Insight

Most company knowledge ends up trapped inside tools you do not own. Notion, Confluence, SharePoint, a wiki that nobody updates. The information is technically there, but moving it, sharing it, or doing anything useful with it outside the original platform is harder than it should be.

Google recently published the Open Knowledge Format (OKF), a specification that takes the opposite approach. Knowledge is stored as plain markdown files with a thin layer of YAML metadata. It can be read without any tooling, versioned in git, and moved between systems without a migration project. It is currently a version 0.1 draft, but the thinking behind it is sound.

The more interesting angle is what OKF implies about how businesses should treat their knowledge. Most organisations accumulate information rather than maintain it. Documents get created, rarely get updated, and slowly become unreliable. OKF treats knowledge as something worth designing: structured, tagged, cross-linked, with a clear type and a modification timestamp. That discipline, knowing what you know, keeping it current, and making it findable, is the actual value. The format is just the container.

Where this connects to AI is practical rather than theoretical. RAG-based systems like Pixel Assist answer questions by retrieving content from your own documents. Before that works well, your knowledge needs to be in reasonable shape. Part of what we do when onboarding clients to Pixel Assist is exactly this cleanup work. OKF does not automate that process, but it gives you a sensible target format, one that benefits employees and customers before AI is even part of the picture.

That said, OKF is not a silver bullet. A well-structured knowledge base that nobody maintains will drift out of date just as quickly as a poorly structured one. The format makes the job easier. It does not do the job for you.

Still have questions? Give us a call or send us a message to understand more without any pushy sales tactics.