This is part 1 of a mulipart series on how to build your first chatbot. The complexity of the tutorials will increase with each iteration.
Building bots have become quite easy, to a point where you can deploy your first chatbot without any knowledge of coding. In this tutorial, we will be publishing a chatbot on facebook that can answer basic questions about you – Resume Bot.
Words you should know
Throughout this (series of) blog post(s) I will be using certain words which you might know. Here’s a jist of what they mean.
- NLP: Natural Language Processing: Used to make the computer understand the human language. So instead of writing a C program to turn off the lights, you could just tell the computer – "Ok Google, Turn off the lights"
- Intents: What you actually mean when you say something. When you say – Can you tell me your name. The intent of the aforementioned sentence is a question regarding a name.
- Entities: Objects/things with certain characteristics. For example in the sentence – My name is Anthony Gonsalves. The entity would be the name and its value – Anthony Gonsalves
- Agent: An agent is the instance of Dialogue flow that you train for specific NLP tasks.
So we will be making a resume bot that will be hosted on Facebook. A resume bot will answer all basic questions about you.
- Head over to dialogflow.com and register if you already haven’t.
- Click on create a new agent
- Let’s start adding intents
- Click on Create an Intent and give it a name.
Let the first Intent be- name
- Click on training phrases and two to three phrases asking someone his/her name.
- Then add the response that your bot will give to those questions.
- Click on save and try testing your new intent in the Try is now section to the right of the page.
- Add more intents and responses to those intents.
- Now it’s time to publish your bot to the world, and what better platform to do that than Facebook?
- Click on Integrations on the menu to your left.
- Enable Facebook Messanger.
- Follow the instructions here to create a facebook page and finish your Facebook integration.