Conversation Analytics administration
Conversation Analytics runs automatically: every captured conversation is transcribed and then processed by the enabled AI tasks — summary, sentiment, topics, custom insights, and (if your subscription includes it) Auto QA. Administrators do not run the analysis; they decide what the AI looks for and review usage. This page maps out what your organization can configure and what your service provider manages.
For the concepts behind these features, see Conversation Analytics in the Overview & Key Concepts guide.
Note
The features described in this chapter depend on your subscription. If your subscription includes Conversation Analytics but a page described here is missing from your Administration menu, contact your service provider.
Where the settings live
The Administration › Conversation Analytics menu contains the areas your organization configures:
| Menu item | What it is | Documented in |
|---|---|---|
| Transcription | Transcription usage and limits (read-only) | Usage and limits |
| AI Tasks | Enable, disable, and customize the AI analyses, plus AI usage, and the Playground | AI tasks, AI Playground |
| Topic Analysis | Define the topics detected in conversations | Topics |
| Data Redaction | Rules for masking sensitive data | Data redaction |
Two related areas live elsewhere:
- Custom fields — the destination for many insights — are managed under Administration › Customization › Custom Fields; see Custom fields.
- Auto QA is configured as an AI task linked to a scorecard; see Auto QA setup.
What your organization configures
- AI tasks — enable or disable each analysis, override the prompts and filters of the built-in tasks, and create your own. See AI tasks.
- Custom insights — define new data points the AI extracts into custom fields, such as CX metrics or sales signals. See Custom insights.
- Topics — the list of subjects the AI detects across conversations. See Topics.
- Data redaction rules — what sensitive content is masked. See Data redaction.
What your service provider manages
Managed by your service provider
The transcription engines and languages, the generative AI engines that execute AI tasks, and your organization's usage limits are configured by your service provider. Your administrators see usage against those limits (Usage and limits) but cannot change them.
Licensing
AI analysis runs on the conversations of users who hold the Conversation Analytics per-user license (shown as Speech analytics license on the user profile). Conversations of unlicensed users are captured but not analyzed. See Users and Subscriptions & Licensing.
Related pages
- AI insights in the User Guide — where users see the results.
- Usage and limits — monitoring transcription minutes and AI token usage.