Intents

What are Intents?

To explain it in a simple language, adding intents to the chatbot, is to train the bot to understand user intents (intentions). Users may enter any type of simple or complex queries to the bot. And therefore bot needs to be capable enough to respond to all variation of the queries. Intents help bot to understand and respond to complex queries. 

How to add intents?

Initially create intent by +CREATE INTENT option from the right of the dashboard.

 

 

While creating intent, you need to fill the below information

 

 

User Says:

Add the possible variants of the user queries to the bot separating it by pressing enter.

 

Bot response:

Set the bot response for the queries of “User Says”. i.e what should be the chatbot response for the queries you added in above user says section.

You can set bot response in the form of text, flow or conversation.

 

 

In the above example of hotel chatbot, an intent named Hotel Room booking is created in which user expressions (queries) are added and the response is set with the flow of “Travel Date”

Floatbot uses context under intent section. 

Context helps to control flow by defining specific intent. Generally, what the user types matches or is close to the trained intent the set action is triggered by the chatbot. With context, the intent will only trigger when the context is active. 

The activation and deactivation of context depend on Input context and Output context. They are the types of context.

• Input context tells the chatbot to match intent if it is close enough to what the user types and if the context is active.
• Output context tells the chatbot to activate context if it is not active.

For example,

You have trained two intents with phrase “Online Booking”. One shows online booking for travel ticket and another for hotel room booking. Now for the bot to trigger an accurate response, one more intent would trigger asking for “book travel ticket” or “book hotel room” output context. Each “Online booking” intent has a corresponding input context to ensure the intent matches the user’s request.

Output contexts

When you apply output context it activates the context if not active or renews if it is already activated. Whenever you add output context it is important to set the intent lifespan. The output intend expires after its corresponding intent matches.

Continuing our above example of “book travel ticket”

User says: Online booking

Bot says: Book travel ticket?

User says: Yes.   - Context is set and activated with lifespan 2

Bot says: Do you want to book travel ticket?   – Reduces to one

User says: Yes, want to book travel ticket.

Bot says: Sure, here is the process. – Reduce to one and becomes zero

Whenever the lifespan turns Zero the intent gets removed.

 

How to add output context?

• Click on Intents.
• Below that, click on the Add output context.
• Add phrases and click on save.

 

Input context.

Input increases the likelihood of the intent being matched after the context is active.

If contexts A, B, and C are active. An intent with A, B, and D is the input, the context will match for A and B, and will not match for D because the context of D is not active.  

How to add input context?

• Click on Intents in the left menu.
• Click on intent.
• Click on the Add input context.
• When you type a list of existing output contexts will appear. Click on the context you want to match.
• Click on save and then build.