User can navigate directly to the Birst reports to see forecasts and additional insights generated by the ML model.

Forecasts are shown at different granularity levels and allow the user to select the appropriate timeframe for the analysis. Other charts include but are not limited to year over year sales, monthly sales, historical sales vs forecasted sales vs benchmark forecasted sales, top contributing departments and locations, to cost savings and to the potential revenue increase.

As data gets incrementally updated, the workflow triggers a series of Coleman quests to run automatically on a daily, weekly, or monthly basis, depending on preset configurations. These quests include data preprocessing, staging observations, data cleansing, feature engineering, forecast generation and ML model evaluation and reporting

The Forecast Generation Quest takes in transformed data as input and after feature transformation, takes care of training the Machine learning model using not only the history and the underlying patterns in the data but also the features. The forecast is then generated at selected levels, evaluated, and compared against a subset of the history, before outputting forecasts that are stored in different data lake objects to be consumed for reporting purposes, by ERPs or other Cloud suite products.