Behavioral science meets machine learning.
As trauma nurses will tell you, continued exposure to strong emotions depletes our ability to connect with other human beings.
Customer service representatives also manage strong emotions all day long, all week long. After hearing the same problems over and over, it is inevitable that CSRs may struggle to connect with callers.
It’s not sufficient to post pick-me-up reminders in the workspace. In no time, agents screen out the reminders—the sticky notes with smiley faces, or mirrors printed with “Smile!” hung near monitors.
Behavioral science tells us that we need corrections in the moment. Take an example from newer cars: safe driving alerts.
Many vehicles tell drivers they’re swerving out of their lanes by flashing a light in the cabin or displaying a dashboard icon. These driving alerts illustrate the use of behavioral corrections in the moment they’re needed.
Similarly, it’s possible to stray from conversational lanes. It’s possible—and, in stressful situations, likely—we’ll stray outside the boundaries necessary for a smooth conversation. More about that in a second.
Automotive technology that keeps us in our lanes illustrates a key difference between artificial intelligence and augmented intelligence. While AI-controlled, self-driving vehicles are getting the most publicity, the real advances are occurring when we use AI to augment our own intelligence and make us better drivers.
In other words, we can use AI to improve human performance. So if we can use it to help us drive, could it help us serve customers on the phone?
To Begin, Let’s Identify Our Conversational Lanes
Stressful conversations on the phone demand high degrees of emotional intelligence. Without eye contact and visual cues from our counterparts in a conversation, we have to rely more heavily on voice, which makes it harder to properly adjust our behaviors.
Customer service representatives are tasked with making emotional connections, on demand, with upset or frustrated people. If we could help stressed customer service representatives correct their behaviors—to stay in their conversational lanes—which behaviors would we assist?
At Cogito, we’ve identified a handful of key conversational trouble areas where we can augment the intelligence of customer service representatives. Avoid the following lane violations and callers (and agents) will be happier with their experiences:
- Interrupting and talking over callers as they attempt to explain their problems
- Going silent for too long, without acknowledging the caller
- Speaking too quickly – at a pace at which the caller can’t process what is being communicated
- Speaking with low energy that reflects distraction or disinterest
- Speaking continuously, not allowing the customer to engage in the conversation or acknowledge comprehension
- Ignoring emotional changes in the customer’s voice and not being empathetic to their situation.
These conversational trouble spots may seem like common sense. However, detecting them under duress, understanding their impact, and adjusting them at just the right time during complex emotional conversations can be challenging. If successful, the results are very powerful in practice.
So, if we were to apply machine learning to help address these challenges, the keys to successfully augment intelligence might look like this:
- Understand what is good and bad conversational behavior within context of a given situation
- Provide agents with behavioral notifications the moment that guidance is needed
- Make notifications intuitive, to the point, and actionable
- Consider and analyze both sides of an interaction —the caller and the agent—and the conversational dynamics amongst them
- Continuously learn and improve
This is the “what” of augmented intelligence, where behavioral science meets machine learning. How is it done? That’s the topic of our next blog post.
Download Your Free Ebook, The Cogito Guide to Emotional Intelligence