Learn how this feature uses Artificial Intelligence to evaluate interactions, where to access it in Zenvia Customer Cloud, and how to interpret the results.

1. What is sentiment analysis?

Sentiment analysis is a feature that uses Artificial Intelligence to evaluate the content of interactions with your contacts. It analyzes the context of words and texts to classify sentiment as positive, negative, or neutral.

2. Where is this analysis presented?

The feature is displayed differently depending on the module:

• Home: Organization overview
  Presented as Satisfaction Index, it provides a general assessment of interactions with contacts, indicating whether they were positive, negative, or neutral.

• Contact profile
  Presented as Sentiment, it evaluates how the contact feels about their interactions with your team. Each interaction is analyzed and classified as positive, negative, or neutral.

• Contact base
  It can be used as a filter called Contact Sentiment to refine the search for contacts based on the predominant sentiment in their interactions. It is also available as a customizable column, allowing the analysis to be viewed directly in the contact base.
• Segmentation rules
  Sentiment analysis can be used as a condition called Customer Sentiment to create segmented groups based on the predominant sentiment in the contacts' interactions.

3. What is the difference between the Satisfaction Index and Sentiment?

• Satisfaction Index (Home Module): Represents a general view of sentiment across all interactions in the organization.
• Sentiment (Contact profile): Focuses on the analysis of a specific contact's interactions.

4. Do the Satisfaction Index or Sentiment consider formal surveys or service duration?

No. Both are calculated exclusively based on the analysis of the content of interactions, without considering formal surveys or service time.

5. How is the score calculated?

The analysis is performed using a Natural Language Processing (NLP) model, which evaluates the context of conversations to identify the predominant sentiment. The model considers not only the words used but also how they connect and the overall tone of the interaction. Based on this, it classifies the sentiment as positive, negative, or neutral.