What Are Slots in Watson Assistant?

If you’re using Watson Assistant to create chatbots or virtual assistants, you may have heard about slots. So what are slots in Watson Assistant and how can they help you build better conversational experiences?

In a nutshell, slots are used to collect specific pieces of information from users during a conversation. For example, if you’re building a virtual assistant for a bank, you might use a slot to collect the user’s account number or the reason for their inquiry.

At their core, slots are just variables that hold information until it’s needed later in the conversation. However, they have some powerful features that make them essential tools for building effective chatbots.

One important feature of slots is their ability to handle multiple values. For example, if you’re collecting information about a user’s hobbies, they might have more than one hobby to share. With slots, you can define a list of acceptable values and allow users to provide as many or as few as they’d like.

Another key benefit of using slots is that they allow for more natural and flexible conversations. Instead of forcing users to provide all the information upfront (which can be overwhelming), you can use slots to ask follow-up questions only when necessary. This allows users to provide information at their own pace and in a way that feels more like a natural conversation.

So how do you actually use slots in Watson Assistant? Here’s an example:

Let’s say you’re building a virtual assistant for an airline and you want to collect some basic information from users before booking their flight. You might start by defining some slots for things like departure city, destination city, and travel dates.

To do this in Watson Assistant, you would create entities for each piece of information (e.g., @departure_city) and then create corresponding slots in your dialog tree. You can specify things like the prompt message (what the user will see), any validation rules (to ensure the user provides valid input), and what to do if the user doesn’t provide the required information.

Once your slots are defined, you can start using them in your conversation flow. For example, you might start by asking the user for their departure city and then using their response to populate the @departure_city entity and slot. If the user doesn’t provide a valid city name, you can prompt them to try again until they do.

You can continue this process for all the required information (destination city, travel dates, etc.) and then use all that information to book the user’s flight. And because you’re using slots, you can handle unexpected inputs or changes in plans without derailing the conversation.

Overall, slots are an essential tool for building effective chatbots and virtual assistants in Watson Assistant. By allowing for more natural conversations and handling multiple values with ease, slots help ensure that your users have a positive experience and get the information they need.