Understanding the Complexity of Real-Time Analytics in the Contact Center
Author: Simon Black | Date: 17/04/2023
Real-time analytics and post-interaction analytics can both be used in the contact centre to improve performance.
They both drive insight-led efficiency savings and assistance for the agent.
And yet, as I’ve discussed before, building real-time analytics services can be far more complex than post-interaction due to the cost, infrastructure requirements and pressure on audio / transcription quality. Conversely, post-interaction analytics can be spun up in hours and can provide a range of insights immediately to operations teams as it does not require significant infrastructure changes.
Organisations need to seriously analyse their customer use cases and decide which method or which combination of methods is most suitable for them.
Post-interaction analytics can quickly provide information such as, how often customers are being put on hold, which agents put customers on hold the most, and the emotion and sentiment of both the agent and customer throughout the interaction. These insights can be used to identify training gaps and improve the process by removing needless steps.
Real-time analytics, on the other hand, can act as a direct assistant to agents. For example, it can be used to determine whether a customer qualifies for a discount, whether the delivery is positive and informative, or whether compliance statements have been read correctly. This type of analytics is also useful for operations teams because it can identify when a compliance statement is read incorrectly or not at all, and live interventions can be made to steer the interaction back on course.
The real value of real-time analytics lies in its ability to provide solutions to actual problems in the contact centre and fix them there and then. For example, if an agent is struggling with a difficult customer, real-time analytics can provide suggestions on how to handle the situation. By providing real-time insights, agents can be more effective in their roles, leading to better customer experiences and increased productivity.
But that all comes at a cost. Very few organisations currently stream calls live (more likely they batch upload them at a later point), have the bandwidth and storage to handle the large amount of data produced with live streaming, which all puts pressure on the cost-base of a function that is already heavily scrutinised for spend.
Ultimately, the choice between real-time and post-interaction analytics depends on who is being helped. Post-interaction analytics are more suitable for operations teams, while real-time analytics are better suited for assisting agents. Whichever type of analytics is chosen, it is essential to ensure that it provides value to the contact centre and improves overall performance, and one method requires significantly more investment than the other.