System entities: they are pre-built by calldesk and can be used without any configuration
Custom entities: they are created by you depending on the data you want to parse
Name Entity Recognizer (NER) are used to recognize an entity inside a query. To better understand how the system works, here is a quick example.
Let’s say that the users says:
<aside> đź—Ł My tracking number is 1A132456789
</aside>
The NER will parse this query to find the entity that matches a tracking number.
There are 2 different kind of entities that can be created in the studio:
System entities: they are pre-built by calldesk and can be used without any configuration
Custom entities: they are created by you depending on the data you want to parse
Custom entities can be of 2 different types:
Simple entity: this is a simple list of values that you want to match in users’ queries. As an example, you want to match a list of fruit. You can also define variants, that are alternative ways of saying the same fruit, or custom properties.
Regular Expression (RegEx): this is an entity based on a pattern. For instance, you would like to match a tracking number that always start with a number between 1 and 3, followed by the letter A, B or C, followed by 9 numbers between 0 and 9. Again, you can define custom properties for each pattern.
Under the hood, there is a NER associated with the entity. Each system entity has its own NER in order to parse data in users’ queries. Regarding custom entities, simple ones use a specific NER that will build a database based on the list of values provided. The regex ones use the pattern provided to look for it inside the users’ queries.
<aside> ⚠️ System NER are also language dependant. They are not supported by all languages of the platform.
</aside>
Your custom NER will then be integrated as a system NER, available on the language it has been developed for.