Articles on: AI Chatbot & Automations

How do I create an AI Chatbot for Facebook Messenger?

Crisp no-code chatbot builder helps you to create your own AI chatbot for Facebook Messenger. Draw complex response flows, with events (eg. user sent message), actions (eg. send response message) and conditions (eg. check segments). With our Bot Builder for Facebook, you can craft customized experiences on Messenger in little to no-time.


This guide goes into explaining how to drive automated conversations on Facebook Messenger. As an official Meta Partner, Crisp offers you to connect Meta Business to our chatbot software, for free.


Note that if you haven't connected Meta Messenger to our chatbot platform, you should follow this guide: How to connect Crisp with Messenger?


If you're not an existing Crisp customer, creating a bot for the Meta Business Inbox is 100% free for the time of your trial (which can go up to 30 days). Make sure to create an account before going through the tutorial.


➡️ To create an account Crisp, please click here.


What do I need to build a Messenger bot?


To build automated responses for Facebook Messenger, you'll need the following requirements:

  • An admin access to Meta Business Suite
  • A Meta Business Account connected to Crisp
  • A subscription to the Essentials plan
  • A Crisp Account set as "Owner"


Create an AI chatbot using Meta Messenger


Now that you're ready to build an AI bot for Messenger, let's get down to the creation process. For the sake of this tutorial, we're going to create a department routing bot that will require WhatsApp users to interact buttons on which he can tap to be routed to the right department in your company.


  1. Train the AI to speak and sound like you do


Five kinds of data sources can be used, they all work together to generate better responses:


  • Answer Snippets : write short Q&A (Questions & Answers) about common themes your customers frequently ask you. Those Q&A are solely used by your AI and are not visible by your users. This where you can enter all your custom and private data that you can't add through the other sources.
  • Web Content : add all your Web domains, and we extract text from all your Web pages for you. We train the AI with the main textual content from your Web pages. If you have an online documentation, adding its domain can bring valuable of knowledge to your AI.
  • Knowledge Base Articles : as you create articles on your Crisp Knowledge Base, we feed them to your AI to train them. Those articles will be visible by your users on your Crisp Knowledge Base and used by your AI to suggest helpful Knowledge Base articles in automated AI-generated responses.
  • Inbox Messages : all messages sent by your human agents to your customers are used to train your MagicReply AI ( note that we do not use your Inbox Messages to generate automated AI answers to your users for privacy reasons, we only use them to suggest AI-generated answers by your human agents within the Crisp Inbox ).
  • Data importer : As you store knowledge internally, such as in .txt files, pdf, or CSVs, you can now upload them into Crisp to power your AI with PDF, CSV or .txt files.


Data Sources can be configured in the Automations section of the Crisp App


  1. Open the "AI Chatbot" Section



  1. Click on "New Scenario" at the top right and hit "Use a Sample Scenario". From there, choose "Department Bot".



  1. Hover the "User Message Matches" block and click on the Pen to edit the script behaviour.



  1. In the Origins drop down, select "Messenger" so the bot triggers on messages coming from Messenger only."


Note that if you want the bot to become omni-channel, that's totally possible as Crisp supports more than 10 different channels.



And voilà, you have created your first chatbot for Facebook Messenger. Now, you can take advantage of the numerous featured offered by Crisp.


How to test the AI Chatbot before pushing to prod?


At that point, you're almost there. And it's vital you check how the chatbot behaves, and answers your key questions so you can refine AI's knowledge about your business.


The chatbot editor interface has an integrated menu to help you test your scenario directly, as covered earlier in this article. However, in some cases you may want to test it live if it contains advanced conditions, or if it is composed of several inter-connected individual scenarios.


A good way to test a scenario "live" without causing interference with your users, is to edit the terms detected by your initial "User sent a message" Event block, and to temporarily make it detect a term only known by you (like a password).



Going beyond with Crisp AI Chatbot Builder


Crisp AI chatbot builder offers a large number of possibilities for Facebook Messenger. For example, you could create an AI FAQ bot through Facebook, using Crisp Knowledge Base.


Note that you can connect Crisp with Instagram and WhatsApp so you have one scenario built for the 3 channels:


Want know more about Crisp Chatbot Builder? Click here to access a dedicated guide.


Frequently Asked Questions


Still have some questions about how to connect the Bot and social media channels that were not covered in this article or related guides? Perhaps you'll find the answer here!
Here is a collection of the most frequently asked questions, you'll also find some extra resources to help you better understand how this plugin works, and to prepare for the most obscure and extravagant use-cases.


Don't hesitate visiting this in-depth article to learn more about the Bot and master your craftsmanship of scenarios


If your question does not figure here, share it with us directly by chat, we'll be available!


Does this also work for Instagram and other, social medias?


Absolutely, the same can be done with Instagram to detect messages from this channel, simply by selecting it in the "origins" of your New User Message event block.


With Instagram, you can also specifically target replies to your story, by instead using the Event block User Message Matches with the value **Reply to story: ** which allows you to detect messages where this label has been automatically appended by Crisp

Updated on: 02/07/2025

Was this article helpful?

Share your feedback

Cancel

Thank you!