Yes! You heard it right. In less than 30 mins, you will be able to create the chatbot of your choice. I will walk you through the entire process of creating a chatbot in detail. Here, I will be creating a chatbot which answers questions related to the Corona Virus. So, without much ado, let’s dive into the process.
Step 1: Create a Dialogflow account
Dialogflow is a Natural Language Processing(NLP) platform which will help us in creating our chatbot without getting much into the technical intricacies of NLP. Go to https://dialogflow.com/ and sign up for free. It hardly takes a minute.
Step 2: Create an Agent
An agent in the context of Dialogflow is someone who will handle the conversation with the end-users. Now that you have signed up and are on the Home page, click on the Create agent button on the left side and you will land up on a page to fill in your agent details. Click Create once you are done. Here, I have named it Covid-19SupportBot because my Chatbot is all about Covid-19 but you can name it whatever you feel like.
Step 3: Modify the default Welcome Intent
As soon as your agent is created, you will land upon the Intents page. An Intent is an action which is being invoked by your end-user. For instance, when the end-user types, Hi or Hello or Holla!, what should the agent reply.
Once, you click on the Default Welcome Intent, you will see a list of Training phrases. You can add some more training phrases if you want to. The agent will identify from these training phrases and will respond from the responses that we configure.
When you scroll down the training phrases, you will see a list of default responses. The agent can randomly return one of those responses. You can configure the responses according to your choice.
1. Publishing your Chatbot on Microsoft Teams
2. 🤖 How to talk to Computers: A Framework for building Conversational Agents — Part 1
If you want the response to be the same all the time, just keep only one response and remove others. Just like what I have done here.
Voila! Now your welcome Intent is set. 😬Let’s get into the next step.
Step 4: Create Entities
An entity is an attribute which will help the Dialogflow to identify the dialogues of the user. Click on the Entities tab on the left side and then click, create Entity. An entity can have multiple reference values and each of which can have synonyms. I have created an Entity named Virus. Now, when referring to the virus, users can use words like corona, COVID and hence all those become part of the synonyms list.
Now create as many Entities as needed in your chatbot.
Step 5: Create Intents
Go to the Intents tab, click on Create Intent. Name your intent with respect to what it is all about. Next is to add the training phrases to help Dialogflow identify the Intent. Add as many phrases as you can think of which users might enter. Always remember, More training phrases => Better Accuracy. Also, notice when training phrases are entered, it will pick up the entities associated with that phrase. If you see that there is no entity associated with a training phrase, it’s time to add that synonym to your entity.
Next, move to the Actions and Parameters section, it would have automatically picked up the entity and value. The text under Value is something which is actually typed in by the user. So, you can use it to customize the response.
Next, is adding the response. This is the text which the bot will reply to your end-user. You can add multiple responses if you don’t wantt your bot to sound monotonous.
Step 6: Create connections with Web Services and Database systems
If you want to transfer information to and fro the backend systems, you can do it via fulfilment. This is specifically useful if you want to take any action in your backend based on the user’s input. It will help you provide dynamic content to your user. For instance, if your chatbot is scheduling doctor’s appointment, then fulfilment will help you get the real-time appointment slot instead of just a static response that your appointment is booked. Fulfilment also provides you with an Inline Editor where you can write your customized code. How cool is that! 😎
Step 7: Configure the Small Talk
Small Talk Feature is something which will help us handle the casual chat which the user does with the bot. This is important and so very easy. You just got to enable it and you are sorted. It has got all the default responses already filled in.
Step 8: The Last one, Launch your Chatbot 📱
Click on the Integrations tab then select the places, you want to deploy your chatbot on. The options are enormous.
This is just a handful of options that I have shown but there are many more. I have enabled the web demo for my chatbot and with that, you get the URL for your chatbot.
Share the URL with your friends and let them have some fun 😃
Take a look at how my chatbot looks like.