The way people find things online is shifting. Where a Google search once returned a list of links to click, an answer engine — ChatGPT, Perplexity, Claude, Gemini, Copilot — now returns a paragraph.
Sometimes it credits the source. Sometimes it doesn't.
That decision — who gets named when an AI summarises the web — is the most consequential shift in publishing since the invention of search. Sites that adapt their structure for that decision get cited. Sites that don't get paraphrased without attribution and lose the traffic either way.
We took the opportunity to build moonshotdigital.com as a working example of what an AI-readable site looks like in this context.
Here is a short tour of what that means in practice.
A plain-text twin for every page
Every page you can read here as a human has a clean, plain-text version waiting at a stable address — the same words, no menus, no scripts, no chrome.
When an AI agent fetches the site to ground an answer, it reads the twin instead of guessing what's body and what's frame. Less noise, better citations.
A clear, machine-readable policy
We publish a short policy explaining how AI systems may use this site — quote it, summarise it, train on it — and how we'd like to be credited when they do. Most sites leave that question to ambiguity. We don't.
A welcome mat for agents
Every page ships a small list of pointers in its response, telling automated visitors exactly where to find our policy, our index, and our content. An agent landing here cold can map the whole site in a single fetch.
Fresh work, broadcast immediately
When we publish or update something, we ping the engines that listen the moment it goes live. New content surfaces in answer engines in hours, not days.
About the playbook
The patterns above are part of a broader twelve-pattern playbook authored by our co-founder Simon Beauloye, drawn from his work building simonbeauloye.com — the reference implementation.
The playbook is open, freely usable under CC-BY-4.0, and intended as a working document for anyone building a content site that wants to be cited rather than paraphrased.
Cite as
Simon Beauloye, "The AI-readable web: a playbook for sites that get cited," https://simonbeauloye.com/