Configure AI topic clustering for cases

Customer Service Insights uses AI to give you insights into your customer service data by grouping semantically related cases and generating a topic. New cases that match the generated topic will be automatically added to the topic group. This can help you identify areas for improvement that can have the greatest impact on system performance.

The AI-driven technology empowers informed decision-making about how to improve resolution rates, reduce wait times, and decrease customer service costs. You can use case resolution insights, backlog trends, and historical comparisons to evaluate agent performance and business impact, and address inefficiencies in your system.

Enable topic clustering for cases

Topic clustering is enabled by default when you enable historical analytics. To enable historical analytics, see Configure Customer Service historical analytics.

Summary and Model run summary views

Summary and Model run

The Summary and Model run summary views provide key information about how the topic model is operating.

View Description
Status Whether the feature is enabled or disabled.
Data attributes used Which text field from the Case entity is used for topic generation.
Topics generated The total number of topics generated by the model.
Cases associated to a topic The percentage of cases that were considered for topic generation and classified to a topic.
Last successful run Timestamp of the last time new cases were processed.
Run frequency The cadence in which new cases are processed and tagged with topics.

Data mapping

Data mapping enables you to choose the text field that the agents in your organization will most likely use to describe the reason why a customer reached out to support. By default, the Case Title attribute is used. The other attributes available to you are Serial Number, Description, and Case Title.

Data Mapping

Improve data quality by cleaning support case data

Clean support case data to improve data quality

The AI Insights charts displayed on the Customer Service Insights dashboards are generated by applying language understanding technology to the titles of support cases. However, the results can be misleading if the titles include extraneous information such as product name, case status, or ticket number tags. You can improve the quality of the results displayed in AI Insights charts by specifying Data Cleaning settings to disregard tags in titles when they are grouped into topics, as well as specific phrases that should be ignored. Note that when you choose to apply both options, sections are ignored first, followed by phrases.

Enable topic automation for Power Virtual Agents

AI discovered topics in Customer Service Historical Analytics are often prime candidates as topics for automation in Power Virtual Agents bots. If Power Virtual Agents is available in the region that your Customer Service organization is in, then the feature can be enabled by switching the Disabled toggle to Enabled.

Topic automation for PVA


Topic automation to Power Virtual Agents bot is currently not supported in Government Community Cloud.

Language availability for topics

The topics capability in the Customer Service historical analytics reports comes with a natural language understanding model that can understand the text semantics and intent in the following languages:

  • English
  • French
  • German
  • Italian
  • Japanese
  • Portuguese
  • Simplified Chinese
  • Spanish


While topic discovery is not prevented and still possible in languages that are not listed above, there may be differences in experience for users who leverage topics in unsupported languages.

See also

Introduction to Customer Service analytics and insights
Dashboard overview
Topics Dashboard