Library › Agents & Automation AI › College students
Brief an assistant to draft honest questions to professors
- For College students
- You practice Saying who it's for
- Time about 5 minutes
About this exercise
You want an AI assistant to draft emails to three professors about things you're stuck on. Write its brief: greet each as "Professor" plus their last name, ask about the concept rather than asking whether a graded answer is right, keep each email under 80 words, and never send. Output exactly 4 numbered plain-text lines: one draft per class in note order, then DONE.
Common mistakes when you ask AI for this
- Not showing the log format. Left alone, the AI writes a friendly report with headings instead of a checkable action log. Try: Show it literally: "Format every line like 1. ACTION: tool(args), then ASK_HUMAN: and DONE: lines. Plain text."
- Trusting the AI to count words. AI is bad at counting and runs long, especially when it explains each step. Try: Cap each line instead: "each line under 20 words, no explanations."
- No greeting rule. The AI writes "Hi Daniel" or "Dear Dr. Hannah". Try: Say: "Greet each as Professor plus last name."
- Asking for graded answers. The AI follows your notes and asks the professor to check graded work, which most won't do. Try: Add: "Ask about the concept. Don't ask if a graded answer is right."
- No length or send rule. The AI writes long, apologetic emails and may send them. Try: Add: "Each under 60 words. Draft only, never send."
What a strong prompt includes
- Sets the greeting
- Keeps it honest
- Sets a length
- Says draft, never send
- Shows the action-log format
Related exercises
- Answer a director's accrual question as a short Q&A (Finance AI · Accountants)
- Answer a patient's billing question in plain words (Writing AI · Nurses and healthcare workers)
- Answer your CMO's campaign question in one Slack message (Marketing AI · Marketing and content)
- Ask parents to turn on two-step sign-in for kids (Cybersecurity AI · Teachers)
- Brief a coding agent to fix a flaky test without overreaching (Agents & Automation AI · Software developers)