I call it the AIADD Method (Agile-AI Driven Development). Its all laid out here in the follow up to the rules generating rules videos: https://www.youtube.com/watch?v=JbhiLUY_V2U
I’ve been experimenting heavily with Cursor AI for building more complex, multi-layered applications and kept hitting the same issues:
• Agents not understanding long-term intent
• Loss of context across multiple sessions
• Burned credits and unnecessary rewrites
• Poor structure and architectural decisions
I realized the root issue wasn’t the tools — it was the lack of structure. So I built a workflow that merges Agile development roles with AI prompt scaffolding.
In Part 1 of a new series, I clearly explain how Agile and AI come together and work with tools like Gemini or Open AI to use deep thought and deep research FOR FREE or very cheap to produce the artifacts that will then be used within cursor to drive the AI to complete success granularly building your app, with minimal need for rules to keep it on the rails, because these artifacts will do the heavy lifting.
Would love your feedback, especially from folks building multi-agent or multi-module systems.
These prompts and the a deeper dive into the workflow will follow up very soon - and check the Auto Rules Generator repo also - I will share the prompts that I use for deep thinking and deep research to produce the docs that drive cursor to build complex apps consistently!