Granularity of Intents

-> to BOTwiki - The Chatbot Wiki

The granularity of intents represents the depth of detail and content of the intents of a chatbot. This means that the more granular a chatbot is built up, the more individually it can respond to certain requests. This is illustrated in the following section using an example:

  • Fine granular intent:

Question: "Does the Bundestag offer group tours?"

Answer: "Group tours can also be booked in the Bundestag. More information and booking can be found here:ührungen

  • Coarse granular intent:

Question: "Is the group tour in the Bundestag barrier-free?"

Answer: "All information about group tours can be found here:ührungen=- sync:ßÇÈâÈâ

How granular should intents be?

The granularity depends very much on the use case and the complexity. Often it makes sense at the beginning not to plan the contents of a chatbot too finely granular. One should pay attention to the interaction of the user with the chatbot during the development and especially during the test phase, in order to then be able to decide how deep users want to go into the details. Of course it is possible to adjust the granularity during and after the introduction of the chatbot and to create new intents to ensure the level of detail.

Which problems can occur with too granular intents?

The more detailed the intents, the more difficult it becomes to maintain the content of a chatbot. Of course, a descriptive naming of the created intents helps to ensure that several developers can work on one chatbot. However, there is another challenge that occurs when the intents are too granular. With very similar and very close requests, it is difficult for the NLP service to identify the right intention behind a user's request. The NLP service must decide, by creating various utterances, how sure it is that this request is part of this intention. If these question possibilities are too similar, the confidence score and the machine is no longer certain of the intent behind this request.

Overall, it should be said that a middle course should be chosen that will be crystallized by testing with real users.


> Back to the BOTwiki - The Chatbot Wiki