01. Analytics of Chatbots in Zenvia Customer Cloud

The Chatbot Analytics module contains a dedicated area for chatbot metrics for tracking purposes. With it, you have access to metrics without the need to create them manually from scratch. This allows you to monitor and analyze your bot's performance, such as customer engagement and satisfaction.

For example, you can gain insights into data such as the number of people who accessed the chatbot and how many of them were transferred to a human agent.

How to Access Chatbot Analysis

The Analytics module is located in the left sidebar of the Zenvia Customer Cloud. Click on this option, and then choose Chatbots.

Just below the screen description, you will see three tabs for navigation. Each corresponds to a specific panel:

Contacts panel: contains metrics related to the service provided by the chatbot.
Clients Dashboard: contains metrics related to the volume of clients who accessed the chatbot.
Quality Monitor: contains satisfaction indicators and interaction metrics.

All three panels have the same handling tutorial, differing only in the metrics displayed on the screen. Further in this article, you'll find all the specific metrics for each panel. For now, see how to handle them:

Set panel parameters

In the chatbot panels, you can set parameters to apply filters, monitor metrics, and indicators related to service, clients, and chatbot quality.

The main benefit of these panels is to provide customizable dashboards, supporting decision-making and ensuring a consistent and continuous user experience.

To start, define the filters according to your needs. This definition is the same for any panel you are viewing.

Chatbot: Choose which chatbot you want to analyze.
All Channels: Select the channel. By default, "All channels" will be selected.
Period: Set the time frame. By default, "Today" will be selected.

After the settings, click on the Apply Filters button next to the Period filter, and the corresponding metrics will be displayed.

Export Chatbot Dashboard

The process of exporting a Panel is the same on all tabs, whether it's the Contacts panel, Clients Dashboard, or Quality Monitor.

There are two options for exporting:

Suppose you are in the Contacts panel tab. For each metric, such as Contacts by channel, there is a download arrow when you hover over it. This option exports metric by metric separately.
In the top bar of the screen, next to Apply Filters, there is the Export Dashboard button followed by the name, such as Export chatbot service dashboard. This option exports all metrics after applying the filters.

In both cases, the file is generated based on the defined filters and the values displayed on the screen. However, you can choose specific metrics to be exported if the CSV format is chosen. The Download File button initiates the transfer through the browser in both options.

💡Tip: : Set up automatic metric updates and define a time interval, between 5 and 10 minutes, so that the panel always shows the latest metrics. Click on Set auto-update, located on the right side of the screen and below Export Dashboard, and define the interval you want.

Contacts Panel Metrics

You can use the Contacts Panel after launching your chatbot to monitor the performance of your new digital service channel. The available metrics are:

Calls: the total accumulated services created in conversation with the chatbot in the selected period.
Contacts by channel: graphical distribution of services per channel or the sum of all channels.
Retention rate: customers who interacted more than once with the chatbot, calculated as (Total conversations - Total abandonment - total conversations that had overflow) / Total conversations.
Abandonment rate: customers who did not complete the service, calculated as (Total conversations - Total conversations where the user failed to respond before triggering an end_chat - Total conversations transferred) / Total conversations.
Throughput: services transferred to human assistance, the percentage of conversations that had an overflow action activated in the bot.
Peak hours: average volume of services per hour, considering all services over the days within the analyzed period.

Clients Dashboard Metrics

You can use the Clients Dashboard after monitoring the Service Panel to complement your analysis with data such as who your active clients are, which ones are recurring, or the volume of clients you reached in a certain period. The available metrics are:

Active clients: the number of clients who had at least one service, the number of contacts created in conversation with the chatbot. In other words, counting each individual who generated one or more services only once.
Repeat clients: the number of clients who had more than one service, the volume of contacts that had more than one conversation in the selected period in the filter. In other words, counting each individual who generated more than one service (conversation) in the selected period.
Recurrence rate: the percentage of active clients who interacted with the chatbot more than once, recurring contacts divided by the total created contacts (in percentage).
Volume of customers over the period: the volume of clients over the selected period.

Quality Monitor Metrics

In the Quality Monitor, you can check the opinions of clients about your company. The main benefit of this functionality is to provide an analysis of customer satisfaction and engagement, enabling improvements in your product or service. The available metrics are:

CSAT: the Customer Satisfaction Score (CSAT) is a scale indicating customer satisfaction when they have just used a product or service offered by the company. The scale ranges from 1 (low) to 5 (high), and the result represents the portion of satisfied customers (ratings 4 and 5).
NPS: the Net Promoter Score (NPS) is a scale measuring the level of customer satisfaction and loyalty with the brand, based on their willingness to recommend it to others. The result is calculated by the difference between promoters and detractors and ranges from -100 (critical) to 100 (excellence).
Average interactions: the average number of messages exchanged between the chatbot and clients.

There you have it! Also, explore the Flow Map of the Chatbot module, which allows you to understand and optimize the journey of clients interacting with the chatbot.