We are evaluating Genesys Predictive Routing. I must extract the specific routing methodology utilized for each interaction via the /api/v2/analytics/conversations/details/query endpoint. I need to distinguish whether an interaction was routed via Standard Routing, Bullseye Routing, or Predictive Routing. The standard views provide aggregate data, but our data warehouse requires the exact routing type per conversation. Which specific dimension or metric in the JSON payload dictates the routing method that ultimately delivered the interaction to the agent?
Greetings! This is a phenomenal use case for the detailed analytics API! When you extract the conversation details, you must look within the acd participant block. Inside that specific block, you will find an array of sessions. Each session contains a routingStatus and a segments array.
The critical property you are searching for is named usedRouting. This property will explicitly state the routing method that was utilized, such as Predictive or Standard or Bullseye.
Additionally, if you are analyzing the efficacy of the predictive model, you should also look for the proposedAgents array within that same segment, which details the specific agents the machine learning model evaluated before making its routing decision!
Hey! Just a quick heads up on this. We tried pulling that same usedRouting field for a gamification dashboard we built to reward agents who handled the hardest predictive calls. Sometimes the field just says Standard even if you have Predictive Routing turned on for the queue! It turns out that if the predictive model times out or if there is not enough historical data for the caller, it falls back to standard routing automatically.
You have to check the routingReason field as well to see if it fell back due to a timeout. Do not just assume every call in a predictive queue is actually routed predictively!