Agent empathy analysis gauges a customer service representative’s level of empathy and emotional intelligence during interactions with customers. This process involves evaluating the language that agents use in their conversations with customers, provides valuable insights about the customer and agent dynamic, and enables organizations to evaluate their Customer Care quality.

Agent empathy classifies each phrase from the agent based on the exhibited empathy by analyzing the language used and categorizes them as empathetic, neutral, or unhelpful.

Agent empathy assigns an overall empathy score to each agent in the interaction, and measures the degree of empathy conveyed. This score reflects the agent’s ability to understand and relate to the customer’s emotions and concerns, and ranges from -100 to +100. 

Agent empathy analysis cannot attribute specific behavior to any individual agent. For that reason, the score calculation does not include portions of interactions that involve two or more agents that simultaneously interact with the customer.

How agent empathy can be utilized in your organization

Training and coaching: Organizations can use empathy analysis results to tailor agent training and coaching programs. This helps agents develop their empathy skills and deliver more compassionate customer service.

Quality assurance: Empathy analysis can be integrated into quality assurance processes to ensure that customer interactions meet predefined empathy standards.

Customer experience enhancement: By identifying interactions with low empathy scores, organizations can pinpoint cases where customers may have had negative experiences. This process allows them to take corrective actions and improve overall customer satisfaction.

Agent recognition: Organizations can recognize and reward high-performing agents with exceptional empathy skills and motivate others to emulate their approach.

Calculating agent empathy score

Empathy score = 100 * [(Number of empathetic phrases – Number of unhelpful phrases) / (Total number of empathetic and unhelpful phrases)]

For example, if an interaction has 2 empathetic phrases from an agent and 3 unhelpful phrases the score is calculated as follows:
100 * [(2-3)/(2+3)] -> 100 * [-1/5] -> 100 * (-0.2) -> -20

  • If an interaction only contains empathetic phrases, the overall empathy score will be 100 for agent involved in the interaction. If the interaction only contains unhelpful phrases, the overall empathy score will be –100.
  • Agent empathy scores can only be accessed using the Speech and Text Analytics API.

Enable agent empathy analysis

  1. From the Admin menu, select Quality.
  2. Select Speech and Text Analytics and verify that the Agent Empathy Analysis is enabled.

Example phrases

Empathetic examples that respond to a negative event

  • Sorry that we couldn’t help you. We will serve you better the next time.
  • I appreciate your patience. I am sorry for the inconvenience caused because the product is not working.
  • I’m sorry to hear about your problem. I can completely understand why you are so upset.
  • Your frustration is completely valid. I appreciate your patience.
  • I can completely understand why you’re so upset.
  • I’m sorry to hear about your problem. I appreciate your patience.
  • You’re a valuable customer to us and your frustration is completely valid. It should not have happened that way.

Empathetic examples that respond to a positive event

  • I’m glad that you got on the right track with it and set it up
  • I’m glad to hear that
  • Oh, I am really happy that I have helped you today

Unhelpful agent behavior

Unhelpful agent behavior encompasses interactions involving instances where the agent uses inappropriate language, displays impatience, fails to listen carefully to the customer, or lacks the knowledge to handle the issue.

  • I don’t even know what venue you’re speaking of
  • I don’t know what to tell you
  • You’ve been nothing but rude ever since ever since i’ve answered the phone
  • Okay but you are not letting me explain
  • Can’t understand your accent
  • There is nothing we can do

Note: Phrases detected as empathetic or unhelpful can be accessed using the Speech and Text Analytics Transcript API.