agent-Native Build.
We build a live MCP server wired to your lead forms, publish an OpenAPI spec + API catalog, and rewrite key content for GEO citation. 4 weeks.
Track.
Monthly citation tests across ChatGPT, Claude, Perplexity, and Google AI Overviews. We report what's moving and what needs sharpening.
Pricing
Fixed fee, phased, or retained.
Fixed-fee packages: **Agent-Ready Audit** is free (2-page PDF + isitagentready.com scan). **Foundations** £1,500–£2,500 (robots.txt, llms.txt, Link headers, skills index, schema — 2 weeks). **Agent-Native Build** £4,500–£8,000 (Foundations + live MCP server + OpenAPI + markdown feed + GEO content work — 4 weeks). **Retainer** £750–£1,500/month (quarterly re-scan, citation tracking, content refresh).
Questions
Things people ask us.
01 What's the difference between SEO and AI-Native Readiness?
SEO is about ranking on Google's 10 blue links. AI-Native Readiness is about being the answer when someone asks ChatGPT, Claude, or Perplexity. Different inputs, different outputs, different tactics. You probably need both — but they're genuinely separate disciplines.
02 Do you have proof this works?
Yes. Kingsland Fabrications — a real UK fabricator we work with — scores a perfect 100 out of 100 on Cloudflare's agent-readiness scanner (Level 5, "Agent-Native"). Happy Webs' own site scores the same. We apply the exact same playbook to clients.
03 What is an MCP server and why do I need one?
Model Context Protocol is the standard Claude, ChatGPT, and most AI agents use to discover and call tools. If your site has an MCP server, an agent helping a user buy from you can call tools like "request_quote" or "submit_enquiry" directly — no copy-paste, no form filling, no friction. If you don't have one, they go elsewhere.
04 Isn't this just vanity — do agents actually send leads?
Right now it's early. Honest answer: the volume isn't huge yet, but the direction of travel is clear. Gartner expects 25%+ of enterprise search queries to shift to AI assistants by 2027. Getting in early means you're indexed, cited, and trusted before the flood. Retrofitting later will cost multiples of doing it now.
05 Can you do this if my site isn't on Cloudflare?
Yes. Most of the work is in your content, robots.txt, /.well-known/ paths, and HTTP response headers — those work on any hosting. Cloudflare-specific features (Markdown for Agents) require Cloudflare, but we can replicate them with a small worker or middleware on other platforms.
06 How do you measure results?
Three ways: (1) isitagentready.com score climbs from wherever you start to 90–100 / Level 5. (2) GA4 shows referral traffic from chat.openai.com, perplexity.ai, claude.ai. (3) We test brand citation monthly across target queries — "best fabrication shop near Manchester", "UK AI consultancy for SMBs" — and track whether you appear in ChatGPT and Perplexity answers.
07 What do you need from us?
Read-only access to Google Analytics 4 and Search Console (we provide a template grant request). FTP or Git access to your website. A 30-minute kickoff call to agree target queries and priority services. That's it.
08 Why is the audit free?
Because it takes us 20 minutes and you get something genuinely useful — a score, a gap list, and specific fixes. Most businesses end up fixing the easy wins themselves. The ones that want it done properly hire us for Foundations or the full Build. We'd rather earn trust with the free audit than sell something nobody needs.
09 How does this help if most of my customers find me via Google?
It doesn't replace Google — it adds a new channel. A year ago 100% of search was Google-shaped. Now a growing share is conversational. Within 24 months that share will be material. Investing £1,500 now for Foundations is cheap insurance; waiting until competitors have lapped you costs more.
10 Is this related to schema and structured data?
Yes — schema.org JSON-LD is part of it, but only part. AI engines read schema, but they also read llms.txt, HTTP Link headers, /.well-known/ discovery paths, and (crucially) whether you expose callable tools via MCP. Schema alone gets you partway. Full agent-readiness needs the whole stack.
Keep reading
Related services.
Let's see if we can help.
A 30-minute chat with Chris & Kay. No slides. No pitch deck. We tell you honestly whether we're the right fit — and if not, who might be.