"Is my bot able to help users quickly and effectively?"

In order to evaluate whether your assistant is able to handle user needs, you will want to monitor the areas and topics in which the assistant is successful and where it fails. The Conversation Metrics view is used for this purpose. 

  1. In the Efficiency graph, you can focus on how user queries are handled in terms of RT (Response Time = average conversation length) and FCR (First Call Resolution = percentage of conversations where the customer's request was resolved). These metrics are available for the 10 most common topics and also include indicators of change (trends) in relation to the previous period. 
  1. In the FCR percentage chart, for example, you can see that within the topic of Lost or Stolen Card, the assistant helped resolve the issue in 80% of conversations, while for the topic of Savings Accounts, this was not achieved in any of the cases. From this, you may conclude that the topic of Lost or Stolen Card has much better coverage than the topic of Savings Accounts. Thanks to these insights, it is easier to focus on improvements to areas with a long-term low FCR, in this case the topic of Savings Accounts. 
  1. It is also useful to track changes over time using trends. If a significant drop in FCR has been recorded for any of the topics, for example from 90% to 35%, this may mean that a change for the worse has been implemented in this area, or that users have started asking questions about a topic that your virtual assistant is not sufficiently prepared for. 
  1. It is equally important to monitor Abandoned conversations. Inquio’s analysis can reveal whether a user left the conversation before their request was satisfactorily dealt with. This metric tells you how many users gave up on communication and left prematurely. An increase in the number of abandoned conversations is a warning sign that the dialogue is not flowing as it should. 
  1. The Empty conversations graph shows the percentage of conversations that were abandoned without the user writing anything. This indicates that the virtual assistant’s initial approach did not interest the user – or even discouraged them – and it is therefore advisable to focus on this part of the interaction from a dialogue design and business perspective. 

Related