Introduction to Entities

Entities are a rule-based recognition mechanism that provide a way of steering Conversational AI Cloud's recognition to a specific article after an intent is matched on the users input, or to match the entities a user may refer to when going through a Conversational AI Cloud dialog. In its core, it checks the contents of an entity against the contents of the user’s input. If there is a match between the contents of the entity and the user’s input, it constitutes an entity match.

Entities consist of a word, a number of words, or a single regular expression that all refer to the same part of a user’s input. This makes entities an ideal complement to intent recognition. Where intent recognition focuses on what the user is trying to achieve, entities focus on the variable(s) that may occur within that intent (the entities).

For example, if a user asks the question "I'd like to buy an iPhone" then the user’s intent would be "buying". The entity in this example is "iPhone”. The entity should contain the word "iPhone", but could also contain words "iPhone Pro", “iPhone SE” and so on.

Conversational AI Cloud allows for different types of word combinations within entities as well, in the example above "iPhone Pro", and "iPhone SE" are considered compound word entities, which we’ll explain more in-depth in our advanced guide.

Aside from word-based entities, Conversational AI Cloud also allows for the definition of entities based on regular expressions. These expression type entities make it possible to capture dynamic input like a postal code, or a phone number without having to think of and maintain all the different variations through a standard entity. If used correctly, expression entities based provide a sustainable way of recognizing complex user input.

Want to learn more about how entities work together and complement other recognition mechanics within Conversational AI Cloud?

View the Recognition topic.