After you start predictive routing free trial, you can evaluate the ongoing value of using predictive routing on a queue using an 80:20 time split method. This method routes interactions using predictive routing 80% of the time interval and the baseline routing method 20% of the time interval. It increases the volume of predictive routing interactions, and allows for an increase in the benefits of predictive routing compared to a comparison test, which has a 50:50 split. You can compare predictive routing against the standard or bullseye baseline routing method for a particular KPI.

Set up ongoing value monitoring for a queue

Note: Before setting up monitoring, ensure that you completed at least a two-week 50-50 comparison testing period and observed a desirable benefit. For more information, see Test predictive routing on your queues.
  1. Click Admin.
  2. Under Contact Center, click Queues
  3. Click a queue name. 
  4. Click the Routing tab. 
  5. Under Routing Method, click Predictive Routing.
  6. In the Key Performance Indicator box, click the KPI to target.
  7. In the Predictive Routing Timeout box, type the timeout period (from 12-259,200 seconds).
  8. Under Predictive Routing Mode, do the following:
    1. Click Ongoing Value Monitoring.
    2. In the Ongoing Value Monitoring pane, click the routing method to compare against predictive routing.
  9. To use workload balancing with Predictive Routing, set Workload Balancing to On.
  10. To use skill matching with predictive routing, set Skill Matching to On.
  11. Under Timeout Routing Method, click the evaluation method to use when an interaction reaches the predictive routing timeout period and requires a fallback routing method. The fallback method is always standard routing.
  12. Click Save. The Queue list page appears, showing “Ongoing Value Monitoring” in the Predictive Routing Evaluation column. For more information about the processing phase, see The KPI processing phase.

Achieve a valid comparison

To achieve a valid comparison, Genesys recommends that you run ongoing value monitoring using the 80:20 time-split in six cycles of six days each. The 80:20 formula repeats as follows throughout the time period:

  • 8 hours using predictive routing; 
  • 2 hours using an alternative routing method; 
  • 8 hours using predictive routing; 
  • 2 hours using the baseline routing method

It takes six days to complete a full “cycle”; that is, every hour of the day used both predictive routing and the baseline routing method. It takes several of these cycles (up to two or three months) before you can consider the accuracy of the comparison as valid. If you change to another mode prematurely, we cannot ensure the accuracy of the comparison and reported benefit.

View results of ongoing value monitoring

Alternatively, you can also view the performance of the comparison test from the queue performance report. For more information, see View performance of queues with predictive routing