TL;DR
A RAND Corporation study of industry AI teams reported that rigid Scrum routines can be a “poor fit for AI projects,” since machine learning work often requires an initial research or data exploration phase of unpredictable length. Forcing exploratory AI development into uniform sprint boxes causes inefficiency. This mismatch can lead to frustration and Agile ceremonies that feel like overhead in AI initiatives. The key issue is that AI development involves iterative data tuning and model experimentation that don’t always deliver tangible increments every sprint. Without adaptation, traditional Agile metrics (like velocity or burndown) may fail to capture progress, and teams risk stakeholder misalignment.
You can deep-dive into this emerging trend as we seek to adapt and iterate over this rapidly changing employment landscape in my newly created blog where I keep things candid, clear, accurate and without salesmanship. Enjoy!