When thinking about the 'things' a bot can do, we talk about abilities. A bot can have one or multiple abilities, depending on what it is to be used for.
Something to keep in mind when adding multiple abilities to your bot is that the more abilities are added, the more complex the Natural Language Processing model (NLP) will be. This could affect the performance as well as the accuracy of your bot.
At the moment any ability can only be added to the same bot once. For example, it is not supported to add two Meetings abilities to the same bot.


Abilities may have what we refer to as targets. These are names or words that refer to a particular 'thing' when triggering an ability. For example, when thinking about the Call ability, we could have the following targets:

  • Facebook
  • Phone
  • Skype
  • WhatsApp

These are all different systems that can be used to complete the Call action.

Another example is when thinking of the Leave ability. Leave is a common way of referring to 'time-off' in some countries, but others, like the USA, refer to Vacation instead of Leave. In this case, the ability can be changed to not make any reference to Leave, but rather use the Vacation terminology that suits the scenario / environment better.


When you add an ability to your bot, it will provide a range of actions. These vary per ability, but an example is when looking at the Meetings ability. Actions for this ability would be:

  • Book a meeting
  • List upcoming meetings
  • Cancel a meeting

As you can see, actions indicate on a more detailed level what a bot can do. Actions can be enabled or disabled individually.


Keywords relate to actions. These are trigger words that can be used by a user to initiate the particular action that they would like complete. For example, when looking at the book meeting action, the keywords could be:

  • Book
  • Organize
  • Schedule
Did this answer your question?