Skip to Main Content
Lightning Talk Intermediate NIl

Making Docs Work for Humans and AI Agents

Proposal status is Review Pending
Session Description

Content that works for human readers may lose essential context when retrieved independently by AI agents.  Missing prerequisites, implicit assumptions, and incomplete examples can lead to unreliable AI responses.

Drawing on my experience documenting cloud-native products such as NeuVector and Dash0, this session explores how to make documentation useful for both humans and AI agents.

Key Takeaways
  1. Structure documentation into self-contained, retrieval-friendly sections.

  2. Make prerequisites, parameters, constraints, and examples explicit.

  3. Understand how RAG, llms.txt, Markdown, and MCP support documentation discovery and access.

  4. Identify risks such as stale content, ambiguous instructions, and prompt injection.

  5. Improve existing documentation for AI consumption without compromising human readability.

References

Session Categories

Technology architecture
Engineering practice - productivity, debugging
Talk License: NIl
Speakers are hidden and anonymized for this event.