Keep Your Docs Fresh: An AI-Assisted Documentation Workflow

This video walks through a practical workflow for keeping product documentation accurate and up-to-date as your codebase evolves. Brian demonstrates how he maintains draft documentation files directly inside his project's codebase (as markdown files), then uses his AI coding agent to automatically update those docs at the end of each development session. The key insight: after making changes in a long conversation with your agent, you can prompt it to review your docs folder and fix any content that's now inaccurate—while the context of what just changed is still fresh.

The second half shows this workflow in action, taking those draft docs and turning them into a polished public documentation site. Brian demonstrates how he:

• Copied his in-progress markdown docs into his Builder Methods site repo
• Prompted Claude Code (using Opus 4.5) to create a full multi-page docs section
• Referenced an existing section (Agent OS docs) as the design pattern to follow
• Got a production-ready result with minimal fixes needed

Key takeaways:

• Keep draft docs in your codebase so your agent can access and update them
• Update docs immediately after feature work, while context is still in the conversation
• Reference existing design patterns in your codebase for consistent results
• Use plan mode in Claude Code for larger structural changes