# intent

### <mark style="color:orange;">`intent(name=String|Array, &block)`</mark>

{% hint style="success" %}

### Configuration

The NLP engine must be enabled and configured in [`bot/config/nlp.rb`](/getting-started/nlp-configuration.md) file in order to implement this block.
{% endhint %}

## **Platforms**

<table><thead><tr><th width="362.78343949044586">Platform</th><th data-type="checkbox">Supported</th></tr></thead><tbody><tr><td>Messenger</td><td>true</td></tr><tr><td>WhatsApp</td><td>true</td></tr><tr><td>Telegram</td><td>true</td></tr></tbody></table>

## Usage

In the following example, <mark style="color:orange;">`MainContext`</mark> will handle 3 different intentions:

* `greeting`: Greeting messages like "Hi", "Hello" and so on.
* `goodbye`: Goodbye messages like "Bye", "GodBye", "Goodnight" and so on.
* `thanks`: Thank you messages like "Thanks", "Thank you", "I appreciate it" and so on.

```ruby
class MainContext < Conversation

  def blocks

    intent "greeting" do
      @reply.text "Hello!"
    end

    intent "godbye" do
      @reply.text "Bye bye!"
    end

    intent "thanks" do
       @reply.text "You're welcome"
    end
  
end

```

### Example of how they look in Wit.ai

As we mention before, each intent must has been created and trained in the NLP Engine (in this case Wit.ai)

![](/files/siVPOIQz3sOFfp5ltSqs)

## Reading Params

This block passes 4 parameters that are: <mark style="color:orange;">`text`</mark>, <mark style="color:orange;">`entities`</mark>, <mark style="color:orange;">`traits`</mark> and <mark style="color:orange;">`confidence`</mark>, which contain additional information to the intention itself.

### Usage example

For the example below, let's assume the incoming message says the following:

*<mark style="color:blue;">"Please wake me up tomorrow at 7am. I'd really appreciate it"</mark>*

Assuming also that we have been created and trained an intent called <mark style="color:orange;">`set_alarm`</mark>, that is linked to the entity <mark style="color:purple;">`wit/datetime`</mark> and the trait <mark style="color:purple;">`wit/sentiment`</mark>.

![Wit.ai screenshot](/files/StJzbf6w2hwdz8rpmk5C)

In the code below, thanks to these parameters the block can send an even more appropriate reply:

```ruby
class MainContext < Conversation

  def blocks
  
    intent :set_alarm do |text, entities, traits, confidence|
    
       unless entities["wit$datetime:datetime"].nil?
       
          entity = entities["wit$datetime:datetime"].first
          @reply.text "I'll wake you up at #{entity[:value]}"
          
       else
       
          @reply.text "To help you with that, I need you to tell me a time for the alarm."
          
       end
       
       @reply.typing 1.second
       
       unless traits["wit$sentiment"].nil?
       
          trait = traits["wit$sentiment"].first
          case trait[:value]
             when "positive"
             
                @reply.text "And thank you for asking so kindly."
                
             when "negative"
             
                @reply.text "But you could try to be nicer next time.."
                
             when "neutral"
                # Nothing here
          end
       end
       
    end
  
end
```

### Params definition

#### <mark style="color:orange;">`text`</mark>

Just the text message <mark style="color:blue;">"</mark>*<mark style="color:blue;">Please wake me up tomorrow at 7am. I'd really appreciate it</mark>*<mark style="color:blue;">"</mark>

#### <mark style="color:orange;">`entities`</mark>

Array with the entities found.

```json
{
  "wit$datetime:datetime": [
    {
      "id": "313292537627827",
      "name": "wit$datetime",
      "role": "datetime",
      "start": 38,
      "end": 53,
      "body": "tomorrow at 7am",
      "confidence": 0.9995,
      "entities": [

      ],
      "type": "value",
      "grain": "hour",
      "value": "2022-04-30T07:00:00.000-07:00",
      "values": [
        {
          "type": "value",
          "grain": "hour",
          "value": "2022-04-30T07:00:00.000-07:00"
        }
      ]
    }
  ]
}
```

#### <mark style="color:orange;">`traits`</mark>

Array with the traits found.

```json
{
  "wit$sentiment": [
    {
      "id": "5ac2b50a-44e4-466e-9d49-bad6bd40092c",
      "value": "positive",
      "confidence": 0.9047
    }
  ]
}
```

#### <mark style="color:orange;">`confidence`</mark>

The percentage of confidence of the NLP engine in associating the message with the intent.

<mark style="color:green;">`0.998`</mark>

## `any_intent` Block

Catch any intent of the message and returns 5 parameters: <mark style="color:orange;">`intent`</mark>, <mark style="color:orange;">`text`</mark>, <mark style="color:orange;">`entities`</mark>, <mark style="color:orange;">`traits`</mark> and <mark style="color:orange;">`confidence`</mark>.

### Usage

```ruby
class MainContext < Conversation

  def blocks

    any_intent do |intent|
      if intent == "gretting"
        @reply.text "Hello!"
      elsif intent == "godbye"
        @reply.text "Bye bye!"
      elsif intent == "thanks"
        @reply.text "You're welcome" 
      end 
    end
  
  end

end
```

{% hint style="success" %}

### Routing to Context

An intent can be routed to a specific context, learn how in [Routing Chapter](/contexts/routing.md#intent).
{% endhint %}


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.kogno.io/contexts/blocks/intent.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.