| Add-on | AI agents - Advanced |
AI agent Satisfaction (BSAT) Rating allows you to measure customer happiness with the AI agent experience and make targeted improvements to intents or replies.
To start collecting feedback on your AI agent's performance from your customers, it needs to be enabled.
You can enable AI agent Satisfaction (BSAT) Rating collection, by navigating to Content > Intents > Collect BSAT reply
This article covers how to:
- Enabling AI agent Satisfaction Score (BSAT) collection
- Filtering by BSAT in Conversation Logs
- BSAT Analytics
To learn more about what AI agent Satisfaction (BSAT) Rating is, read System Replies Explained.
Enabling AI agent Satisfaction (BSAT) Rating collection
You have two ways to collect BSAT: triggered by resolution states or as linked replies.
BSAT collection with resolution states
By Clicking on Collect BSAT reply in the Content > Intents menu, you can navigate to Triggers. Here, you can turn on feedback collection, by clicking on the ‘Enable BSAT collection’ switch. If you would like to turn off this system reply for your AI agent, you can do it with the same switch anytime.
In the ‘Trigger reply when’ section, you can set up when should the BSAT collection prompt appear. Collecting BSAT this way relies on resolution states, so make sure you have them set up.
You can select any resolution state and add how much time should pass before the customer sees the question.
The BSAT reply triggered by resolution states only appears once per conversation.
BSAT collection with linking to another reply
You can choose to trigger the feedback reply by following any node in the Dialogue Builder with a Link to another Reply and adding a BSAT reply. In this case, you have to link the BSAT reply to every scenario where you would like to have it triggered.
Disabling triggers for BSAT replies on the Triggers tab does not affect linked replies. It is not recommended to use both the resolution state-based triggers and linked replies at the same time.
Configuring AI agent Satisfaction (BSAT) Rating collection
On the Settings tab for Collect BSAT reply in the Content > Intents menu you can select what type of scale should be displayed in your conversations when the BSAT reply is triggered. By default, Collect BSAT reply uses emojis, but you can also select a Star⭐️ rating system. In conversations, all options are rendered as buttons.
If you want to set up different wait times for different resolution states, just add another criterion with the ‘Plus’ button and select the relevant resolution states.
After navigating back to ‘Replies’ and clicking on the reply, you can configure the dialogue.
- On the ‘BSAT Question’ block, you can select how the AI agent will ask for a rating. In addition to the three predefined questions, you can add a translated version for all supported languages by your AI agent. To do this, create a new reply with the appropriate language selected and add a translated question. Customers have the option to rate the AI agent on a scale of 1 to 5 (with stars or emojis) or click on the ‘Not now’ options.
- On the ‘Not now’ flow, you can simply add an AI agent Message to close the conversation.
- When the customer rates the conversation, you can access the score using the {{bsatScore}} parameter.
- With a Conditional Block, you can set up AI agent messages for different scenarios, like for scores greater than or equal to 3 and scores below that value.
Filtering by BSAT in Conversation Logs
In Conversation Logs, by clicking on the ‘Add Filter’ button, you can access the BSAT filter. 
By filtering by responses given to the BSAT Question, you can deep-dive into these conversations to look for opportunities to improve.
BSAT Analytics
To review the BSAT rating collected from your customers, navigate to Analytics > AI agent Analytics > Performance Summary.
In the Performance Overview section, navigate to the Areas to Investigate widget and select the BSAT KPIs
| Navigation | View |
 |
 |
| Visual |
Description |
 |
The BSAT score - Overall indicates how satisfied customers are with your company's services. It's measured through customer feedback and expressed as a percentage (100% would be the highest and the higher the better) The overall BSAT is the total number of conversations with a rating of 4 or 5 divided by the total number of conversations with feedback provided, then multiplied by 100. |
 |
The Average BSAT Score is the sum of all collected ratings divided by the total number of conversations where feedback was provided by your customers. The metric provides an overview of where the majority of the rating falls. In this example, it seems that the average score is 2.27 |
 |
The Response Rate metric overviews the percentages of conversations where your customers provided rating feedback. example: out of the 948 conversations where the AI agent triggered the BSAT reply and asked for a rating, only around 16.4% provided a rating. |
In the AI agent Performance Analysis section, you can review the AI agent's performance through the collected BSAT. On the left side, use the dropdown list to select the relevant widget.
| Description |
Visual |
|
AI agent Overall Satisfaction Score - Overall Trend A trend view to review how the overall AI agent satisfaction Score fluctuates over time, week over week. Through this view, you can easily see how the satisfaction rate is progressing or regressing. |
 |
|
BSAT by Intent A stacked bar chart to review the breakdown of the collected BSAT score per intent where you can analyze the AI agent performance through intents and BSAT scores. The order of the intents is listed by the ones with the most BSAT scores provided. In other words, conversations where the intent was triggered and your customers provided a BSAT score. You can use the slicer on the left side to focus on the intents you are looking for. |
 |
On the right side, use the dropdown list to select the relevant widget.
| Description |
Visual |
|
AI agent Overall Satisfaction Score Trend by Language A trend view to review how the overall AI agent satisfaction Score fluctuates over time, week over week, segmented by language. Through this view, you can easily see how the satisfaction rate is progressing or regressing for each language. |
 |
In the Intent Performance section, you can review the average BSAT score per intent.
The BSAT score is the sum of all collected ratings divided by the total number of conversations where feedback was provided by your customers. It is visualized in the intent performance table from the perspective of the intents where each intent in the conversation inherits the same rating of the conversation.
Limitations
Due to limiting the number of buttons for each message, the Collect AI agent Satisfaction (BSAT) Rating feature is not usable with WhatsApp. This is a limitation due to the design of WhatsApp. You can refer to their documentation on the subject here.
0 comments