Most AI lesson-planning advice treats planning like content generation. It is not. Planning is judgement: what did last week reveal, what does this class need next, where is the next zone of difficulty. AI is good at the parts that surround that judgement, not at replacing it. Here is the workflow, on the clock.
7:00 PM: review the week (4 minutes)
Open Mission Control's weekly summary. It surfaces three things you would otherwise dig for: which standards your class struggled on this week, which students need a check-in, and which assignments landed flat. Read it. Do not act on it yet.
7:04 PM: draft the arc, not the lessons (6 minutes)
Open a blank doc and write the five-day arc in your own words. What is the destination Friday. What is the on-ramp Monday. What is the hinge on Wednesday where you check for understanding before the back half of the week. This is the part AI cannot do, because it does not know your kids.
Keep this part analog
Type it, write it, voice-memo it. The point is that the arc is yours before the AI touches it.
7:10 PM: hand the arc to the planner (8 minutes)
Paste the arc into the Lesson Planner assistant inside the AI Assistants tool. Ask it to fill in: warm-ups, exit tickets, differentiation suggestions, a slide outline per day. Edit the output ruthlessly. You are looking for an 80 percent draft, not a final.
7:18 PM: post and close the laptop (2 minutes)
Push the lessons to your classes, print whatever needs printing, and stop. Monday morning is set.
What this workflow protects
- Your Sunday evening
- The teacherly part of planning, which lives in the arc, not the slides
- Students from generic AI-generated lessons that do not respond to last week
When to break the workflow
When you are starting a new unit, or coming back from a break, or teaching a topic for the first time, take longer. The 20-minute version is for steady-state weeks, not transitions.