Agentic AI workflows expose 'vision drift' gap in current project management tools

A developer building an issue tracker called Epiq has identified a problem he terms 'vision drift', where AI-driven agentic workflows cause implementation to stray not just from architecture but from the original product intent. Existing issue trackers capture only a current snapshot of work, making it difficult to trace how decisions and tasks evolved over time — unlike code, which has Git history. The author observed this gap firsthand after an AI agent independently completed a feature in two hours, leaving him unable to reconstruct how the work had unfolded without real-time monitoring. Epiq addresses this through an event-sourced architecture that allows users to time-travel through project history and replay sequences of board activity. The author argues that as multi-agent collaboration becomes more common, traceable issue history will shift from a novelty to a critical oversight tool.