Skip to main content
Documentation
Documentation
Adaptive Planning
Last Updated: 2026-05-15
Steps: Set Up Anomaly Detection

Steps: Set Up Anomaly Detection

  • Complete a financial, sales, or workforce planning model.
  • Ensure you have 24 months of data in your actuals version.
Anomaly Detection is an Intelligent Planning capability that uses machine learning. Anomaly Detection helps you build more precise plans by identifying profitable opportunities and avoiding unknown risks up to 24 time periods into the future.
To use Anomaly Detection, you must first generate predictions into a prediction plan version. The algorithm calculates predictions based on your actuals data. After you generate the machine learning predictions, you can:
  • Compare the predictions to your budgets.
  • Build outlier reports to identify anomalies in your data.
  • Use the Anomaly Detection capability in sheets.
You can't use anomaly detection with accounts:
  • With master formulas.
  • That are setup to read only.
  1. Select
    Month
     for your default time strata.
    Security:
    Configure Calendar
     permission.
  2. The machine learning version is a special plan version that holds the predicted values of the algorithm.
    Security:
    Versions
     permission.
  3. Build a modeled sheet so that you can define your own thresholds for each account, version, and level combination.
    Security:
    Model includes sheets, accounts, dimensions, and formulas
     permission.
  4. (Optional) Create Accounts.
    To set up existing and new general ledger and custom accounts for anomaly detection, select these account settings:
    Option Description
    Account Type
    Select
    Periodic
    .
    Data Entry Sheet Type
    Select
    Standard
    Start Expanded
    Select to expand accounts by default to help users identify anomalies.
    Security:
    Model includes sheets, accounts, dimensions, and formulas
     permission.
  5. To highlight account anomalies in standard sheets, remove custom dimensions from the sheet.
    Security:
    Model includes sheets, accounts, dimensions, and formulas
     permission.
  6. Security:
    • Access Sheets
       permission.
    • Model includes sheets, accounts, dimensions, and formulas
       permission.
  7. Machine learning predictions can take a few days to complete.
    Security:
    Model includes sheets, accounts, dimensions, and formulas
     permission.
The predicted data populates the prediction version. Intelligent Planning automatically compares the forecast to your other plan data and highlights anomalies in standard sheets. In sheets, anomalies are 40% more or less than the generated predictions unless you defined different thresholds with your modeled sheet.