When you create an agent, there are two intents already created and configured for you— the Default Welcome Intent and the Default Fallback Intent. The Default Welcome Intent is called whenever the user begins a conversation with your agent via supported one-click integrations, and should return a response that lets users know what your agent does or what users can say to begin a conversation. The Default Fallback Intent is matched if your agent doesn't recognize user input.
You can delete these pre-configured intents; however, it's recommended to use these in your conversation and tailor them for your particular agent. Then, you can add your own custom dialog intents to create a complete conversational interface.
Default Welcome Intent
The Default Welcome Intent is created for you when you create your agent. This intent has a special
WELCOME event attached to it, which is triggered whenever the user begins a conversation with your agent via supported one-click integrations. The Default Welcome Intent can also be matched through its training phrases, which are pre-populated with common greetings.
When the Default Welcome Intent is matched, your agent will respond with one of the pre-populated text responses in the Responses section. You should delete these pre-existing response options and specify responses tailored for your conversation. You can define as many responses as you'd like in the Default Welcome Intent. Each time this intent is matched, your agent returns one of the responses. For each subsequent matching, your agent selects a different response until all responses have been used; then, it begins again choosing responses randomly.
Default Fallback Intent
When you create an agent, the Default Fallback Intent is automatically configured with a variety of static text responses, like "I didn't get that. Can you say it again?" and "Sorry, what was that?" This intent is matched when your user's input does not match any other intent; in other words, it's a kind of catch-all for any unrecognized user input.
For example, say that your agent has only one custom intent named
weather that recognizes user input like "What is the weather today?" or "Forecast tomorrow". If a user speaking to your agent says "I like the color purple", your Default Fallback Intent is matched because the agent isn't able to match the input to the
You can change the pre-defined text responses in the Default Fallback Intent to anything you'd like, but they should communicate to the user that their input was not recognized. Just as with the Default Welcome Intent, your agent randomly selects a response variant to send to the user when the Default Fallback Intent is matched, and selects a different response for each subsequent matching until all the responses have been used. Then, your agent will begin again selecting responses at random.
You can add training phrases to fallback intents that act as negative examples. There may be cases where user queries have a slight resemblance to your training phrases, but you do not want these queries to match any normal intents.
For example, a room booking service may have a training phrase like "I'd like to book a room". If the user wants to purchase a book about rooms, they may say "I'd like to buy a book about rooms." To ensure that the user query does not match your intent, you can add that phrase as a negative example.