Session abstract
What you’ll learn
Documentation is treated as an afterthought at most companies. It gets a paragraph in the launch checklist, a junior engineer in a hurry, a platform nobody owns, and a quality bar set by whoever shipped the feature. That was always a problem. AI made it impossible to hide. When we migrated Commerce's developer documentation (covering three products: BigCommerce, Makeswift, and Feedonomics) from a legacy platform to Fern on a hard launch deadline, the migration only worked because we stopped treating docs as a deliverable and started treating them as a product. That meant a roadmap, an owner, a quality bar, and a contributor workflow that scaled. AI showed up in three roles along the way: as a contributor accelerating the migration itself, as an editor surfacing gaps in content we thought was fine, and as a reader, the one sitting between every developer and our docs through Cursor, Claude Code, ChatGPT, and our embedded Ask AI. This talk walks through that migration honestly. What changed when we treated docs as a product. How AI helped us ship and what it exposed when we did. A four-layer model (content, structure, surface, contributor workflow) we used to keep the work coherent. Takeaways: - Why "docs as a product" is a real organizational shift, not a slogan, and what changes when you make it - The three roles AI plays in modern docs work: contributor, editor, reader - A four-layer model (content, structure, surface, workflow) for organizing AI-native docs work