Recent searches
No recent searches
Generative AI - Explore reporting
Posted Apr 01, 2024
Currently I see a limitation with using the generate a reply for articles in messaging. When an article is offered to a customer and presents the “Was this helpful?” question, we currently do not have a way to report on this. We have done thorough testing in our sandbox and do not see any of the responses being recorded in any reports in explore. We would like the ability to report on the accuracy of the articles being suggested from our help center and currently are limited on this report.
23
6
6 comments
Tina Yates
I can't emphasize enough the importance of having this type of reporting and sadly it is not there if you use generative ai.
2
Shawna James
Thank you for taking the time to provide us with your feedback and Tina for sharing your feedback here as well. This has been logged for our PM team to review. For others who may be interested in this feature request, please add your support by upvoting this post and/or adding your use case to the comments below. Thank you again!
1
Andrew Barnard
This is essential from our perspective.
1
Luke Bradshaw
Yes this is critical for our usage too.
The part that I'm most confused about, is there seems to be some data in the ‘Insights’ tab of the relevant conversation bot, as explained here that pertains to whether Bot responses included a generated reply.
Clicking “view all bots' performance”, goes to the prebuilt Answer Bot dashboard which contains nothing about generated replies.
At minimum, I wish I could change the date range (instead of 7-day performance), and specify ‘automated resolutions’ to only include responses where a reply was generated (instead of an Article Recommendation).
It's honestly bizarre that Zendesk don't have any reporting for such a basic metric, even more-so because it does appear to be already implemented in the ‘Insights’ tab, albeit with limited functionality.
1
Ele Hoare
Hi Shawna James - are you able to confirm then if the resolutions that come from a generated reply are included in the Explore data set for the flowbuilder? Or are they on top of that?
I assumed we must be missing something in the reports as this seems like an absolutely essential piece of reporting.
If there is any update on whether this will be added to the roadmap, please let me know.
1
皆川 真聖
Additionally, I would like to get the number of cases where the AI was unable to generate an answer.
This is necessary to calculate the AI's response rate.
2