Loading lesson…
Tell the AI agent 'write a plan' before doing anything. Then approve it.
Before an AI agent acts, ask it 'write your plan first.' Read the plan. Approve only if it looks safe and right.
Tell an AI: 'Make a plan to clean up my notes folder. Don't do anything yet, just plan.' Read it.
Here's why "Make AI Agents Write a Plan First" matters: AI agents don't just answer questions — they can do things, like looking things up, writing files, or talking to apps. Tell the AI agent 'write a plan' before doing anything. Then approve it — and knowing how to apply this gives you a concrete advantage.
15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-explorers-agentic-AI-and-the-write-a-plan-r10a5
An AI agent has been asked to help organize your school notes. What is the first thing you should ask it to do?
You asked an AI to make a plan for cleaning up your files. The plan says 'Step 1: Do stuff. Step 2: Fix things. Step 3: Make it nice.' What is wrong with this plan?
Which of these is a reason why you should read an AI's plan before it acts?
What should a good plan always include about the tools the AI will use?
You receive a plan from an AI that includes a step where it will delete old files. What should the plan clearly say about this step?
An AI wrote a plan but it's very vague — you can't really understand what it will do. What should you do?
What does it mean to 'approve' an AI's plan?
Why is it important for a plan to have numbered steps like 'Step 1, Step 2, Step 3'?
You asked an AI to make a plan to help with your homework. The plan says it will use a calculator app and a spell-checker. What should you think about these tools?
What could happen if you let an AI agent act without first reviewing its plan?
A friend says they just tell AI agents to do tasks without checking any plans first. Based on what you learned, why is this risky?
You asked an AI to plan a birthday party for you. The plan includes 'Step 1: Order food from a restaurant.' Why is this a good plan?
What should you do if you read a plan and see a step that seems dangerous or wrong?
An AI agent wants to edit your writing. What is one thing you should look for in its plan?
Why is it better to catch a problem in a plan rather than after the AI has already acted?