Intelligent Forecasting Overview

Edit online

Intelligent Forecasting introduces advanced time-series forecasting techniques into Apptio Planning, empowering teams to produce more accurate, consistent forecasts with less manual effort. The feature analyzes historical spend patterns and automatically applies the most suitable model for each line item.

Key Features:
  • Automated model selection based on historical trends

    Intelligent Forecasting analyzes each line item’s time‑series pattern and automatically selects the most appropriate forecasting model.

  • Outlier detection to improve forecast accuracy

    Historical anomalies are detected allowing users to adjust them, so the model is trained on cleaner, more stable data—leading to more reliable forecast results.

  • More accurate and consistent forecasts

    Automated pattern detection (trend, level, seasonality), parameter tuning, and optimized smoothing produce stronger accuracy and reduce forecast variation between planners.

  • Increased transparency and control, with the ability to review and refine results

    Planners can inspect model settings, smoothing parameters, and standard accuracy measures. Forecasts are fully editable: users can accept, adjust, or override values as needed.

Enable Intelligent Forecasting

Note: Admin or Budget Process Owner roles are required to perform these tasks.
Steps to Enable Intelligent Forecasting
  1. Navigate to Settings (Gear icon) → Company Profile.
  2. Ensure New Expenses Page Experience (Beta) is enabled first—Intelligent Forecasting cannot be used without it. To enable the feature, follow these steps:
    • General Configuration → Access & Permissions → Select Enable New Expenses Page Experience (Beta)
  3. Select Forecasting → Enable Intelligent Forecasting
  4. Click Save and Exit button

Apply Intelligent Forecasting

Note: ITPPredictiveForecastingFeatureFullAccess permission is required to use the Intelligent Forecasting feature. Admin, Budget Process Owner and Cost Center Owner roles already have this permission.

Users can create or modify forecasts by defining the forecast periods and historical data in the Forecast dialog box.

  1. Navigate to the Expenses page and enable the New View toggle.
  2. Select the Expenses (Summary, Other, Assets, Contracts, Labor or Labor Activity) tab.
  3. Go to the Other tab.
  4. Select one or more line items by checking the box next to each item.
  5. Then choose Forecast either from the right‑click menu or by clicking the Forecast button.
  6. In the Forecast dialog, select the Historical Date Range by selecting the Start Period and End Period. This data will be used to generate the forecast.
  7. A minimum of 3 months of data is required to run Intelligent Forecasting.
  8. In the same dialog, configure the Forecast Period by selecting the Start Period and End Period.
    • The Forecast Start Period automatically defaults to the month immediately after the Historical End Period but can be overridden.
    • The Forecast End Period can be any month between the Forecast Start Period and the Plan End Period.
  9. To review the generated forecast, click Preview.
  10. You may switch between forecast models to compare outputs and statistical details.
  11. When satisfied with the forecast, select the model to apply and click Apply Selected Model to update the chosen line items.

Supported Expense Types for Intelligent Forecasting

Other & Summary

  • Other expenses can be forecasted from either the Other tab or the Summary tab.
  • Applying a forecast updates the corresponding financial values automatically.

Labor & Labor Activity

  • Supported for Labor and Labor Activity (Hours).
  • The forecast preview displays:
  • Headcount for Labor
  • Hours for Labor Activity
  • Applying a forecast automatically updates both labor and labor activity financials.

Contracts & Assets

  • Supported for Manual Amortization Contracts and Manual Depreciation Assets.
  • Forecasting is disabled for unsupported contract and asset rows.
  • Applying a forecast generates updated financials for both contracts and assets.

Important Notes

  • Intelligent Forecasting is not available for:
  • External lines
  • Generated financials
  • Submitted expenses
  • Actuals
  • Support for using Actuals as historical data for forecasting is planned for a future release.
  • Support for Intelligent Forecasting on grouped expenses is also planned for a future release.

Reviewing and Interpreting the Forecast Preview

Note: ITPPredictiveForecastingFeatureFullAccess permission is required to use the Intelligent Forecasting feature. Admin, Budget Process Owner and Cost Center Owner roles already have this permission.

The forecasting preview shows what the results for the forecast look like, based on your input. This gives you the opportunity to analyze and work with projected forecast data before applying it, helping you make more informed decisions.

  1. The Forecast Preview is available after clicking the Preview button on the Forecast dialog.
  2. The Preview Chart displays the visualization of the historical data followed by the forecasted data, which is represented as a dotted line.
  3. The Preview Chart shows predicted values, as well as a confidence interval that indicates high and low bounds within a shaded area.
  4. Prediction Accuracy score shows the overall forecast accuracy, indicating how closely the model’s predictions match historical data.
  5. Click on the Statistical details tab to review the details for the forecast.

To learn more about the models and statistical details, please see Forecasting Models

Identifying and Adjusting Forecast Outliers

Note: ITPPredictiveForecastingFeatureFullAccess permission is required to use the Intelligent Forecasting feature. Admin, Budget Process Owner and Cost Center Owner roles already have this permission.
Outliers

An outlier is any data point within the last 20% of your historical values that falls outside the model’s confidence range for that point in time. These points indicate moments where the actual value deviated more than expected from typical patterns in the data.

During forecasting, multiple models are evaluated and compared. Although the model‑fitting confidence intervals for historical points are not displayed, each fitted value has an estimated confidence range—similar to the intervals shown for the forecast.

Most historical points usually align closely with their model‑estimated values. However, when a point falls noticeably outside the expected confidence range, it is identified as an outlier. This signals that the historical value was unusually high or low compared to what the model anticipated.

Outliers can occur due to one‑time events, unusual spend spikes or drops, or data entry anomalies. Highlighting them helps you understand which historical values may be contributing to unexpected model behavior or reduced forecast accuracy.

Adjusting Outliers
  1. When Intelligent Forecasting identifies outliers in your historical data, you can view the outliers in Forecast Preview and Outliers Detected button is activated.
  2. Click the Outliers detected button on the preview. The Outliers detected pane opens.
  3. Click Adjust outliers to update the forecast and see the impact of the adjustment.
  4. When multiple outliers are detected, they are adjusted sequentially from the beginning of the data set. This means that each adjusted outlier value is applied to the calculation of the adjustment for the subsequent outlier, resulting in a more accurate forecast.
  5. When satisfied with the forecast, click Apply Selected Model to update the chosen line items.