Understanding Intents in Dialogflow

Dialogflow is a highly robust and dynamic natural language understanding platform that allows you to quickly build highly scalable conversational chatbots by leveraging the power of Artificial Intelligence (AI).

You can easily create highly customizable chatbots for variety of channels like web applications, mobile applications, Facebook page, social media apps and much more. It provides seamless integration with variety of channels like Facebook, Viber, Telegram, Slack and much more.

But in order to start working on building a chatbot yourself, lets connect the dots to get a clear picture of entire chatbot development process in Dialogflow. To create a chatbot, let us understand some of the nuances of the Dialogflow.

What are Intents?

Intents help Dialogflow ascertain the real intent or motive of the user messages. You can provide a set of certain training phrases that can help Dialogflow get trained to recognize similar such phrases and trigger a specific intent whenever such phrases are encountered.

Lets look at this example, if you want to create a chatbot that can display current score of Indian or any cricket team and for that you need to first know the name of the user who is interacting, then first you can create an intent name GetName (you can give any name as you want) and then provide different training phrases or user expressions that will help Dialogflow to trigger this intent.

Creating Intents

To create an intent inside a Dialogflow bot, follow the steps below:

1. Select the desired chatbot.
2. Select the Entities option.
3. Enter the desired intent name.
4. Add many training phrases to trigger this intent.
5. Click the Save button.

Training Phrases

Lets dive deep into the actual training phrases that will be used to train Dialogflow to invoke this intent when user inputs exact or similar phrases. From the above example where we have to fetch user name from the user expression, we can provide different expressions that will broaden the scope of training and extract the exact name of the user. You can provide expressions like as shown below:

1. Nikhil is my name
2. My name is Nikhil
3. Nikhil
4. I am Nikhil
5. This is Nikhil
6. Nikhil and how are  you? 

      and so on.

There is no limit to the number of user expressions you can enter as training phrases. You can train Dialogflow to extract certain entities from these phrases, which we will cover in our subsequent blogs on Entities.

You can also test out your intent by enter the relevant training phrase in the try it now text box in the panel on right hand side.

Any response that you set gets displayed in the Default Response section.

In the subsequent parts of the blogs, we will dwelve deep to further know concepts like Entities, Fulfillment, Webhooks, Contexts and much more.

Till that time, happy coding!