Signals in Budget Manager Experience

What is Predict Signals?

Predict Signals detects and flags potential anomalies in your data and enables you to take corrective actions through data-driven insights. These anomalies could arise from typos, mistakes, erroneous formulas, or conscious changes in planning assumptions. It gives you a detailed context around each signal by using an intuitive graph and lets you collaborate with your colleagues and peers to make more informed decisions.

How do Predict Signals work?

The Predict Signal's AI Engine, powered by machine learning and statistical/mathematical modeling, generates a predicted range representing the normal data ranges for a given dimension intersection. Any values falling outside of that range are tagged as Signals. These Signals are categorized as High risk, Medium risk, and Low risk depending on how far off they are from the predicted normal range.

AI Engine of Predict Signals

Predict Signals has a robust machine learning pipeline that handles the entire data science cycle, starting from data pre-processing till model deployment and management.

The Model training phase in this ML pipeline is equipped with many time-series algorithms, statistical methods, and classical ML techniques with advanced ensemble and stacking methods to ensure consistent and accurate results. The Model training time is optimized to a maximum extent using multi-threading and parallel executions. You cannot change the algorithmic settings for the Model training as they are designed to adapt to varied customer data.

Advantages

Predict Signals allows you to analyze scenario data, spot anomalies, and improve your projections’ accuracy by providing more context around the spotted signals. In addition, it provides a consolidated data analysis report on the dashboard, and lets you collaborate with colleagues or respective owners of various data sets, and facilitates you to take corrective actions.

• Timely identification of Anomalies: Signals could be displayed due to erroneous data entries like typos, mistakes in formulae, or fat-finger errors. Predict Signals helps in the timely identification of these anomalies and facilitates avoiding these expensive mistakes.

• Efficient Planning: Makes work more efficient by helping you fix errors in time and by letting you collaborate with peers. It helps you customize your dashboard according to your requirements and view only the data that you deem necessary.

• Collaborative Decisioning: If a signal is ever a result of a conscious change in planning, you can easily collaborate with your colleagues to take appropriate action.

• AI-Powered Actionable Insights: Helps to turn your data into actionable insights and empower businesses to make data-driven decisions.

Predict Signals in Budget Manager Experience

Budget managers can use the Predict option on the menu bar or select any cell in the template and right-click on it to access the Predict Signals options. The following options are available to generate signals:

• Check All Lines generates signals instantly for smaller reports of up to 50 lines.
• Check Selected Line(s) allows you to check signals for a maximum of 10 lines in one go.

The cells having signals are highlighted, and budget managers can right-click on that cell to either Resolve the signal or use the Signal Context Screen option to navigate to the Signal Context screen to analyze the signal.

Resolve a Signal

Following are the steps to resolve a signal:

1. Right-click on the cell with the signal.
2. Select the Resolve Signals option. The Resolve Signals pane will appear. 
3. Select the reason for resolution from a list of pre-existing options. If you resolve multiple signals simultaneously, the same reason will be applied to all.

The resolved signal will now appear as a comment in the report. The comment will display the original cell value, the reason for resolving the signal, and any additional information provided while resolving the signal.

Signal Context

The Signal Context screen displays the dimensions of the signal, a data table, and a graph that showcases all the required information on the signals. The Data table contains the following:

• The Budget/Forecast scenario for the entire fiscal year of the selected row.
• The upper and lower ranges generated by the AI engine.
• The historical actuals used for Model Training.

The graph displays the same information visually intuitively, which justifies why it was marked as a signal. By default, the graph has the upper and lower ranges and the forecast scenario, but it lets you see the historical actuals by clicking on the actual data label displayed below the graph. This screen displays data for all the months in the entire fiscal year to demonstrate the trend and seasonality present in the data; this helps understand the signal better.

You can even view the data trend of a selected dimension for the entire fiscal year and compare it with the historical actuals. You can view actual data for multiple years and view their trend along with the selected scenario, upper range, and lower range. It gives you a better understanding of every signal.

You can click on the PDF icon available on the Signal Context screen to download a snapshot of the signal context that you have chosen. This allows you to view and analyze the signal context offline, thereby reducing the dependency on application availability.

Budget managers can right-click on any signal in the signal context table and resolve the signal.

Comment on the Signal

Budget managers can right-click on any signal in the Signal context table and provide a comment specific to that cell and even tag other users in the comment. They can even assign the comment to another user and track the comment’s status.

Refer to the Collaborate with Other Users section available here to know more about the comments feature.

Drill Through

Budget managers can right-click on any signal in the Signal Context table to view the Drill Through option. This allows them to view and analyze all financial, operational, capital, workforce, and transactional data for that signal from a single place.

Budget managers can also use the History option in the Drill Through report screen to know the detail of the data changes on the source data. They can view information like who changed the data, the timestamp of the data change, and the details about the changes.

Following are other options on the Drill Through screen that budget managers can use to optimize the data they see on this screen:

• Label- It defines the way the data is shown in the drill-through report. The options available in this drop-down are Code, Name, and Label.
  • Code - Displays the code of the value mapped to the respective column if applicable.
  • Name - Displays the name of the value mapped to the respective column if applicable.
  • Label - Displays the code and name of the value mapped to the respective column. 
• Output - Exports the drill-through report in excel format.
• Filter - Filters the report based on the defined filter criteria.
• Settings - Uncheck the column names to view only the selected columns
• Save - Use the option to save the changes made to the report or reset the view to default view if you made any changes.

You can refer to this drill-through section if you want to know more about this feature.