Predictive routing benefit assessment

Prerequisites
  • Routing > Assessment > All permissions
  • Routing > Key Performance Indicator > View

The benefit assessment evaluates the potential for improvement per key performance indicator (KPI), per queue. The assessment directs you to the queues and KPIs that promise the best results. If a queue contains different media types with different optimization potential results, the benefit assessment displays only the result with the highest optimization potential.

Considerations

  • Running a benefit assessment does not change any queue settings or impact current queue operations in any way. 
    • Running a benefit assessment incurs no cost. To view your bill and verify that there are no charges, navigate to Admin > Subscriptions.
    • Running a benefit assessment does not require a product subscription. 
    • The benefit assessment runs on all queues for which you have permissions. Filtering the queue list does not affect which queues are included in the benefit assessment. For example, if you filter the view to show only the queues in a selected division, the benefit assessment still evaluates all queues you have permissions to view.
    • After you run a benefit assessment for an organization, the Learn More link no longer appears in the Predictive Routing Evaluation column header. After this initial benefit assessment runs on all queues, you must run later benefit assessments on a queue-by-queue basis.

    Start a benefit assessment

    The following video shows you the steps to run a benefit assessment. The written procedure follows the video.

    1. Click Queues from the Admin console. The Queue list page opens. If you have not run benefit assessment in your organization yet, the Predictive Routing Evaluation column includes a Learn More link. 
    2. Click Learn More to start the benefit assessment for all queues configured in your Genesys Cloud account. The Optimize Routing with AI side pane opens.
    3. Understand the various phases of predictive routing and click Start Evaluation. 
    4. Ensure you have the necessary permissions to begin benefit assessment. Click Start Queue Evaluation. 
    5. The pane closes and the Predictive Routing Evaluation column on the Queues page shows a Processing label on each queue. 
    6. To view updated status information, refresh the Queues page.  
      • A green icon indicates high optimization potential.
      • A yellow icon indicates low optimization potential.
    7. To view the optimization potential for each KPI, click the Optimization Potential link in the Predictive Routing Evaluation column.
      • The benefit assessment result pane shows the optimization potential for each available KPI. It also lists the factors used to assess each KPI for its optimization potential. This detailed breakdown enables you to see where data is inadequate or missing, opening the potential for better results based on other data.
    8. To rerun the benefit assessment for a queue, click the Re-run Assessment icon in the upper-right section of the Queue Benefit Assessment pane.
    9. The options in the next step depend on whether you have purchased AI Experience tokens earlier:
      • If you have not purchased the AI Experience tokens –  Click Purchase AI Experience Tokens. The option redirects you to AppFoundry, where you can purchase AI Experience tokens. After your purchase is processed, return to the Queues window, click Learn More in the Predictive Routing Evaluation column. See Next steps (at the end of this article) to proceed.
      • If you have purchased the AI Experience tokens –  Click Start Routing Comparison Test or Activate Predictive Routing. See Next steps (at the end of this article) to proceed.

    If you create a new queue after initially running a benefit assessment, a Start Evaluation link appears in the Predictive Routing Evaluation column for the new queue. Click this link to run a benefit assessment on the new queue. 

    Understanding your benefit assessment results

    Genesys Cloud runs various checks for each predictive routing KPI, such as the following:

    • Whether there are enough inbound interactions for the most recent 90 days. For robust results, each assessed queue must each have the following minimum numbers of interactions in the most recent 90-day period:
      • 45 days within the required 90-day period having at least one inbound interaction recorded per day.
      • At least 900 inbound interactions in total.
    • Whether the current KPI values show potential for optimization:
      • Average handle time is greater than 180 seconds. Calls shorter than 180 seconds generally have a lower potential for optimization. 
      • The custom KPI value is above 2 percent for the queue.
    • Sufficient variation in agent performance on the queue to allow for predictive routing to provide benefit.
    • Whether AI model tests succeed:
      • If the checks listed previously are successful, predictive routing creates an AI model and tests it. If these tests fail, it means that, based on the currently available data, predictive routing cannot accurately predict agent performance. This can change over time as queue conditions – such as volume, type of workload, and number of agents – change.
    • You can activate predictive routing on a queue that shows low optimization potential. However, the results are likely to show less benefit from predictive routing than you can expect from queues assessed to have high potential.

    Next steps