Planning Fundamentals

Master how AI agents break down complex goals into executable steps

Breaking Down Complex Tasks

Task decomposition is the process of breaking a complex goal into smaller, manageable subtasks. It's like building a house: you don't just "build it" – you pour foundation → frame walls → install plumbing → finish interior.

Interactive: Goal Tree Builder

Click parent tasks to expand/collapse their subtasks. See how complex goals decompose hierarchically.

Book Vacation
Arrange Transportation
Find Accommodation
Plan Activities
💡 Tip: Each parent task contains 2-3 subtasks. This prevents cognitive overload.

Decomposition Principles

🎯

Top-Down Approach

Start with the main goal, then break it into 2-5 major steps. Repeat for each step.

Book Vacation → [Transportation, Accommodation, Activities]
🔗

Identify Dependencies

Some tasks must complete before others. Always do dependent tasks first.

Must book flights BEFORE booking airport transfer

Atomic Actions

Keep decomposing until each task is a single, executable action.

"Find hotel" → search_hotels() → compare_options() → book_hotel()

Real-World Example: Code Review Agent

Goal: Review Pull Request
Top-level goal is too vague to execute directly
Subtask 1: Analyze Changes
• Fetch PR diff
• Identify modified files
• Detect patterns (bug fixes, features, refactors)
Subtask 2: Check Quality
• Run linter
• Check test coverage
• Look for security issues
Subtask 3: Generate Feedback
• Summarize changes
• List issues found
• Post review comment