A product manager asked where the engineering team saw the most issues, so an engineering lead used Claude on Jira via an MCP connector. One team had about 50% of sprint time spent on “bugs,” versus roughly 25% for everyone else, but context revealed almost none were bugs—they were manual workarounds for a missing product capability. Not shipping an item restore feature was burning roughly 1.5 engineers’ worth of capacity. The analysis took 45 minutes and was possible because data was organized, tagged by team, connected to contributors, accessible through MCP, and protected by role-based access. The author, a CTO, calls the value of context data layers “contextual intelligence,” noting Anthropic’s engineering team has called similar work “context engineering” since late 2025. At a recent company hackathon, nine engineering teams built AI tools on an MCP server, but initial demos had the same problem: Claude could talk to data, but answers were either generic or confidently wrong.