Setting up a chatbot to collect feedback
Description of the features
In this section we'll look how to teach the bot to save customer reviews in a Google Spreadsheet, so that later it can analyze the most frequent requests and improve the business.
The bot is able to find reviews in the correspondence that:
- The customer wrote himself, for example:
- «I liked the product, but it will be great if the delivery to the door appears».
- «I have a wish: if the prices will change, please send a new price list to the mail»
- «Great product! Easy setup and installation».
- An employee of our company asked to write to the client during the conversation, for example: «Hello. How are you? Did you manage to configure the system according to the instructions? If you didn't do this, please write where the difficulties occurred.»
- The system automatically requested the customer at the certain stage of the automation rules, for example, after the delivery of the product.
Setting procedure of the chatbot to collect feedback
|IMPORTANT! Next, it's assumed that you have already:
Configuration of the chatbot is a technically complex process. If you don't have a specialist with the necessary level of competence or you want to save your time, contact support and we will configure the bot for you.
Creating an Intent to collect feedback
To teach the bot to save reviews:
- Download the ready-made intent #Quality-Assessment Intent.
- Import the Intent into your bot. To do this, go to the Dialogflow website and in the menu on the left, make sure that you're making settings for the desired bot. Its name should be displayed.
- Go to the Intents section.
- Click on the button with three dots next to the Create Intent button and select Upload Intent from the list that opens.
- Click on the Choose File button and load the Intent.
- Go back to the list of intentions. To do this, click on the Intents section again. Intent #Quality-Assessment appears in the list.
- Open the Intent and, if it necessary, add or change training phrases in the Training phrases field. IMPORTANT! For the bot to work correctly, don't change the name of the intent.
- In the Action and parameters field:
- In the Responses field, adjust the response that the bot will send after it receives the feedback.
- Save your changes.
NOTE. We have prepared several popular Intents for the bot. You can download them from the example agent link and import them into your bot. For more information, see «How to download an agent».
Creating a Google Spreadsheet for saving reviews.
After setting up the intent:
- Go to the Google Sheets website and create a new file. For more information, see the description in the Google Help.
- To allow the bot to save reviews to the created table, grant it access rights to the document. To do this, click on the Share button.
- In the window that opens for the user email@example.com set the Editor access level. For more information on how to grant permissions, see the Google Help.
- In the browser's address bar, copy the link to the document.
- Go to the Bitrix24 portal and select «BotApp24 for open channels» in the menu on the left.
- On the control panel, click on the Bot Settings button.
- In the window that opens, in the Link to Google Feedback Table field, insert the link to the document.
- Save your changes.
Creating a bot to request feedback at the certain stage of working with the client
If you have completed all the previous settings, then your bot is already able to save the reviews that the client wrote himself during the correspondence and that the client was asked to write by an employee of your company.
In order for CRM to automatically request customer feedback at the certain stage of the automation rules, configure the ChatsApp24 bot for sending messages for the transaction, lead, or task stage. For an example of this setting, see the section «Bot for sending messages to WhatsApp». The bot parameters may look like this:
Let's say we added such bot in a transaction at the Delivery stage. In this case, when the transaction is at this stage, the message specified in the bot parameters will be automatically sent to the chat with the client. The response that the client writes will be considered a review and will be recorded in the Google Spreadsheet.
You have configure the bot that collects reviews. Sometimes it will correctly recognize the phrase with the review, and sometimes it will not.
To improve the recognition of reviews, periodically train the bot. For more information, see the section «Training the bot», subtitle «Train Chatbot».
Review Table analysis
The link to the Google Table with reviews doesn't need to be saved in the bookmarks. It's enough to select the section "BotApp24 for open channels" in the left menu on the Bitrix24 portal and click on the Feedback Table button.
Each month, a new sheet is created in the document, which stores the reviews of this month.
If you click on the link in the Client ID column, the client card in the Bitrix24 opens. This way you can find out who wrote the review. In addition, the card displays a detailed history of correspondence with the client. You can try to understand the reasons for such review or rating.