Visualizing Predictive Insights

Objective

After completing this lesson, you will be able to visualize the model runs as different out-of-the-box analytical graphics options and plot them using custom designed dashboards

What are Insights?

Insights are widgets that display data as charts and turn predictive data into clear, actionable analytics, helping you to quickly assess patterns in your predictive model results, understand where risk and opportunity lie both across time and among segments. With just a few clicks, you can add, configure, and interpret these widgets to inform business decisions.

The AI Workbench enables you to transform predictive model outputs into actionable insights through interactive dashboards and out-of-the-box (OOTB) analytical graphics. Whether you're evaluating customer churn risk or estimating Customer Lifetime Value (CLV), these built-in graphics make it easy to visualize trends, distributions, and risk profiles.

In this lesson, you’ll discover how to bring your AI model insights to life using dashboards in the AI Workbench, including how to create, customize, and organize dashboards to visualize your model results and make data-driven decisions with confidence.

Out-of-the-Box Churn Insights

The AI Workbench provides several out-of-the-box (OOTB) analytical charts to visualize Churn model predictions. Let’s briefly look at each.

Churn Probability by Risk Level

Chart Type: Pie Chart

Purpose: Shows what percentage of your customers fall into high, medium, or low churn risk.

Usage: Instantly assess overall risk composition in your customer base.

Pie chart labelled Probability by Risk Level over a 2 month period. The chart segments are 43,987 for Low, 7,0113 for Medium, and zero for High.

Churn Probability Distribution

Chart Type: Histogram

Purpose: Displays the number of profiles associated with low to high churn probability.

Usage: See if churn risk is concentrated among a few customers or spread across the base.

Bar chart showing the Churn Probability distribution. It shows 42,987 profile as values less than 0.1, 6,064 between 0.11 and 0.2, 982 between 0.2 and 0.3, and 7 between 0.4 and o.4. All other rows are zero.

The other OOTB Churn Insights are Churn Risk Over Time, which visualizes how churn risk levels (high, medium, low) change over days, weeks, months, or years; Churn Risk by Segment, which compares churn risk levels across customer segments (e.g., region, demographic group).

Out-of-the-Box CLV Insights

Similarly, the AI Workbench offers several OOTB graphics for Customer Lifetime Value (CLV) predictions:

Average Customer Lifetime Value (CLV)

Chart Type: Pie Chart

Purpose: Illustrates how average CLV is distributed among high, medium, and low value tiers.

Usage: Understand revenue risk and opportunities within your customer base.

Pie chart showing the proportion of High, Medium, and Low Average CLV values. 3.5K are High, 1.3K are Medium, and 243.7 are Low.

Average Customer Lifetime Value (CLV) Over Time

Chart Type: Line Chart (Time Series)

Purpose: Shows the trend of average CLV across levels over time.

Usage: Identify if your average customer value is rising or falling.

Line chart with horizontal axis showing calendar weeks, and vertical axis showing CLV values. Three lines are at zero until the 4th week, at which point the line labelled Low rises to about 200, the line labelled Medium to about 1200, and the line labelled High to about 3500.

The other OOTB CLV Insights are Customer Lifetime Value (CLV) Profiles, which shows the proportion of customer profiles in each CLV segment (high, mid, low); Customer Lifetime Value (CLV) Profiles Over Time, which tracks how the composition of CLV segments changes over time.

Designing New Dashboards

Dashboards are powerful tools that let you view all your key information—like Customer Churn and Customer Lifetime Value (CLV) predictions—on a single screen. This makes it easy to monitor your most important analytics at a glance.

In this section, you’ll learn how to design new dashboards tailored to your needs, add specific insights such as Churn and CLV into dedicated dashboard areas, and efficiently manage your dashboards by cloning or erasing them as your requirements change.

Creating a New Dashboard

To create a new dashboard, please follow the step-by-step instructions.

  1. Go to Dashboards.

    Click the Dashboards icon on the AI Workbench side panel.

    Fragment of the Dashboard page highlighting the Dashboards button on the left.
  2. Enter Split Screen mode.

    Click the Split Screen icon to view dashboards alongside other workbench areas.

    Dashboard page highlighting the Split Screen icon at the top left. below it, various split screen configurations are displayed.
  3. Create a New Dashboard.

    Click Create to open the Create Dashboard dialog.

  4. Enter Name and Description.
    • Enter a name for your dashboard.
    • (Optional) Add a description to summarize its purpose (e.g., "Churn & CLV Monitoring").
    The Create Dashboard dialog has two properties: name (Demo Dashboard) and Description (Churn & CLV Monitoring). The OK button at the bottom right is highlighted.
  5. Save the new dashboard.

    Click OK to save your new dashboard.

Customizing Dashboard Layouts

You can adjust how your insights are arranged on the dashboard.

To change the layout:

  • Click the Edit Layout icon on the top bar.
  • Select a new layout (e.g., 2x2 grid, two row/column, etc.).

    Existing insights will be rearranged automatically.

    Note:

    If you select a layout with fewer placeholders than the number of insights, only the first insights will be shown.

Adding Churn and CLV Insights into Dashboard Areas

You can add Insights to your dashboard to display specific metrics, such as Churn or Customer Lifetime Value (CLV).

  • Open your dashboard.

    Select your newly created dashboard from the dashboards list.

  • Add an insight.

    Click Add Insight.

    Dashboard page showing the Demo Dashboard, which contains four tiles, each one with a highlighted Add Insight link.
  • Choose the Insight Type.

    In the Insight Gallery, choose the insight you want to display, such as Churn or CLV.

    Choose Insight dialog showing available insights, such as CLV Profiles, CLV Profiles Over Time, Average CLV, Average CLV over time (highlighted), Curn risk over time, and Churn Probability by Risk Level.
  • Add the Insight to the dashboard.

    Click a specific Insight to place on your dashboard layout.

    Dashboard containing four tiles: Average CLV Over Time, Average CLV, CHurn Probability by Risk Level, and Churn Probability Distribution.
  • Configure Insight details.

    Click the vertical three-dot More icon to adjust:

    • Solution: Choose between Profile or Group predictive indicators.
    • Predictive Indicator: Select your Churn or CLV indicators.
    • Run ID: Pick a specific model run to visualize.
    • Period: Set the time range (Year, Month, Week, up to 1 year).
    • By: Choose the X-axis unit (e.g., Month, Week, Day).
    Settings pop-up invoked when clicking the vertical three dots showing settings: Solution set to Profile, using the clv_2mo predictive indicator over a month by weeks.

Copying and Deleting Dashboards

If you want to clone an existing dashboard as a starting point:

  1. Open the dashboard you wish to copy.
  2. Click the Copy icon.
  3. A duplicate will appear in your dashboard list titled with "…→ Copy".
  4. Edit the name and insights as needed.
Dashboard page showing four tiles. At the top right of the page, the Copy icon is highlighted.

To remove a dashboard:

  1. Open the dashboard.
  2. Click the Delete icon (trash can).
  3. Confirmation the deletion in the Confirmation dialog.
Dashboard page highlighting the Delete icon at the top right of the page and showing a confirmation dialog for the deletion action.

Summary

In this lesson, you learned how the AI Workbench allows you to visualize predictive model outputs—such as customer churn risk and Customer Lifetime Value (CLV)—using a range of out-of-the-box analytical graphics and customizable dashboards. You explored key insight widgets, including pie charts, histograms, and time series, which help you quickly interpret risk levels, value distributions, and trends for both churn and CLV predictions.

You also discovered how to create, customize, and organize dashboards to display your chosen insights, adjust layouts, and manage dashboards by copying or deleting them as your needs evolve. These capabilities empower you to monitor critical predictive analytics at a glance and make informed, data-driven business decisions.

Learning

SAP FacebookSAP YoutubeSAP LinkedIn