Databricks opens predictive analytics functions in beta

Databricks has put Predictive AI Functions into beta, adding four SQL functions for identifying drivers of metric changes, detecting anomalies and predicting numerical or categorical values. It also updated its existing forecasting function, which now supports holidays and external factors such as pricing or marketing spending. Analysts can call the functions from SQL editors or notebooks and use results in dashboards and pipelines. Databricks maintains the underlying models centrally. Plain-language access through Genie is narrower: the function for identifying metric-change drivers is available today, while forecasting, anomaly detection and the two prediction functions are coming later.