Conversations

Here, you can oversee the quality of interactions and gather valuable data to optimize the assistant's responses and the overall user experience.

1. Filters

Located at the top of the interface, filters allow for searching and classifying conversations based on various criteria:

Main fields

• Text search: Search for specific messages within conversations.

• Language: Filter conversations by language.

• Channels: Classify by the originating channel (e.g., web, phone, etc.).

• Date range: Define a time period for analyzing conversations.

• Additional buttons:

  • Open a sidebar with advanced filters.

  • Reset all filters.

Advanced filters sidebar

• Help requests: Conversations where the user requested human assistance.

• Interventions: Conversations attended by an agent, with options to filter by a specific agent or department.

• Ratings: Classification based on user ratings:

  • Extremely positive (90%-100%).

  • Mostly positive (60%-90%).

  • Mixed (40%-60%).

  • Mostly negative (10%-40%).

  • Extremely negative (0%-10%).

• Intent confidence: Average score (0 to 10) assigned by the AI to all messages.

• Tags: Filter by specific tags.

• Origin URL: Filter conversations based on the message's originating URL.

2. Conversation list

Located on the left of the central area, this displays a summary of recent conversations:

• Origin channel (icon).

• User name or phone number.

• Date of the last message.

• Tags assigned to the conversation.

3. Conversation details

In the center of the interface, this displays the messages of the selected conversation with the following features:

• Messages differentiated by position and color based on the sender (user, bot, or agent).

• Origin channel icon.

• Additional information such as user name, tags, and date of the last message.

• Buttons to navigate to the start or end of the conversation.

• Button to open a side panel with detailed information.

4. Conversation information

In the right side panel, it displays detailed information and allows for additional actions:

Top section

• Conversation Identifier.

• Buttons:

  • Create a support ticket.

  • Open a modal with options like navigating to training sections, sharing the conversation (screenshot or URL).

Expandable sections

• General: Start and last message dates, and the number of messages.

• User information: Data such as phone number, email, and country of origin.

• Ratings:

  • User ratings (classification).

  • Intent confidence average.

  • Number of fallbacks (default intent triggered).

• Help requests: Information on help requests directed to specific departments.

• Interventions: Agents involved, average handling time, and number of messages sent.

• Surveys: Responses from surveys conducted during the conversation.

• Tags: Current tags with the option to add more.

• Selected message information:

  • Sender (user, bot, or agent).

  • Sending date.

  • Message origin (URL).

  • Detected intent and confidence score.

  • User rating (positive/negative).

  • Executed actions (e.g., help request sent).

5. Pagination

Located at the bottom, it allows:

• Displaying a defined number of items per page.

• Navigating between pages: first, last, or a specific page.