Skip to main content

Track conversation topics


Conversation Topics tracks the topics you define based on keywords and phrases. Tracking topics can provide insight into trends in your chatters' behavior. Understanding these trends helps you build better bot content for an enhanced customer experience, and can even inform the development of your products and services.

Use Conversation Topics

The Topics view is found under Conversations on the Ada dashboard. From there, you can create, track, and analyze the topics your chatters discuss.


Create a topic

To create a new topic:

  1. On the Ada dashboard, go to Conversations > Topics.

  2. In the Topics view, click New Topic.

  3. In the Define a new Topic dialog, add a name and description for your topic, then click Next to proceed the topic setup page.

  4. On the topic setup page, in the Keywords to track field, add the keywords and phrases you want to track.

    • Place a comma after each keyword or phrase, even if you include only one keyword. The text will become a pill icon.

      Example: If you were tracking board game titles, you would enter Monopoly, Stratego, Risk,.

    • Use quotation marks to track an exact match only.

      Example: If you were tracking board games with phrases for titles, you would enter "Snakes and Ladders", "Mouse Trap",.

  5. [Optional] To exclude keywords, select the Exclude Keywords checkbox. In the field that appears, enter the keywords to exclude from the topic tracking.

  6. [Optional] On the Additional Conditions panel, create filters to refine which conversations the topic applies. You can filter by Answer, language, or variable. To add more filters, click Add new filter.

  7. Click Save. The dialog closes and your new topic is added to the topics list.

Multilingual support

Conversation Topics is a multilingual feature. It supports every language enabled for your bot. Conversation Topics does not perform translations, however. If you want to track a topic across multiple languages, you need to manually add the translation of that topic in every language you want to include in your search.

Example: Let's say you have a bot that supports French and Spanish, as well as English, and you create a Conversation Topic called Pricing. In addition to the keyword price, you would also include prix and precio to track the keyword in French and Spanish, as well.

Tracking keywords in specific languages means you can filter a topic for those languages. This helps you identify any trends that might be language specific.


Though Conversation Topics is multilingual, conversation summaries, found on the lower half of the conversation topic card, are generated solely in English. This is especially important to note for topics tracking only non-English keywords. For those topics, you won't see any relevant conversation summaries.

Individual topic insights

The Topics view contains a column of topic cards for each of the topics you are tracking. Each topic card displays the number of conversations in which it was triggered, the number of those conversations that resulted in a handoff, and the last person to update the topic settings.

The progress bar for each topic shows the volume of that topic relative to all the other topics you are tracking.

Click any topic card in the list to view a more detailed breakdown of the data collected for that topic. This topic insight view is made up of three panels. Use them to:

  • Explore volume, handoff, and customer satisfaction (CSAT) trends to help understand the topic's performance

  • Review conversation summaries and CSAT comments to gain insight into your customer's sentiment

  • Uncover opportunities for content and training improvements based on unanswered questions related to this topic

  • Filter and download insights.


Conversation Insight

The bottommost panel on the topic page contains the Conversation Insights. Use the left-hand menu to select which insight you wish to see. The following descriptions will help you understand what each item measures.

  • Conversation summaries uses Ada's language engine to distill a conversation down to its key points. This feature helps you understand what it is your chatter is looking to achieve. The listed conversation summaries are identified by chatter ID. Click on a summary to open the entire conversation associated with it.

  • Unanswered questions lists the unanswered questions from conversations in which the topic was triggered. This helps you uncover content gaps and improvement opportunities for your bot. To view training suggestions for your topic, click View all Questions. This opens the Unanswered view under Improve on the Ada dashboard.

  • CSAT Comments displays your chatter's ratings and comments. This provides a targeted way to review CSAT comments related to a topic, and identify opportunities to improve your product and services based on user feedback. This data is only available for bots using Ada's CSAT feature.

Have any questions? Contact your Ada team—or email us at