Skip to main content

PRD Planning Workflow

The PRD Planning workflow converts product requirements documents into Azure DevOps work item hierarchies, decomposing requirements into the four-level structure (Epic > Feature > Story > Task) that Azure DevOps supports natively.

When to Use

  • 📄 A product requirements document needs conversion to work items
  • 🏗️ Building an initial backlog from a specification or design document
  • 🔗 Requirements need traceability from document to backlog items
  • 📊 Converting a large requirements set into a structured work item hierarchy

What It Does

  1. Accepts a PRD, specification, or requirements document as input
  2. Delegates to the @AzDO PRD to WIT agent for parsing and decomposition
  3. Maps requirements to Azure DevOps work item types (Epic, Feature, User Story, Task)
  4. Builds parent-child relationships following the four-level hierarchy
  5. Produces a handoff file with the complete work item hierarchy ready for execution

NOTE

PRD Planning delegates to a specialized agent (@AzDO PRD to WIT) that handles the document parsing and hierarchy construction. The ADO Backlog Manager orchestrates the handoff and provides the execution path.

Hierarchy Model

Azure DevOps supports a four-level work item hierarchy. PRD Planning maps requirements to the appropriate level based on scope and granularity:

LevelWork Item TypeTypical Scope
Level 1EpicBusiness initiative or major objective
Level 2FeatureFunctional capability or component
Level 3User StoryUser-facing requirement or scenario
Level 4TaskImplementation step or technical work

Requirements that span multiple features become Epics. Requirements with clear user value become User Stories. Implementation details become Tasks under their parent Stories.

Output Artifacts

.copilot-tracking/workitems/prds/<prd-name>/
├── artifact-analysis.md # Extracted requirements and field mappings
├── work-items.md # Proposed work item hierarchy
├── planning-log.md # Decomposition decisions and progress
└── handoff.md # Execution-ready operations

How to Use

Option 1: Handoff Button

Click the "PRD" handoff button in the ADO Backlog Manager agent. This delegates to the @AzDO PRD to WIT agent with your document context.

Option 2: Direct Reference

Reference your requirements document in a conversation with the ADO Backlog Manager:

Convert this PRD to Azure DevOps work items: [path/to/requirements.md]

Option 3: Inline Content

Paste requirements directly into the chat:

Create a work item hierarchy from these requirements:
1. Users can search by keyword
2. Search results display in a paginated list
3. Results can be filtered by date range

Example Prompt

Parse the product requirements document at docs/prd-v2.md and create
an Azure DevOps work item hierarchy. Use Epics for major features and
decompose each into Stories with Tasks for implementation steps.

Tips

  • ✅ Provide a structured document with clear requirement boundaries for best results
  • ✅ Review the proposed hierarchy before executing to verify parent-child relationships
  • ✅ Use the execution workflow to apply the hierarchy after review
  • ✅ Combine with sprint planning to assign the created hierarchy to iterations
  • ❌ Do not mix PRD planning with manual work item creation in the same session
  • ❌ Do not skip hierarchy review before execution (parent-child errors are harder to fix)
  • ❌ Do not expect PRD planning to handle ongoing triage (use the triage workflow instead)

Common Pitfalls

PitfallSolution
Requirements too vague for decompositionAdd specificity to the source document before conversion
Hierarchy too deep or too shallowAdjust the decomposition level in your prompt
Duplicate work items from repeated runsCheck existing backlog items before re-running PRD conversion
Missing parent-child linksVerify the handoff file before execution

Next Steps

  1. Review the proposed hierarchy in the handoff file
  2. Use the "Execute" handoff to apply the work item hierarchy to Azure DevOps
  3. Continue with Sprint Planning to assign iterations to the new items

🤖 Crafted with precision by ✨Copilot following brilliant human instruction, then carefully refined by our team of discerning human reviewers.