The AI set it up. We made it scalable.

How it came about.

The learning platform is based on Moodle. But nobody in-house knew how Moodle works in detail. So an AI did the initial setup.

The result worked – on a single server. The next step was clear: the platform was to become scalable, moving from a single server to a Kubernetes cluster.

For that, you need to know what has actually been set up. That's exactly the knowledge that was missing.

What we did.

Before the move came the question of what was actually supposed to move. Only then did we turn to the cluster.

1. Traced the setup: What had the AI set up? We first understood the configuration, database and file storage of the existing installation.
2. Built the target environment: Moodle now runs in a Kubernetes cluster instead of on a single server.
3. Migrated the data: Database, files and settings moved from the old installation to the new environment.
4. Switched over: The learning platform has been running on the cluster ever since.

Now the platform can grow.

What others can learn from it.

An AI can set up a platform. Someone still has to understand it – at the latest when it needs to grow, an update is due or something fails.

If you use AI for the initial setup, write down what was set up while it's still fresh. Otherwise every change starts with archaeology.