This feature is a no-code solution. 🍰
Even if your chatbot offers buttons, quick replies, or other options for users to click in every step of the flow, they might still type in their own text sometimes. And when your bot doesn't know how to respond to these custom text inputs, the user will be stuck, and may even get confused or frustrated enough to abandon the chat. This feature can help solve that problem through the power of artificial intelligence (AI).
How it works
In Chatfuel's Grow tab, you'll see a table called Teach your bot to respond to custom text inputs. This is where a list of all the custom messages users type in and send to your bot will appear.
The first column on the left lists the message text itself. This is the message user(s) have typed out and sent to your bot.
The second column lists the frequency for each message, or the total number of users who have ever sent that word or phrase to your chatbot.
The third and fourth columns and the blue buttons on the end allow you to set up rules for how you want your bot to reply to those words/phrases in the future.
How to set up AI chatbot responses
Now that you've seen the list of custom messages that users commonly send your bot, set up rules so it can respond accurately to them in the future. Here's how:
Navigate to the Grow tab and locate the Teach your bot to respond to custom text inputs table.
For any custom text input listed, choose an option from the drop-down menu under the Bot replies with column: Text, Block or Flow, or Existing Intent.
In the box to the right, type out or select the response you want your chatbot to reply with.
Click the blue Set up button.
There are three options in the Bot replies with drop-down menu. Here's how to set up each one:
Text. If you want your bot to reply to a custom input with some text, select Text from the drop-down and type your text in the box. Then click Set up. For example, if the custom user input was "Hello," you could choose Text and write "Hi there!" in the box.
Block or Flow. If you want to send users to a certain block or flow in your Messenger bot after they send a certain custom input, use this option. You'll just need to select the block or flow from the drop-down and then click Set up. For example, if the custom user input was "shipping," you could send the user to a flow that gives information about your brand's shipping times and costs.
Existing intent. Have an existing AI rule that it makes sense to apply to a custom input on the list? Use this option. Once you select this option under Bot replies with, choose an AI rule that you've already made from the drop-down menu that'll appear in the next box, and click Set up. For example, if you already have an AI rule that sends anyone who types "customer service," "speak to a human" "human," "talk to a human," etc. to a Live Chat session, you can make that rule apply to the custom text input of "help" on the list.
Managing your rules
If you want to see all the rules you currently have set up, click View all AI rules in the bottom-right corner of the table. This will take you to the Set Up AI tab, where you'll see all your existing rules listed. You can add more rules here by clicking +Add AI rule, whether they're for words/phrases listed in your Teach your bot to respond to custom text inputs table or not.
The purpose of the table is just to show you the most commonly typed-in responses from your users, but in the Set Up AI tab, you can make AI rules for any words or phrases you like. Learn more about this in the video below. 👇
Use cases for this feature
This feature can benefit any type of Messenger chatbot for any industry. The more accurately your bot can respond to users, the more helpful it will be to them, and the less drop-off you'll have.
If you want to be proactive, go to the Set Up AI tab and create rules for words and phrases related to common topics that you think users will ask. Then, check the table in the Grow tab regularly to see if there are new words/phrases that users are entering but that you don't have rules for yet.
💡 For example, an ecommerce shop selling protein bars could set up rules for words like ship and shipping. If they later notice in the Teach your bot to respond to custom text inputs table that people are also asking about the nutrition and ingredients of their bars, they could add rules for those and other similar ones, like organic, gluten-free, vegan, nutrition facts, etc.