Operational Forecasting now includes enhancements that improve forecast configuration, transparency, validation, access management, and data residency.
The following features are now available:
How Forecasting Works
A new How Forecasting Works section has been added to the Job Setup page.
This section explains the main stages performed behind the scenes when a forecast is generated using the Recommended Path:
Time Grid Creation. Missing dates are filled in to create a complete timeline for each series
Feature Engineering. Additional features are generated automatically, such as calendar based features, lag values, and rolling window calculations
Intelligent Routing. Each series is analyzed to determine whether the demand pattern is stable or intermittent
Model Assignment. The most appropriate forecasting model is selected based on the series classification
The section is collapsed by default and can be expanded by clicking the section title.
.png)
Forecasting workflow enhancements
What’s new?
Create New Job
When creating a job, users can now select the forecasting execution path.
The Recommended Path option is selected by default and uses the existing forecasting engine and standard job creation workflow. When this path is selected, users can also add an optional covariate dataset to include external business drivers in the forecast calculation.
Users can select a file classified as a Covariate and choose one or more numeric columns to use as external drivers. The covariate dataset must include a date column, use the same hierarchy columns as the target dataset, and cover the required forecast horizon.
.png)
Users can also select the new Univariate Approach, which introduces multiple statistical forecasting algorithms that can be executed at the same time within a forecasting job. Unlike the Recommended Path, which returns one forecast generated by the forecasting engine, the Univariate Approach allows users to compare outputs from different forecasting techniques for the selected hierarchy.
Covariate datasets are not available for jobs created using the Univariate Approach.
.png)
Job Executions
The Job Executions section now includes a Type field.
This field shows which forecasting execution path was selected when the job was created: Recommended Path or Univariate Approach.
Job Results
The Job Results window now displays additional information and validation options based on the forecasting approach used to create the job.
When applicable, the job information area displays the external drivers included in the forecast calculation. This allows users to confirm which additional datasets were used to generate the forecast.
For jobs created using the Univariate Approach, the results table includes one forecast column for each executed algorithm. The column headers correspond to the algorithm identifiers, allowing users to compare algorithm outputs in the same results table.
The selected calculation method and hierarchy level are applied consistently to all forecast columns.
(1).png)
For jobs created using the Recommended Path, the new Holdout Results tab displays forecast-validation results. Holdout validation automatically compares forecasted values with actual historical values from the target dataset, helping users evaluate forecast accuracy.
The tab includes key performance metrics such as MAPE, Forecast Error, and WAPE. The displayed values and metrics are recalculated according to the selected calculation method, hierarchy level, and frequency. The target dataset must contain at least three years of historical data for holdout results to be generated.
.png)
Business benefits
These enhancements help planners:
Include relevant external business drivers in forecast calculations
Review how target and covariate datasets are combined before creating a job
Confirm which external drivers were used to generate a forecast
Compare multiple forecasting methods side by side
Validate forecast outputs more easily
Export results for downstream analysis
Identify the most effective forecasting method for their data
Use consistent aggregation and drill-up capabilities across forecast results.
Centralized access management for Foresight and Operational Forecasting
Applies to: Company Admins
Foresight and Operational Forecasting access can now be managed centrally from the Board Subscription Hub using the existing Developer and Power User licensing model. Administrators can grant or revoke access from the Licenses section for eligible users, with Operational Forecasting access requiring both Foresight entitlement and a Developer or Power User license.
Access granted through the Subscription Hub applies across all environments associated with the subscription, reducing manual license management and helping keep user permissions aligned across Board, Foresight, and Operational Forecasting.
For more details, see Licenses.
Multi-region support
Applies to: Company Admins
Operational Forecasting now supports customer environments hosted in multiple geographic regions, currently the United States and Germany. When a new environment is provisioned, it is assigned to a region based on the customer’s geographic location and contractual requirements. Operational Forecasting data is then stored within that assigned region.
This enhancement helps organizations meet data residency, governance, and sovereignty requirements by ensuring customer data remains in the appropriate geographic region and is not automatically shared, synchronized, or transferred across regions.