[Extension] How to Use the Waizer Extension in Blip to Analyze Your Chatbot Conversations

1. Introduction

2. Features

3. Installation and Configuration

4. Ask Waizer

5. The Dashboard

6. Agent Profile and KPIs

7. Integrations (MCP)

8. Settings and Governance

9. Support

Brief description: Waizer analyzes your chatbot and human service conversations in Blip, generates dashboards on sentiment, errors, and overflow, and answers questions in natural language with Ask Waizer.

Introduction

Waizer is Wiv’s conversational analytics platform, available as a contractable extension within Blip. It reads the bot conversations where installed, evaluates each session using artificial intelligence, and transforms this volume of messages into indicators, sentiment analyses, and practical recommendations to improve both automatic and human service.

For example, instead of opening hundreds of conversations to understand why customers abandon scheduling, you ask Ask Waizer directly: "Why are customers abandoning the scheduling flow?". The tool identifies related data, responds based on real conversations, and shows examples for you to validate the hypothesis. The complete product documentation is available at help.waizer.ai.

Features

• Dashboard with key operation indicators: total contacts, total sessions, error rate, resolution index, overflow rate, and predominant sentiment.

• Performance score from 0 to 100, a consolidated rating of the bot’s performance over the period.

• Sentiment analysis of sessions displayed as bubbles, quadrants, and heat maps by day and time.

• Problems and failures classified by severity and number of occurrences, including impacted contacts.

• Most discussed topics and word cloud with sentiment associated with each term.

• Insights and suggestions generated from conversations, prioritized by impact.

• Ask Waizer, an AI agent that investigates the operation through natural language questions and exports reports in Word, Excel, PDF, and PowerPoint.

• Integrations via MCP (Model Context Protocol) to query external systems during analysis.

• Configurable agent profile and KPIs that feed the analysis context with your business goals.

Supported on the following channels: all channels available on the Blip platform. Waizer is an independent tool and is not limited to operations linked to Blip: besides Blip, it also analyzes data sent via API or imported by CSV file.

Installation and Configuration

Install Waizer on the bot you want to analyze.

1. Open Waizer in the Blip Store.

2. Click Start free trial and install the extension on the desired bot, Router, or service bot. The extension will then appear among the installed extensions of the bot.

3. Open Waizer. The extension connects to the bot’s Blip account and displays the data import screen.

4. Click Start import. The import begins and may take up to 1 hour to complete. You can close the page and return later.

5. When finished, the completion will be indicated on the panel.

6. Click Go to dashboard. The dashboard will already display all available data for the bot.

Note: when creating the project through Blip, the same project is available at app.waizer.ai, where you can also access Waizer.

7. Use the period filter at the top of the dashboard to set the analysis interval.

By default, the filter shows the last 30 days. To compare the charts with each other, keep a specific and consistent period. The comparison is made relative to the equivalent previous period.

Attention: when installing through the Blip Store, the extension starts with a free trial, with a limited preview of the last days. For information about plans, costs, and full release, contact Wiv’s sales team at comercial@wiv.com.br.

Ask Waizer

Ask Waizer investigates the operation through natural language questions. Instead of manually creating reports, describe what you want to analyze and deepen the investigation based on Waizer’s data.

You ask as you would ask an analyst, without knowing database structures or configuring filters. Some examples:

• What were the main reasons for contact this month?

• At which step of the flow do customers most often abandon?

• Which topics most generated overflow to human service?

• Which agents received the most complaints?

• Was there an increase in negative sentiment compared to the previous month?

The tool maintains context between questions. You start with a broad question and follow up with complementary questions on the same topic, analyzing causes, conversation examples, periods, queues, or agents. When there is insufficient data to support an answer, Ask Waizer notifies you, asks for a more specific period, or suggests another cut, reducing the risk of unsupported conclusions.

To export an analysis, request the desired format from Ask Waizer. It generates the file in Word, Excel, PDF, or PowerPoint and makes it available for download, making it easy to share results and bring them to meetings.

The Dashboard

Main Indicators

The cards at the top of the dashboard summarize the operation during the selected period:

• Total contacts: unique contacts that interacted during the period. The same contact is counted once, even if they send multiple messages.

• Total sessions: each contact can generate multiple sessions. A session is a continuous block of interaction, separated by periods of inactivity.

• Error rate: percentage of sessions with some failure or unexpected scenario, such as processing errors, integration failures, or misunderstood messages.

• Resolution index: percentage of sessions resolved according to user needs, such as a completed sale or support provided.

• Overflow rate: percentage of sessions that required human intervention.

• Predominant sentiment: the predominant final sentiment in contacts during the period.

Performance Score

The Performance Score is a consolidated rating from 0 to 100 that summarizes the bot’s performance based on operation, experience, and quality indicators. Use it to track service evolution over time and communicate results between technical and business teams.

You can also export the performance report.

Sentiment Analysis

The Session Sentiment chart compares the user’s initial sentiment with the final sentiment after interaction. The horizontal axis shows the initial sentiment and the vertical axis the final sentiment, from negative to positive. The bubble color indicates sentiment (red for negative, yellow for neutral, green for positive) and the size indicates how many conversations fall into that classification.

Click on the bubbles to see the corresponding conversations.

Clicking a conversation directs you to Blip, to the contact and conversation history.

The session quadrant is an exclusive Wiv methodology that organizes conversations by crossing the initial and final sentiment:

• Q1, successful positive conversations: started well and ended well.

• Q2, conversations that regained trust: started poorly and ended well.

• Q3, critical poorly resolved conversations: started poorly and ended poorly.

• Q4, conversations with frustration: started well and ended poorly.

Many bubbles in Q3 and Q4 signal that the flow needs to be reviewed.

The heat map shows when interactions occur and the sentiment at each moment. The vertical axis shows days of the week and the horizontal axis times of day, with the same sentiment colors. Concentrated green blocks indicate times with good performance; red blocks at specific times may signal overload or flow failures. For example, if on Mondays between 9 a.m. and 12 p.m. red predominates, the volume during that period may be overloading the bot or team.

Click a block to open contacts from that moment.

Problems and Failures

The Problems and Failures field lists the main errors identified in interactions, organized by failure type, severity (high, medium, or low), and number of occurrences. The total number of impacted users appears at the top. Address high-severity problems with many occurrences first, as they directly affect the customer experience.

You can also see the contacts impacted by each problem.

Most Discussed Topics

The Most Discussed Topics field shows the most recurring subjects in conversations. The percentage next to each topic indicates the proportion of conversations in which it appeared.

You can sort the topics.

Click a topic to open the affected contacts and analyze the conversations.

Insights and Suggestions

The Insights and Suggestions field presents recommendations generated from conversations, classified by impact (high, medium, or low). Each suggestion comes with a practical description of how to apply it to the bot.

Click an insight to see the identified problem, recommended solution, expected result, and impacted contacts.

Word Cloud

The word cloud displays the most used terms by customers and the sentiment associated with each. The screen shows the total words analyzed, total occurrences, average sentiments, and distribution among positive, neutral, and negative. The color of each word indicates the predominant sentiment (green positive, yellow neutral, red negative).

Use the sentiment distribution to filter words by positive, neutral, or negative.

Click a word to see the context in which it appears, with the words said before and after. In this example, the word is "order".

The detail also shows occurrences, sentiments, and usage examples of the word.

The examples show conversations in which the term appeared.

Agent Profile and KPIs

The Agent Profile defines the AI behavior and how it interprets data. Fill in the company information according to your business rules and click Save.

The KPIs are the basis for Waizer monitoring. Some dashboard metrics are guided by KPIs defined for the bot, keeping monitoring aligned with business goals. You can add, edit, or remove KPIs according to your operation’s reality. These data feed into the AI agent’s context and make the analysis more precise.

Integrations (MCP)

Besides conversation analyses, Waizer connects external systems through MCP (Model Context Protocol) servers. Thus, Ask Waizer crosses conversation data with information from CRMs, databases, ERPs, BI tools, and proprietary APIs.

In Integrations, click Add integration, choose Custom URL, and fill in:

• Name.

• Server URL.

• Headers, such as authorization.

Click Add to save the integration.

Attention: configure the integration with someone familiar with the MCP server used. Limit permissions to what is necessary for each use case, treat credentials and tokens as confidential information, and test the connection before releasing it to the entire operation. Usage must comply with your company’s security, privacy, and data governance policies.

Settings and Governance

In project settings, you find general information, Blip configurations, and the danger zone.

In Project Information, edit the name and view the project ID and dates.

In the Danger Zone, delete the project. Once deleted, this action cannot be undone.

Manage team members through your organization.

Click Add user to include new members.

Define each member’s Role to control permissions.

Edit permissions or remove users when necessary.

Adjust the panel theme between light and dark mode from your account menu.

Support

For questions, documentation, and technical support, visit Waizer’s Help Center at help.waizer.ai.

For information about plans, costs, and free trials, contact Wiv’s sales team at comercial@wiv.com.br.

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