How AI Helped Jeff Sheltren Rebuild Tag1’s Internal Documentation

The project started when AI training materials and related documentation began to accumulate in Tag1’s Notion workspace without structure. Although the content was high quality, it became increasingly difficult to navigate. Faced with the daunting task of auditing and reorganizing it, Jeff turned to Claude Code integrated with the Notion MCP API. AI quickly inventoried the documentation, highlighted duplicates, and mapped the existing structure, allowing for rapid analysis that would have taken days manually. With that inventory, Jeff collaborated with AI to develop a reorganization strategy that considered different user journeys and structural approaches, such as whether to continue using Notion databases or flatten the architecture. He iterated on the plan with AI but maintained firm control to prevent unintended content generation. 

Once a plan was finalized and version-controlled, AI executed the migration into a sandbox environment. Claude copied content, preserved formatting, removed redundancy, and updated internal links. Multiple rounds of review followed, including checks by AI editor and Notion expert agents. Once validated, the reorganized content was pushed live, resulting in immediate usability improvements across the team. Sheltren emphasized that the project succeeded because AI was treated as a tool, not a decision-maker. The structured approach, human oversight, and technical integration allowed AI to shine in repetitive but critical tasks, transforming a long-overdue cleanup into a practical win for internal operations.