However you measure business success, customer experience is a big factor. And in many cases, it’s the call center experience that directly affects the customer experience – and  in very measurable ways: customer satisfaction scores, churn, social media ratings and brand reputation.

And yet, on any given day, a significant portion of call center agents are new, burned out or simply not adept at discerning and responding appropriately to the caller’s emotional state in real-time. Such information would clue them in on whether to escalate a call, for instance, or when it’s best to veer from the standard script.

Cognitive computing systems can recognize the sentiment behind a caller’s words and phrases in real-time and guide call center agents to respond effectively. Converting irate callers to neutral or happy callers as quickly as possible increases the likelihood of customer retention. It also relieves agent stress that contributes to burnout and attrition.

Responding to Caller Emotion

Imagine a cognitive computing service that understands both content and emotion by combining audio-to-text conversion, natural language processing and auditory signatures. Based on these inputs, such a service could identify the caller’s emotional state, flashing a red light for negative, yellow for neutral, green for positive. At the same time, it could present a script that’s had a high rate of success for similar callers.

The goal is to elevate the caller from the red zone to the yellow or green zone as soon as possible. The conversation might go like this:


Taking Up Where Agent Training Leaves Off

Could you train agents to be as sensitive as the agent in the scenario above? Not reliably. Because of high agent turnover, continuous training is impractical and expensive. Further, some people are not as naturally attuned to caller emotions as others. Cultural differences between caller and agent (economic, social and geographic) compound the challenge. The caller, agent or both might have limited language proficiency, or may be preoccupied about a personal matter.

The fact is, cognitive systems are better suited than humans are to make complex, real-time decisions – call after call and day after day. By doing the urgent work, cognitive systems free up human agents to focus on the important work. In the case of contact centers, that’s first-call resolution.

Cognitive computing services for call centers also collect valuable information about the customer interaction that, when mined for meaning, yields insights that can improve the customer experience in the following ways:

  • Predict customer churn: If the system detects a high likelihood for defection, it can prompt proactive incentive offers. Even better, it can identify the factors causing the customer’s dissatisfaction and work to prevent those factors in future customer interactions. 
  • Deliver hyper-personalized experiences: Cognitive computing services can also consider historical data when guiding agents. For example, the system might map the caller ID to the customer profile – say, male, age 40 to 60, on the Eastern seaboard. Historical data would reveal the phrases most and least successful in moving this type of customer from the red to a yellow state.

Call centers could even create baseline voice profiles for regular customers. After a few words or phrases, the system could detect whether the caller sounded more positive or negative than usual, and then adjust the agent script accordingly. 

  • Avoid the need for the call: Cognitive science provides the most value in call centers by reducing call volume. The tools to accomplish this are diagnostic, predictive and prescriptive analytics: 
    • Use diagnostic analytics to identify triggers for dissatisfaction, such as a service being turned off, or a late product shipment.
    • Use predictive analytics to anticipate the customer response. Customers who believe the agent can’t help them often ask for a supervisor, for example. Having to ask deepens the irritation, especially if the agent initially resists.
    • Use prescriptive analytics to automatically take action. This could mean prompting the supervisor to intervene if the caller has remained in a negative state for a certain amount of time. 

The Call Center of the Future

To compete in the real-time business world, organizations need to invoke the help of cognitive computing systems. By processing what callers say and how they say it in real time, these systems can guide agents to de-escalate tense situations, resulting in higher customer retention, lower agent turnover and the insights to create a better customer experience. In the near future, cognitive computing-based customer service will increasingly be a make-or-break factor for succeeding in a fast-paced, competitive business environment.

To learn more, visit the AI & Analytics section of our website or our conversational AI practice.

Jerry Smith

Jerry Smith

Jerry A. Smith is Vice-President of Data Sciences at Cognizant. He is a practicing data scientist with a passion for realizing busi­ness... Read more