Ben_Wyatt | 2021-04-21 17:57:09 UTC | #1
Within the Conversation Details Job/Query (/api/v2/analytics/conversations/details/{jobs/query}) as well as the Evaluation Aggregate query (/api/v2/analytics/evaluations/aggregates/query), the oTotalScore metric for evaluations is provided as an interger. At the same time, the score is available via the frontend (specifically within /directory/#/analytics/interactions) with 2 decimal places. Is there an API call that will return the score on evaluations with greater precision consistent with the frontend? Failing an alternative solution, this lack of precision is going to be a huge problem for us creating evaluations on the cloud as we need to get that precise score for performance management.
Thanks in advance!
Jerome.Saint-Marc | 2021-04-21 19:51:39 UTC | #2
Hello,
I am not expert in Quality Evaluations (I mean not something I have practiced much) but just to clarify.
When you say "the score is available via the frontend (specifically within /directory/#/analytics/interactions)", I assume you mean selecting a specific interaction in the Performance - Interactions view, then clicking on the Quality Summary tab (of this interaction), and then clicking on one of the evaluations of that interaction.
If that's correct, the API endpoint which is leveraged there, and which brings back the total score with 2 decimals is: GET /api/v2/quality/conversations/{conversationId}/evaluations/{evaluationId} [Get an evaluation].
There is a second API endpoint which brings back totalScore with 2 decimals: GET /api/v2/quality/evaluations/query [Queries Evaluations and returns a paged list]. You can search by conversationId, queueId, agentUserId, evaluatorUserId to return multiple evaluations. You will need to set the expandAnswerTotalScores parameter to true to get the scores back.
Regards,
Ben_Wyatt | 2021-04-21 19:48:22 UTC | #3
Jerome, your clarification is correct, and the information around the expandAnswerTotalScores on GET /api/v2/quality/evaluations/query is helpful. I can't say I'm enthused by the idea of querying for a list of IDs to then query for details, all just to get precise scores, but I guess it will work. I'll probably use GET /api/v2/quality/evaluations/query grouped by the evaluatorUserIds since that should be the lowest cardinality of results and therefore the fewest extra calls.
system | 2021-05-21 19:48:24 UTC | #4
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: 10701