It might be convenient to assign different sets of KPIs to different categories of conversations. For example, sales conversations might be measured by Upsell (abitliy to upsell new products, services, or warranties while selling another product), while technical support conversations could be measured by Troubleshooting (following the proper troubleshooting process with the customer).

In order to do that you have a few capabilties combined together: defining different KPI sets, defining different categories of conversations, automatic AI categorization of the conversation, and, lastly, assignment of KPIs to categories. Let's take a deeper look how to get those pieces together.

Categorize a conversation

This step assigns a category to a conversation, using predefined Category Collection. It is used for defining which KPIs to be analyzed for this type of conversation. A Category Collection contains list of categories. It is assiciated with KPI collection. It is possible to define which KPIs should be calculated and which not for a particular category.

This capability is useful when several different types of conversatons are analyzed, e.g. sales and tecnical support conversations. KPI for ability to upsell an extended warranty might be important for a sales call, while KPI about technical expertise might be more imprtant for a technical support call.

defining KPIs

It is required to create a KPI collection: Data/

To create a new KPI collection press "New" button and enter the fields describing the KPI collection.

![KPI Collection fields] (./assets/kpi-collection-fields.png)

Do not forget to press "Save" button to save the new collection.

On the next step, it is required to create KPI definitions for the KPI Collections. To do so, click on the KPI collection name.

![KPI] (./assets/kpis.png)

assigning KPIs to categories

![categories] (./assets/categories)