The answers above cover the basics. Before a discovery call it helps to settle a few things inside your own team, because the answers shape what we build. Who owns the decision, who handles safeguarding and who looks after your systems? Having those names ready makes the conversation shorter. If you are comparing providers, ask each one who builds, who runs it afterwards and what happens when the agent is unsure.
Some questions are really about trust, such as what happens if an agent gets something wrong. The short answer is that your people approve sensitive actions, actions are logged and staff can step in at any point. Asking to see those controls before you commit is reasonable, and we are glad to show them.
A short list saves time and keeps expectations grounded. It does not need to be polished. A few lines from the person who feels the problem most are worth more than a long document, because they tell us where the work really sits. We use them to shape the discovery call and the blueprint that follows from it.
The questions here are the ones schools and edtech teams raise most often. Yours may be more specific, such as how a particular system connects or how a policy applies in your region. Ask it on the discovery call. We would rather say plainly that something needs checking than guess. You will also see the main points written into your agent blueprint, so you can read them again before anything goes live.
The tutor gives hints, not answers, and teachers can lock it to hints-only mode for homework and tests.