Only 37% are utilising advanced analytics effectively
Despite the clear value advanced analytics could bring to contact centre organisations, research suggests that only 37% are utilising it effectively. But those who have applied advanced analytics including its real-time applications were able to reduce average handling time by 40%, boost conversion by 50%, and improve self-service by 5-20%. One of the ways in which real-time analytics unlocks these benefits is by preventing script non-compliance.
In most contact centre scenarios (particularly outbound), agents are equipped with a script that will guide the conversation in the right direction, making sure that thorny areas are avoided while the essential conversation points are mentioned. Importantly, contact centre scripts aren’t 100% structured end-to-end with no room for flexibility.
They are more like a checklist of dos and don’ts that a contact centre agent must adhere to when interacting with customers. Telephony-based centres and live chat support both require script adherence.
In this context, script non-compliance can be defined as instances where agents have deviated from the preset interaction script, either mentioning an element that was to be avoided, forgetting to mention a necessary element or following a script not suited to the interaction at hand. While there is always a little wiggle room around script adherence, non-compliance can pose a variety of risks:
For these reasons, it is advisable that contact centres use real-time analytics to monitor script adherence during live conversations and flag non-compliance immediately.
The solution uses a combination of speech analytics, real-time data processing, data visualisation, and alert delivery systems to monitor ongoing interactions and flag non-compliance, reporting the instance to a supervisor.
Speech analytics is used to compare the ongoing conversation (audio speech) against predefined script rules (text), often applying AI technology for this purpose. The data is streamed and processed in real-time so that it can be visualised in the supervisor interface and an alert is sent automatically. Let’s say the agent was to inform the customer that the call would be recorded within the first 45 seconds of call initiation. Real-time analytics map the conversation against the predefined script for that slot to immediately flag non-compliance (i.e., the failure to mention the call recording disclaimer).
The non-compliance alert should be sent to the agent as well as their supervisor, giving the agent an opportunity to self-correct. The supervisor can step in if there are too many alerts in sequence, preventing the call from getting derailed before the interaction can end.