• by Laurie Petersen , November 21, 2016

In the heat of a customer service call, it’s possible to turn things around using the empathy and understanding that come by really listening. Artificial intelligence (AI) is a tool increasingly employed to support call center reps by doing just that. I spoke with Josh Feast, CEO and co-founder of real-time emotional intelligence technology provider Cogito, about the impact on the bottom line. Here’s an edited account of the conversation.

Q: What is real-time emotional intelligence? 

A: Emotional intelligence refers to one’s ability to understand their own emotions, the emotions of another person, and to then adjust one’s own behavior based on that information. Real-time emotional intelligence refers to the ability to do these things in the moment; for example, on a phone call between an agent and a customer.

Q: How can artificial intelligence assist with this?

A: For low-value, low-complexity service inquiries, AI can improve the accuracy of responses in self-service channels, detecting what your inquiry may pertain to in a website or app, and automatically presenting a resolution it deems most aligned to your inquiry. For sophisticated inquiries — which make up about 75% of the total — AI augments call-center agent skills to better interpret the customer’s needs and respond in a way that builds rapport, trust, and a deeper emotional connection.  AI helps agents become better attuned to customers’ perceptions on calls and helps better guide them to adjust their speaking style. This ultimately ensures a better call for the customer and increased job satisfaction for the agent.

Q: Isn't there a danger of becoming too robotic? How do you prevent that? 

A: That is a danger when applications focus only on the words someone is speaking or only on empowering self-service channels in place of humans. In these instances, companies actually risk losing the connection to their customers and becoming a commodity service provider, which has a damaging effect on loyalty.  Also, when companies choose to push self-service without offering a simple escalation to a human, they risk alienating their customers. This can leave customers with the impression that they don’t care enough to help them solve their problem. They push the burden onto the customer.

AI that is not based in behavioral science can force a poor deployment.  In that situation, a technology may assume it can find the best resolution by simply churning through enough data.  This often leads to false results and inaccurate applications. To be effective, AI has to be based on a theory that it can be used to enhance the application over time.

Done well, technology can augment human insights by providing behavioral guidance to call center representatives. This is done by analyzing the voices (customer and agent) in conversations, detecting subtle signals of frustration, tension, anxiety, low energy, and nudging agents to help them adjust their style to communicate with empathy and build better rapport with customers. It detects the richness of human behavior (not just words spoken) to ensure each conversation gives considerations to the individuals interacting and their specific situations.

Q: What are some concrete examples of customer service AI and analytics in action?

A: Humana spoke at Dreamforce recently about their experience applying AI in the customer service world. Because many of the calls they handle impact a person’s health, calls tend to be emotional and Humana uses AI to help agents recognize key moments in calls and respond in an empathetic way. Agents receive live alerts to adjust their communication. Little messages like “You know that you’re talking over the customer” or “You’re speaking really quickly right now. Get back to normal.” It’s extremely applicable to a broad base. 

This is also helping agent satisfaction because when the customer views the agent as someone invested in solving their problem, they treat the agent with respect and are far more likely to provide information and engage in conversations. It’s a technology about raising all boats. 

Some people are better at reading social signals than others. On any day, you may have 50 calls and you may have four difficult calls. You have to be cheery for all of them and as you get tired, you don’t read the signals as well. This allows for a consistent level of service, even in a difficult work day.  

Q: Are there any bottom-line numbers you can provide on customer experience and the bottom line?

A: From a bottom-line perspective, we see increased customer loyalty, increased agent satisfaction and reduced churn, and increases in efficiency such as reduced call times and increased first contact resolution. Humana’s stated results are quite compelling:

• 28% improvement in customer satisfaction

• 63% improvement in agent engagement

• 6% improvement in first contact resolution