Mintlify automates 95% of Decagon docs
Decagon built a self-healing docs-as-code agent workflow atop Mintlify to automatically draft 95% of documentation updates from code commits and Slack announcements. The setup increased monthly merged pull requests by 200% while keeping customer-facing AI agent Duet grounded in fresh, accurate knowledge.
Documentation is shifting from an afterthought into critical infrastructure because AI agents fail immediately when underlying knowledge bases go stale. Decagon's workflow demonstrates that automated docs maintenance is becoming mandatory for teams operating in-production AI agents.
- –Closed-loop agent reliance: Decagon's customer agent (Duet) answers queries directly from docs, turning documentation freshness from an editorial chore into an operational requirement.
- –Docs-as-code enables agentic edits: Because Mintlify stores documentation as GitHub Markdown, background agents can parse commit diffs and open pull requests directly.
- –Multi-team synchronization: Over 10 internal teams at Decagon have adopted the docs agent to keep internal guides, runbooks, and presentation decks synchronized.
- –Enforced guidelines via AGENTS.md: Structured prompt instructions and custom skills maintain voice and consistency, keeping engineers and PMs focused on reviewing rather than writing.
DISCOVERED
4h ago
2026-11-16
PUBLISHED
4h ago
2026-11-16
RELEVANCE
AUTHOR
mintlify