Back to glossary

Sprint Velocity

A measure of the amount of work a team completes during a single sprint, typically expressed in story points. It serves as a planning tool for forecasting how much work the team can handle in future sprints based on historical performance.

Sprint velocity is a team-specific metric that should never be used to compare teams or as a performance target. Its sole purpose is improving forecast accuracy for sprint planning. By tracking how many story points the team consistently completes, product managers can make more realistic commitments about upcoming work and identify when capacity is being strained.

For AI product teams, velocity tracking has nuances because AI work often involves research spikes and experiments with unpredictable outcomes. A team might allocate story points to a model evaluation task, but the actual effort depends on how many iterations are needed to reach acceptable performance. Growth engineering teams should track velocity separately for deterministic feature work and experimental AI work to improve planning accuracy. When velocity drops unexpectedly, it often signals hidden complexity in data pipelines or model integration that should inform future estimation practices rather than pressure to work faster.

Related Terms