Grooming Fundamentals
Why AI-driven sprints still need refinement
Story grooming in BMAD takes raw stories from the PM phase and makes them genuinely ready for implementation — unambiguous, correctly sized, and free of any interpretation choice the AI would otherwise make silently.
Grooming is often treated as an optional refinement step. In BMAD it is mandatory, because AI agents do not ask clarifying questions the way a human developer would. A human developer reading an ambiguous story might ask "does this mean X or Y?" before starting. An AI will pick one interpretation and implement it. If it picked the wrong one, the work needs to be redone.
The grooming session takes a story from "written" to "ready." Ready means: the acceptance criteria cover both success and failure paths; the scope is appropriate for a single session; all dependencies are identified and either satisfied or explicitly flagged; and any term in the story that could be interpreted more than one way has been resolved before implementation begins.
BMAD supports several grooming patterns — Party Mode, Sequential Grooming, and Scope Hammering — each suited to different kinds of ambiguity and team contexts. All produce the same output: a story that is unambiguous, correctly sized, and ready for a single-session implementation. The right pattern depends on how much ambiguity the story contains and how many perspectives are needed to resolve it.
Key Concepts
- Grooming is mandatory — AI does not ask clarifying questions before implementing
- Ready means: AC complete, scope correct, dependencies explicit, terms unambiguous
- Grooming resolves every interpretation choice before the AI makes them silently
- Multiple grooming patterns exist for different types and depths of ambiguity
- A story that fails any readiness check is not ready — grooming is not done