DocumentationCustomer Journey AnalyticsCustomer Journey Analytics Guide

Create or edit configurations

Last update: October 9, 2026

CREATED FOR:

  • Admin
  • User

Conversation Insights enables you to analyze conversations from the agent experiences you offer to your customers. Those agent experiences can be based on large language models (LLM) or based on human conversations. For example, a chatbot interacting with a customer or call center transcripts.
Through Conversation Insights you are able to understand the impact of agents on actual user outcomes.

Through the Conversation Insights configuration interface you can quickly create or edit a configuration and the associated artifacts (connection, data views, and more).

When you create or edit a Conversation Insights configuration, you specify the sandbox and the event datasets that contain prompts, responses and feedback data. You also select the Customer Journey Analytics connection to which you want to add these datasets. And the data view to which you want to add the Conversation Insights metrics and dimensions.

Only system administrators can create or edit Conversation Insights configurations.

You create or edit configurations from the Conversation Insights Configurations interface.

Restore missing blended dataset

If you edit a configuration and the blended dataset that has been generated for the configuration no longer exists, select Restore to regenerate the blended dataset.

Configuration steps

For each configuration:

  1. In the Details section, specify the following information:

    Conversation Insights Details

    table 0-row-2 1-row-2 2-row-2
    Field Description
    Name Specify a name for the configuration.
    Sandbox Select the Experience Platform sandbox that contains the prompts, responses and feedback event datasets that you want to add to your connection.

  2. In the Datasets section, specify the following information:

    Conversation Insights Datasets

    table 0-row-2 1-row-2 2-row-2 3-row-2
    Field Description
    Prompts event dataset Select the dataset that contains the prompts event data.
    Responses event dataset Select the dataset that contains the responses event data.
    Feedback event dataset Select the dataset that contains the feedback event data.

  3. In the Connection section, if no connection is already configured, use Select a connection to select a connection.

    Conversation Insights Connection

    If a connection is already configured, select Edit Edit to select another connection.

    Conversation Insights Edit Connection

    In the Select a connection dialog:

    Conversation Insights Select Connection

    1. Select the checkbox next to the connection to which you want to add the prompts, responses and feedback event datasets.
    2. Select Use connection.
    • To search in the list of connections to select from, use the Search field.
    • To configure which columns to display in the table, select ColumnSetting . In the Customize table dialog, select the columns to show. Then select Apply.

  4. In the Data views section, if no data views are already configured, select Select data views to select data views.

    If data views are already configured, select Edit Edit data view selection to reconfigure the selection of data views.

    In the Select multiple data views dialog:

    Conversation Insights Select data views

    1. Select one or more data views you want to use for the Conversation Insights configuration.

    2. Select Use data views to use the data views. Select Cancel to cancel.

    • To search in the list of data views to select from, use the Search field.
    • To configure which columns to display in the table, select ColumnSetting . In the Customize table dialog, select the columns to show. Then select Apply.

  5. To finish the configuration:

    • Select Discard for a new configuration that is not created.

    • Select Save for later for a new configuration you want to save but you do not want to create the artifact for (updates to data views for example). You can revisit the configuration later and finish the actual creation of the configuration.

    • Select Create to create the new configuration.

    • Select Save to save the modified configuration.

    • Select Restore to restore the configuration to regenerate a new blended dataset for the configuration.

    • Select Exit to ignore any change to the configuration.

Data view verification

The data views you have configured in Configuration steps, have Conversation Insights as value for Integrations in Data views.

For each of the configured data views:

  • Containers: The Containers tab contains a new Container name: conversation with Display name: Container as an additional System Container type.

  • Components: You see additional schema field folders. For example: agentExperience and conversation. Additionally the following components are automatically added:

    table 0-row-3 1-row-3 2-row-3 3-row-3 4-row-3
    Metrics Schema data type Schema path
    Customer Feedbacks String eventType
    Positive Sentiments String Derived Fields
    Recommendations String eventType
    Turns String eventType
    table 0-row-3 1-row-3 2-row-3 3-row-3 4-row-3 5-row-3 6-row-3 7-row-3 8-row-3 9-row-3 10-row-3 11-row-3 12-row-3 13-row-3 14-row-3 15-row-3 16-row-3 17-row-3 18-row-3 19-row-3 20-row-3 21-row-3 22-row-3 23-row-3 24-row-3 25-row-3
    Dimensions Schema data type Schema path
    Agent ID String agenticExperience.agents.agentID
    Agent Name String agenticExperience.agents.name
    Concierge Name String agenticExperience.name
    Concierge Version String agenticExperience.version
    Conversation ID String conversation.conversationID
    Conversation Name String conversation.conversationName
    Conversation Signal Name String conversation.signals.name
    Conversation Summary Boolean Value Boolean conversation.signals.values.booleanValue
    Conversation Summary Confidence Double conversation.signals.values.confidence
    Conversation Summary Metadata Key String conversation.signals.values.metadata.key
    Conversation Summary Number Value Double conversation.signals.values.numberValue
    Conversation Summary Qualifiers String conversation.signals.values.qualifiers
    Conversation Tone Signals String conversation.signals.attributes.tones.values
    Environment String agenticExperience.environment
    Feedback Classification String Derived Fields
    Feedback Rating Classification String conversation.feedback.rating.classification
    Feedback Section Purpose String conversation.feedback.raw.purpose
    Feedback Source String conversation.feedback.source
    Phrase String conversation.signals.attributes.subjects.values.phrase
    Response Raw Text String conversation.response.raw.text
    Response Source String conversation.response.source
    Sentiment Classification String Derived Fields
    Skill Name String agenticExperience.agents.skills.name
    Skill Version String agenticExperience.agents.skills.version
    Value String agenticExperience.agents.skills.parameters.value
recommendation-more-help
analytics-platform-help-main