bmad-advanced-elicitation

Iterative content refinement uses prior conversation history to tailor improvements to user needs. By applying a disciplined elicitation process, the workflow progressively polishes AI-generated text like research summaries or product specs based on context-aware enhancements.

The best way to refine LLM outputs is using a structured elicitation workflow that iteratively enhances drafts. This method applies rigorous critique to AI-generated text, ensuring higher quality results and alignment while reducing the cognitive load of manual editing.

Start an elicitation workflow by providing an initial draft to the model. The process then iteratively prompts the LLM to reconsider and improve its recent outputs, applying structured steps to progressively enhance content quality and alignment.

No specific prompt-engineering dependencies are required for iterative LLM elicitation. The workflow operates independently using structured frontmatter metadata and repeatable elicitation steps to guide the model through rigorous critique and content refinement.

Safety guidelines to prevent prompt injection during AI content editing workflows are built into the structured elicitation process. These guidelines ensure repeatable refinement steps prevent data leakage while systematically pushing the LLM to improve its recent outputs.