Detecting Problematic Conversations in a Call Centre

Nearly 3/4 of consumers switch providers after poor contact centre experience

2
Poor Contact Centre Experience (1)
Data & Analytics

Published: March 12, 2021

Anwesha Roy - UC Today

Anwesha Roy

A bad conversion or a negative customer experience can cause irreparable damage to your brand. Research suggests that nearly 3/4 of consumers will switch providers after a poor contact centre experience, costing companies approximately $168 billion per year. For this reason, it is vital to monitor conversations in real-time and nip poor experiences in the bud by:  

  • Equipping agents with the information they need to solve customer queries
  • Intervening in real-time or escalating to a supervisor
  • Switching scripts and tonality to suit the customer’s preferences
  • Following up on a negative experience to reassure the customer and prevent churn

How do You Detect Bad/Problematic Conversations? 

The most common approach for identifying bad conversations in a contact centre, before they “go south,” is through live monitoring. A supervisor or the team leader listens to ongoing conversations in real-time (usually discreetly) to identify a change in customer attitude, negative words/phrases, and other indicators of a looming problem. 

There are also technology-enabled methods suggested to spot potential problem calls in a contact centre.  

For example, experts from Tata Consultancy services proposed a visual method that identifies a structure associated with what you’d call a “normal call.” If a conversation veers from this structure, it is automatically flagged as potentially problematic, and the team leader or supervisor is alerted.  

Another research paper suggests that bad/problematic conversations are detected based on hesitation, reaction, Interruption and cumulative voice activity. This allows you to spot poor CX independent of the language (and its specific words or phrases) being used.  

Getting Started with Call Monitoring and Interaction Improvement Technology  

The most popular approach to detecting problematic conversations is by using digital tools such as:  

  • Real-time speech analytics – Real-time speech analytics uses artificial intelligence and natural language processing to convert raw speech data into structured keywords and phrases. These are then matched with a prebuilt keyword database to understand positive and negative experiences. Apart from intervening in a bad conversation, speech analytics also helps to find the right moment to discuss offers/promotions and new products
  • Call centre whisper mode – Whisper mode is a telephonic technology that lets you join an ongoing call without having to initiate a conference call or making the caller aware of another person’s presence. You can stay in the background and discreetly convey information, tips, and arguments, or simply let the agent know that you are there in case assistance is needed
  • Real-time agent-assist – Real-time agent assist combines speech analytics with knowledge management to identify problematic conversations, analyse the type of information needed, and guide agents with the information and tips in real-time. A customised pop-up or email alert appears in front of the agent, telling them how they should pivot to prevent the call from turning into a negative experience
  • Barge into calls – Some tools let you barge into an ongoing conversation so that you can take over from the agent, reassuring customers and de-escalating the situation

Apart from this, recording all calls and a strong QA practice will help to better equip agents to handle problematic conversations independently in the future.  

 

 

Agent AssistArtificial IntelligenceCRMWorkforce Management
Featured

Share This Post