Hello everyone! I am so excited to be building a new supervisor dashboard! My goal is to send a proactive alert to our team leads before an SLA breach actually occurs. I want to calculate the ‘Predicted Wait Time’ for each interaction and trigger a notification if it exceeds eighty percent of our SLA threshold. Does the analytics API provide a ‘Predicted Wait’ metric, or should I be calculating this based on current queue occupancy and average handle time?
Welcome to the world of predictive analytics. I managed this in PureConnect using some complex SQL, but in Genesys Cloud, you have to do it all via API. There is no ‘Predicted Wait’ metric available out of the box in the analytics query results.
You have to build a logic engine that looks at the nOffered and nWaiting counts and factors in your tHandle average. It is a bit of a mathematical puzzle to get it right.
I have been troubleshooting some SSO issues lately but I took a look at this for our reporting team. You should check out the ‘Estimated Wait Time’ (EWT) available in the Routing API. It is designed for IVR announcements, but you can pull it for your dashboard too.
However, EWT is just an estimate based on recent history. For a true proactive alert, you might want to look at the ‘Workload’ metrics in WFM to see if you have a staffing gap coming up in the next thirty minutes.
I have seen these predictive alerts fail during massive spikes. If your logic is too sensitive, your supervisors will just ignore the alerts. You should implement a ‘Smoothing’ algorithm to avoid alerts for temporary micro-spikes.
And definitely use the Notification API for the real-time interaction counts; polling the Analytics API for this will eventually cause rate limiting issues when your dashboard scales to multiple supervisors.