Machine Learning is a matching system that uses Natural Language Processing and algorithmic probability to analyze user input and to determine its similarity to available interactions.
Overview
-
Matching systems used in User Input interactions analyze every message sent by the customer.
-
Machine Learning is the prioritized matching system for the User Input interaction. There are a few exceptions to this rule. You can check the list of exceptions in this article.
-
The default Confidence Score for all your stories is
0.5. You can adjust this number by following steps from this article.
How Machine Learning works
When a customer talks to your bot, the system picks all of the queries and analyses them. The result of this analysis is expressed as a number called the matching score. This score is then compared with the Confidence Score, which defines the minimum value that has to be reached to trigger an interaction successfully.
As the Machine Learning system is powered by artificial intelligence, the query doesn’t have to match 100% the User Says field. Typos and various syntaxes are allowed.
When to use a Machine Learning
Machine System analyses the whole user input. That’s why this matching system is recommended for most of the interactions such as:
-
General questions, like:
-
Where are you located?
-
How can I contact you?
-
-
Providing data, like:
-
My order was not delivered.
-
I need more information about this product.
-
Read more:
