Matching sources

Inbenta tracks the actions that users perform with every product (ChatbotKnowledge Management and Search). There are two categories of tracking: Atomic and Sessions tracking. This data is logged separately, and it is used to generate the reports that you see in the product app's Dashboards section.

This section describes possible Matching Sources. To see the different categories of logsclick here. To see what kind of Actions they track, click here.

Matching sources

By default, all searches tracks matchings as internal results, because all results come from a single source, which is the product instance. However, if you set up your application to connect to other instances, this results in multiple sources. In this case, matchings from these other sources will be tracked as external

In Chatbot, where the application triggers a search event with every message, matchings from within the Chatbot content are tracked as "no external" (i.e. internal results), whereas AIML or Federated Bot matchings are tracked as external.

Example: You set up your SDK from Instance X, but use the Builder to add an extra component to also connect to and run searches in Instance Y. Instance X acts as your main instance, which means that matchings from this instance are tracked as internal in the logs. Instance Y is a secondary instance, and matchings from that instance are tracked as external

Suppose that you have a support page created offering FAQs with the Knowledge Manager, and you want to add additional results from your website. You set up a Search instance to deliver these website results. In this scenario, you could set the KM instance to be the main instance (internal) and the Search instance to be secondary (external), or vice versa.