General: Contextual and advanced responses

Inside an agent's settings, under the "General" section, you'll find several features that can expand your Agent's versatility.

Contextual Responses

This section lets you enable features that allow the Agent to access contextual information during conversations, improving the relevance and personalization of its responses. These settings are essential for content that depends on the user's location or the current date.

Location

When enabled, the Agent can respond based on the user's geographic location. If location is required, the same question can receive different answers depending on where the user is.

Date and time

When enabled, the Agent can provide responses based on the current date and time. If date is required, the same question can receive different answers depending on when it's asked.

Accessible to authenticated users only

When enabled, the Agent only responds to users who are logged in, ensuring controlled access to its features.

Advanced Settings

The advanced settings let you configure specific Agent behaviors related to content and interaction management. To access these options, click the expandable "Advanced Settings" section.

Ask for feedback on content-based responses

When enabled, the Agent automatically adds a feedback request at the end of content-based responses (e.g., "Was this answer helpful?"). The collected feedback is used to improve the match between questions and content.

Disable AI content collection

Enabling this option prevents the Agent from accumulating drafts and unanswered questions.

If your Agent mostly uses generative AI for dynamic responses rather than content-based ones, we recommend disabling collection.

Suggest questions when the user is inactive

If the user doesn't interact for a period of time, the Agent suggests questions drawn from its content, respecting the configured timeouts. You can exclude specific questions by marking them as "Not random" in the content settings.

Allow users to report an incorrect response (content only)

Lets users flag when a content-based response doesn't match their request. The feedback refines the match between question and content: in similar future requests, that content won't be triggered automatically. The correction is local to the Agent and doesn't affect generative responses.

Confidence Threshold

Defines the level of certainty the Agent requires before providing a response. You can choose from five levels:

• Very low;

• Low;

• Normal;

• High;

• Very high.

With a higher threshold, the Agent only responds when it's very confident it has the right answer. This results in more accurate responses, but increases the cases where the Agent says it doesn't know.

Changing the confidence threshold can significantly affect response quality. Only modify it if you know what you're doing.

DCM Integration

This section lets you connect the Agent to an external DCM (Digital Content Management) system. Fill in:

• DCM User: the user identifier for the integration;

• Secret: the secret key for authentication;

• App Context: the specific application context.

Once you've made your changes, click the purple "Save changes" button at the bottom right to apply the new settings.