Over the past few months, I've been experimenting with a simple shift in how I work with AI coding tools like Cursor and Claude Code: I ask them to plan before they execute.
The difference is striking.
When the AI takes time to map out its approach—thinking through architecture, dependencies, and implementation steps—the code quality improves significantly. And it's not just about better code. This pattern forces me to think through the solution more carefully too.
You get two wins
Better planning and architecture on your end
Dramatically better execution from the AI
Tools like Cursor and Claude Code have even built planning modes directly into their interfaces now, making this workflow even smoother.
If you haven't tried this approach yet, I'd strongly encourage giving it a shot. Yes, you'll spend time iterating on the plan upfront, but it saves you from going back and forth with the AI multiple times trying to course-correct toward what you actually wanted.
What patterns have you discovered when working with AI coding tools?