
Per Colibri's product documentation, the platform runs a Best Fit algorithm that selects a closest-fitting statistical model per demand series, supplemented by external variables, correlation analysis and intelligent clustering. The output is one forecast number per item-period pair, plus safety stock optimization, constrained-plan automation and a conversational AI assistant that, per Colibri's homepage, provides intelligent user support, application control and contextual analysis on top of the planning data. The forecasting layer stays point-based, not probabilistic.