Ben_Wyatt | 2021-05-20 14:51:53 UTC | #1
From a single agent participant segment on an inbound conversation:
"segments": [ { "conference": false, "queueId": "20538dc3-21c1-400b-b3ff-ee17888e9440", "segmentEnd": "2021-05-05T19:41:30.790Z", "segmentStart": "2021-05-05T19:41:27.356Z", "segmentType": "alert" }, { "conference": false, "disconnectType": "transfer", "queueId": "20538dc3-21c1-400b-b3ff-ee17888e9440", "segmentEnd": "2021-05-05T19:43:36.891Z", "segmentStart": "2021-05-05T19:41:30.790Z", "segmentType": "interact" }, { "conference": false, "disconnectType": "transfer", "queueId": "20538dc3-21c1-400b-b3ff-ee17888e9440", "segmentEnd": "2021-05-05T19:43:40.445Z", "segmentStart": "2021-05-05T19:43:36.445Z", "segmentType": "wrapup", "wrapUpCode": "6e385b31-2603-402a-8054-406dd1bf06a5" } ]
The interact and wrapup segments overlap by almost 500ms. Should this even be possible? Is there likely a configuration issue, or is there a possible defect to explain this behavior?
Thanks!
Ben_Wyatt | 2021-05-20 17:02:05 UTC | #2
Similarly, I have an example where there is a 463ms gap between the end of the interact segment and the start of the wrapup segment.
"segments": [ { "conference": false, "queueId": "773b581e-f722-41aa-8fa8-aef4fef6c2e4", "segmentEnd": "2021-05-05T19:42:37.817Z", "segmentStart": "2021-05-05T19:42:33.360Z", "segmentType": "alert" }, { "conference": false, "disconnectType": "client", "queueId": "773b581e-f722-41aa-8fa8-aef4fef6c2e4", "segmentEnd": "2021-05-05T19:45:21.584Z", "segmentStart": "2021-05-05T19:42:37.817Z", "segmentType": "interact" }, { "conference": false, "disconnectType": "client", "queueId": "773b581e-f722-41aa-8fa8-aef4fef6c2e4", "segmentEnd": "2021-05-05T19:45:27.052Z", "segmentStart": "2021-05-05T19:45:22.052Z", "segmentType": "wrapup", "wrapUpCode": "2f3b1df9-e4f3-4482-bcf6-4f381fe34862" } ]
Is it fair to say that the segmentStart of the wrapup segment is imprecise relative to other segmentTypes? If so, are there other segmentTypes where I should expect this lack of relative precision? As a counterexample, the end of any applicable alert segmentType appears to match perfectly with the start of the interact segmentType.
tim.smith | 2021-05-24 19:13:46 UTC | #3
It should be fine that the segments overlap by a small amount; neither a conversation's nor a participant's flow are strictly linear. If you need further investigation into these timestamps, please open a case with Genesys Cloud Care; customer data cannot be investigated via the forum.
Ben_Wyatt | 2021-05-24 18:46:08 UTC | #4
I don't follow that response. Are you saying it ought to be possible for an agent to be interacting with the customer and on wrap-up concurrently? Or vice-versa, that there can be a time gap after an agent is done interacting with a customer but before they begin their wrap-up?
tim.smith | 2021-05-24 20:11:11 UTC | #5
Is it possible for a human to simultaneously be on the call with the customer and have ended the call and be in wrap up? No. But it is possible for non-linear asynchronous processes to overlap slightly while doing those things. The latter is what the timestamps are measuring. This is definitely a pedantic distinction, but it's important to understand because it explains how the data can seemingly show something contrary to the physical events that took place.
system | 2021-06-23 19:13:27 UTC | #6
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.
This post was migrated from the old Developer Forum.
ref: 10981