How I'm making our documentation agent-friendly (and how I'm not)

At Fern, the sole technical writer crafts documentation that caters both to human users and AI agents by leveraging a strategy that involves creating content that is not only readable but executable by AI. This approach entails embedding additional context in the form of llms-only tags to aid AI agents in executing tasks like setting up a documentation site autonomously while keeping human-facing content clean and streamlined. The documentation is designed to support AI agents in understanding and navigating systems with added procedural and contextual information that humans might not require. By breaking down long pages into focused sections and ensuring that content is up-to-date and easily accessible, Fern Docs ensures that both humans and AI agents can efficiently find and use the necessary information. The integration of platform-level scaffolding is key, as it indexes and structures the content in a way that is accessible to AI, highlighting the interdependence between content creation and platform support to facilitate effective AI interaction.