Improve topic generation by excluding unnecessary phrases
Important
This content is archived and is not being updated. For the latest documentation, see Microsoft Dynamics 365 product documentation. For the latest release plans, see Dynamics 365 and Microsoft Power Platform release plans.
| Enabled for | Public preview | General availability |
|---|---|---|
| Admins, makers, marketers, or analysts, automatically | - | 
Oct 1, 2020 |
Business value
The analysis of support issues helps businesses optimize support delivery and improve customer experience but can be time-consuming. This feature improves the quality of the AI-driven clustering by identifying and excluding common phrases that can skew the clustering.
Feature details
Dynamics 365 Customer Service Insights automatically parses support case data and uses natural language understanding to group semantically similar cases into autogenerated topic clusters. The knowledge of the top-trending topics allows managers to optimize their support and provide feedback to the core product teams. However, common phrases, such as the product name, can dilute the strength of the topic-clustering algorithm, making the support topic clusters less useful. To help alleviate the problem, users can specify delimiters that are used as tags (among others) within the case title that can be ignored. However, there's no way to ignore non-delimited text.
With this release, additional keywords and phrases, such as company name and irrelevant words, that are specified by the users will be ignored when the text is parsed. The exclusion of the irrelevant words will result in a more accurate and meaningful topic generation.