Skip to main content

[Extension] How to Analyze Customer Sentiment in Your Bot Conversations on Blip with the Waizer Extension

  • Updated

  1. Introduction

  2. How to Analyze Customer Sentiment

  3. Usage Example

  4. Best Practices

  5. Support

 

Brief description: Analyze customer sentiment in your bot conversations on Blip with Waizer by comparing the beginning and end of each conversation and cross-referencing with the heat map and the most discussed topics.

 

Introduction

Customer sentiment is one of the most sought-after indicators, and for a straightforward reason: it captures how the person felt after interacting with the company, something that resolution rate alone does not show. A service can be technically resolved yet still leave the customer frustrated due to tone or delayed responses.

Waizer applies sentiment analysis to each bot conversation and compares how the contact felt at the start of the interaction with how they felt at the end. This comparison appears in the Sentiment Bubbles chart, which positions each conversation between initial and final sentiment. You can also cross-reference this reading with the Heat Map by day and time and with the most discussed Topics, understanding not only that the customer was dissatisfied but also why. For those who have tried to do this on their own, the difference is clear: there is no need to build an NLP pipeline, train models, or integrate a sentiment API. Waizer delivers the analysis ready on the dashboard.

Before you start

Before analyzing conversation sentiment, make sure that:

  • The Waizer extension is installed and active on the bot. Activate Waizer in the Blip Store.

  • The bot already has conversations recorded for the period you want to analyze.

  • You have access to the Waizer Dashboard.

  • The analyzed period contains enough conversations to generate a reliable reading.

 

How to Analyze Customer Sentiment

  1. Open Waizer on the desired bot and click on Dashboard.

Tela inicial do Waizer com o Dashboard de análise aberto

  1. Locate the Session Sentiment card. Waizer classifies sentiment as Positive, Neutral, or Negative and measures both the start and end of each conversation.

Card Sentimento das sessões do Waizer com a legenda de positivo, neutro e negativo

  1. In the Sentiment Bubbles tab, read each bubble by its position: the horizontal axis shows the Initial Sentiment and the vertical axis shows the Final Sentiment. The bubble size reflects the volume of messages. Hover over the information icon to see the quadrants, which separate conversations that ended well from those that regained trust, unresolved complaints, and those that caused frustration.

  2. Open the Heat Map tab to see session sentiment distributed by day of the week and time. This way, you can discover if negative conversations concentrate during a specific period.

Aba Mapa de calor do Waizer com o sentimento das sessões distribuído por dia e horário

  1. Click on a bubble or a cell in the map to open the conversations from that group and read what caused the sentiment.

  2. Cross-reference this reading with the Topics card to identify subjects linked to negative sentiment. In the Sort by selector, switch between Most Discussed, Error Rate, Resolution Rate, and Overflow Rate to see the topics from each perspective.

 

Usage Example

Suppose many contacts start the conversation with a neutral sentiment and end with a negative sentiment. In the Sentiment Bubbles, this pattern appears in bubbles that drop from start to end. This usually indicates that the bot did not resolve the request, the responses were unclear, or the contact was directed to incorrect paths.

Bolhas de sentimento mostrando conversas que começam neutras e terminam negativas

Based on this analysis, you review the messages, adjust the flow options, and monitor if the final sentiment improves.

 

Best Practices

  • Analyze sentiment together with the Error Rate and Overflow Rate.

  • Observe which themes in the Topics card concentrate the most negative sentiment.

  • Review messages that may cause confusion or frustration.

  • Compare sentiment before and after each flow adjustment.

  • Use the most recurring topics to improve bot content and responses.

 

Support

If you have any questions or issues with the extension, contact us at support@wiv.com.br.


 

Need more help? Explore our content at Blip Academy or Blip Community, watch tutorials on our YouTube channel or get your questions answered in our support channel 😃

Related articles