People used to find a business one way. They searched, and they clicked a link. AI changed that. Now an assistant might read your website, pull from a knowledge file, ask a live connection a question, or look you up through a standard that didn't exist six months ago. The number of ways AI can find and use a business keeps growing, and the big platforms keep adding more.
Keeping track of all that isn't realistic when you run a clinic or a studio or a restaurant. You have a business to run. You're not going to follow AI standards for a living.
So we handle it. Every Courtyard customer gets a full AI-discoverability surface, built automatically from the knowledge base you already keep with us. It covers the ways an AI might find, read, or use your business today. And when a new standard shows up and starts getting real adoption, we add it for you. Here's what's in that surface right now.
The map and the full text
The starting point is simple. Make everything about your business easy for any AI to find and read. Courtyard publishes an llms.txt, which is basically a map that tells AI where your information lives. It also publishes an llms-full.txt, the complete text of your business in one clean file, plus a normal sitemap and feed. Any AI can read the whole picture today: your services, hours, pricing, and policies, all accurate and current.
A live connection
A file is a snapshot. Sometimes an assistant needs to actually ask a question, like whether there's a 6pm class on Thursday. So Courtyard also runs a live connection to your knowledge base. It's built on the Model Context Protocol, or MCP, the emerging standard for letting AI query a source directly. We pair it with a set of agent skills that tell an assistant how to use it. Now an AI can get a current answer about your business on demand, instead of going off whatever it cached last.
Structured knowledge: the Open Knowledge Format
In June, Google Cloud published the Open Knowledge Format, or OKF. It's an open standard for writing a business's knowledge as clean, labeled files that any AI can read and carry with it. Courtyard now builds an OKF bundle for you automatically. Every topic gets its own file, so pricing, hours, and each service are separate, and each one is tagged so an assistant knows what it's looking at. It's the same knowledge you already have, in the format the industry is settling on.
A discovery catalog: Agentic Resource Discovery
Also in June, Google and Microsoft, along with companies like GitHub, NVIDIA, Hugging Face, and GoDaddy, published Agentic Resource Discovery, or ARD. It gives an AI agent a standard way to find out what a business offers. In practice it's a short, machine-readable menu at a fixed address, and it ties the rest of the surface together. It points an agent to your live connection, your agent skills, and your full text, and it spells out the real things someone can ask about you. Courtyard publishes this catalog for every customer and keeps it in step with everything else.
Between them, these pieces cover the questions an AI tends to have. Where does the information live, what does the business actually know, what can I ask it, and how do I use it. They answer in whatever format the assistant on the other end understands.
These standards are early, and that's the point
Some of this is brand new. ARD and OKF are only a few weeks old. But they come from the companies building the AI everyone already uses, and adoption is picking up fast. Getting on them early matters. As this becomes normal, the businesses that were already set up are the ones AI finds and trusts first.
The hard part for a small business has always been keeping up. A new standard lands every few months, each with its own file and format and address. Our whole point is that you don't have to watch for any of it. We build the surface, refresh it whenever your information changes, and add each new standard as it arrives. It runs in the background, and it doesn't cost extra.
For years, getting found meant being easy for Google to read. That still counts, but it's no longer the whole story. More often now, someone asks an AI about you before they ever land on your site, and the answer they get comes down to what AI can find. We make sure it finds you, whichever way it looks. That's what being the business AI understands best actually takes.
Not using Courtyard yet? Take a look at getcourtyard.ai.
