Trackers is a powerful Conversation Intelligence feature you can use to automatically recognize phrases and their meaning in conversations.

An individual tracker is a group of phrases identifying a characteristic or an event you want to track in conversations.

For example, if you are a Product Manager interested in tracking pain points from all customer conversations in your organization, you can create a tracker named Pain Points. You can add phrases such as "we struggle with", "we find it difficult", "blocker", and "need help with". These keywords and phrases are commonly used by customers to describe their pain points.

When you process a conversation using the Async API or Streaming API, finds phrases that have similar meaning to the ones you defined in the Tracker. For Pain Points, phrases with similar meanings might include "we are having trouble", and "there is a problem".

How Trackers Work

The Trackers algorithm uses zero-shot learning to efficiently find matches for your Trackers in conversations without any training data. It first understands the contextual meaning or intent of the vocabulary that you provide for a tracker, then creates a robust tracking context. When you process a conversation, the Trackers algorithm tries to match the intent of messages in the conversation with the intent of your tracker vocabulary. If the intent matches, the detected trackers are added to the response.

For example, “I don’t have any money” is contextually similar to “I ran out of budget” as both represent similar inherent meaning. The Trackers algorithm generalizes the scope of meaning and finds phrases related to running out of money.

Trackers Use Cases

Trackers provide the flexibility to choose your own vocabulary so that you can tailor them to your use cases. You can use the occurrence of Trackers in conversations to trigger automatic or manual workflows, identify coaching opportunities for employees, and identify customer interests and issues.

The following examples illustrate some specific use cases:

Sales: Sandra is the training head of the Sales team in her company. She uses insights from the Intro Tracker to look at how great sales agents start conversations. She uses the Pricing Tracker to understand how great agents share the pricing details. She then uses these insights to train other agents.

The Intro Tracker Sandra uses tracks contextually similar phrases such as:

  • "Thank you for taking some time to speak with me"
  • "Have you ever noticed"
  • "Have you ever considered"

Quality Assurance: Tom is the Chief Quality Assurance Manager and wants to see how many agents are using negative phrases on the customer calls. He gets insights from the Negative Phrases Tracker by tracking the following contextually similar phases:

  • "I am not interested"
  • "This makes me uncomfortable"
  • "Not so good"
  • "I am skeptical"
  • "My only problem is"
  • "My only issue is"
  • "This does not fit"

Business Leads: Samuel uses insights from the Leads Tracker for all the dial out calls at his car insurance company. By tracking phrases that show buying intent, inquiries about policy options, quotes, discounts, and so on, he generates more leads for his business. The Leads Tracker detects the following contextually similar phrases:

  • "I need"
  • "Policy options"
  • "Require"
  • "Use case"
  • "Fits our requirements"
  • "Quotes"
  • "Discounts"

Next Steps

  • Getting Started with Trackers helps you quickly add, edit, and work with Trackers on the API Explorer Platform.
  • Managed Trackers Library gives you the option to choose from predefined Trackers that best suit your conversations. This saves the time and effort required to manually create a Tracker.
  • Custom Trackers you can use to create entirely new Trackers from scratch.
  • Recommendations are new Trackers that suggests to you based on conversations you have processed in the last 30 days.

Did this page help you?