Think of this as a small company of AI helpers. Each helper (the animals) has its own room and one kind of job. You only ever talk to one of them: the CEO.
- Tell the CEO what you need in the box at the bottom, in your own words. For example: “Write a blog post about phishing” or “Check if our labs work”.
- The CEO picks the right helper and hands over the job. You will see that helper wake up and start working.
- Watch it work. The “Live” panel shows who is busy and what they are doing right now.
- Read the result. When a helper finishes, a card appears under “Needs you”. Press Full result to read it, then Approve when you are happy.
- Come back later. The History tab keeps every question and every result.
- Big jobs. Ask the CEO for something large, like a whole course. It splits the job into steps in the Plan tab. You look at the steps, press Run all, and each helper hands its result to the next one.
- Repeat or queue work. The Plan tab lets you schedule a helper (for example every Monday morning) and keep a list of jobs waiting for their turn. Nothing starts unless you scheduled it or pressed Run.
Good to know
- Nothing is published, sent or paid for by itself. Helpers only write drafts. Sending, posting and paying stay with you.
- Nothing runs while you are away. No work, no cost.
- Every result is checked by a second helper before you rely on it: code by the Engineer, blog posts by the Researcher, social posts by the Writer, visuals by the Marketer, lessons by QA. Their verdict (APPROVE, CHANGES NEEDED, BLOCKED) shows up as its own card.
- “Check sources” in a result looks up the CVE numbers and links the helper mentioned, so you can see if they exist.
- A daily limit stops the office from using too much AI time. You can see it on the Cost tab.
- Something looks wrong? Open Settings → Health. It tells you in plain words what is missing.