Giving an AI access to your business platform is a big step, and we built the service assuming things will occasionally go wrong — so that when they do, you can roll back in seconds. Here is the protection model, plainly.
1. The dangerous actions are simply impossible
The AI cannot delete your site, change your password, create new admin users, or run code that endangers the platform. This doesn’t depend on the AI’s judgment — those actions are blocked in our code, on the platform side. Even if you explicitly ask for them, they won’t happen.
2. Before every change, the previous state is saved
Every time the AI edits a page, a snippet, a course structure or a form, the version that existed before the change is saved automatically — up to three versions back per item. If a result isn’t what you wanted, tell the AI “roll it back”. No tickets, no forms, seconds.
3. Big actions require your approval
Deleting content, emailing a large list, publishing live — the AI shows you exactly what it’s about to do and waits for your yes. No significant action happens behind your back.
4. Your data stays on your platform
Client details, orders, members, content — it all lives on your platform. We don’t copy your data to our servers and we don’t store it. The connection credential is stored encrypted, and every connection runs over HTTPS.
5. You control what’s allowed
You choose which capabilities are open to the AI: read-only answers, edit-but-never-delete, or full access. The permission settings are flexible and you can change them at any moment from your customer portal.
6. Instant disconnect
One click in the portal cuts the connection and invalidates the credential. Reconnecting later is just as simple. You can also revoke an individual team member’s connection without touching your own.
What the AI layer is not
- It doesn’t act on its own. Everything starts from a conversation you (or a team member) initiate.
- It isn’t Schoolyland watching your business. The AI conversation happens between you and your AI provider; our layer executes the platform actions you approve.
- It doesn’t replace your judgment. It’s a very fast pair of hands — review what it produces like you’d review a new assistant’s work, especially in the first weeks.
The fine print, if you want it
The AI layer has its own short legal documents: AI Layer Terms and AI Layer Privacy. You accepted them at checkout; they’re always available at those links. You can disable the AI layer entirely at any time — the rest of your platform is unaffected.
