Rob Richardson

Here are 5 mental models I teach:

1. Reverse Prompting
Stop guessing what the perfect prompt should be. Give AI the outcome and context, and have it design the reusable prompt for you.
2. Prompt Optimization
Already have a prompt? Ask AI to identify ambiguity, add constraints, clarify assumptions, and improve the instructions.
3. Chain of Verification
Don’t just say, “Make sure this is accurate.” Require claims to be verified against specific sources and cited.
4. Adversarial Prompting
Ask AI to attack your design. Find the holes. Challenge the assumptions. Surface risks you didn’t consider.
5. Multi-Perspective Simulation
Have AI examine the problem from competing perspectives before synthesizing the strongest coherent answer.