Document cognition insights
This section provides insights generated by the documents uploaded in the Studio section while training the bot. Document cognition (Doc Cog) offers analytics on search result performance, search categories, and more.
Detailed documentation on document cognition is available here.
In this article, you will learn about:
- Number of queries that went to document search: Number of queries answered and unanswered by document search for your bot over a period of time.
- User feedback: Cataloging the feedback received for search results for your bot into positive, negative and no feedback.
- User searches: Searches performed by the users that are grouped based on the similar queries under a clusters.
- Document performance statistics: Track the relevance of documents by inspecting the performance metrics - suggestion count of the document, click through rate and percentage of positive feedback.
Access doc cog insights
To view document cognition insights:
Open Insights > Doc cog.
Use the available filters to view insights for a specific time period by selecting the desired range or custom dates.
Number of queries that went to document search
Select the period and view the number of queries that were answered/unanswered during that range when the search was performed.
User feedback
After a query is answered by the bot using documents, users can give feedback for active learning. You can view the number of positive/negative responses provided by the user for the bot responses.
Group/categories that the users are searching for
Understand the categories or topics users are searching for, along with the corresponding bot responses and feedback received.
Document performance statistics
Analyze the relevance and performance of your documents through various metrics. Track suggestion count, click-through rate, and percentage of positive feedback. The performance statistics are presented in two tabs: "High Rated Documents" and "Low Rated Documents."
