You can think of a workspace as an isolated environment for your flows and deployments. Those could be used e.g. to isolate dev vs. prod vs. staging environments. They could also be used to provide separation between different teams - say, your data scientists don't want to look at jobs loading data to a data warehouse and instead they want a custom workspace for their ML work.
But I agree, this should be somewhere in the docs 🙂
@Marvin open "Orion: add concept documentation explaining the purpose of workspaces in Cloud 2"