One of the easiest ways to get bad results from AI is to ask it to “make me a course.” That usually gives you generic content, weak sequencing and a lot of filler. A better workflow is: 1. Start with the outcome What should the learner be able to do by the end? 2. Define the starting point What do they already know? What are they likely to misunderstand? 3. Build the sequence Ask AI to break the journey into the smallest logical steps from current state to desired outcome. 4. Add practice For each step, ask for one realistic task, decision or example the learner has to complete. 5. Add feedback Ask AI to create the criteria for a good response and the most likely mistakes. 6. Edit like an expert Remove anything that sounds generic, repetitive or disconnected from the real context. COPY/PASTE PROMPT: “Help me design a short learning sequence for [AUDIENCE] who need to achieve [OUTCOME]. They currently know [STARTING POINT]. Create the smallest logical sequence of steps needed to reach the outcome. For each step include: - the key idea - one practical task - one likely misconception - one piece of feedback or guidance Do not pad the course with unnecessary content. Prioritise application over explanation.” The important bit is that AI is not replacing the learning design. It is helping you structure, test and improve it faster. If you try this, post what you’re building below and I’ll help you refine the sequence.