Automated Emotional Response in Customer Service
I love my Typinator. It’s faster and far more accurate and consistent than I am on my own. In fact, since I started working tech support in March 2015, my Typinator has saved me an estimated 41 hours of typing. 41 hours! That’s equal to an entire week of vacation, all thanks to this simple, text automation tool.
When I think about how much time my Typinator has saved me, I sometimes find myself wondering: Why do we have a job?
More specifically, why isn’t a piece of software doing our jobs? After all, a machine would be faster, and a lot of the tasks involved in our work can feel robotic: tag this, categorize that, copy, paste, verify information...
Then again, other tasks, no matter how subtle, are far beyond the capabilities of computers. Because, when you're working in Customer Service, you are working with people.
People are irrational and often don’t even know how to express what it is they want or need. Machines simply can’t handle that kind of ambiguity (at least, not yet). And, when you are dealing with people, you’re going to have to deal with emotions.
Working with emotionally charged customers can prove especially challenging and, at times, very draining, which seems obvious but it was only in the 1980’s that customer service became the subject of study. It was then, in 1983 , that a sociologist name Arlie Hochschild coined the term “emotional labor.”
Emotional labor is a requirement of a job that you display certain emotions towards customers or others.
Hochschild found that customer service agents had three ways of dealing with emotional labor:
- The first is surface acting, what Hochschild called the “painted on smile.” The kind of inauthentic but expected behavior of a flight attendant laughing at a passenger's bad joke (1)
- The next method she called Deep Acting, what we might call empathy, when customer service agents really try to put themselves in the shoes of the customer to understand their point of view.
- The third is real emotion, which is spontaneous and authentic, but not always appropriate. So, on the one hand it’s lovely if you really feel like helping someone, but when you are feeling an inappropriate emotion (such as annoyance) you have to actively suppress it and turn to surface acting.
Another study that followed Horchild found that while surface acting requires less effort, it results in higher stress and is linked to greater burnout and job dissatisfaction. (2)
Deep Acting, on the other hand, though it requires more emotional intelligence, time and mental energy, it actually results in less stress and lower rates of burnout. They also found that it can reinforce the identify of the customer service agent and increase their job satisfaction.
So, let’s look at these two methods as they apply to modern customer care at a technology company, starting with Surface acting.
Let’s say you are working tickets one day, but you just... don't feel like doing it. Maybe you had a bad morning, or ate a bad burrito, either way, you aren’t feeling much like helping anyone. But you’re not going to tell the customer that, instead, you’ll say something like "I'm happy to help!"
How about deep acting? When I started this research, I was only working in email, and since I was reading these studies, I figured that deep acting is what I should be aspiring for if I wanted to prevent burnout. And yet, I found myself asking, Do I have time to empathize?
Let’s say I get an emotionally charged ticket with lots of back and forth, but I know that if I want to keep a good pace, I only have about ten minutes to send the ticket. Do I really have time to close my eyes and think about a time in my life that might help me understand my customer’s anger over their social media account feed not automatically updating every two minutes?
Then there is a bigger question, which is whether or not it is even possible to really empathize in a digital setting.
For starters, eye contact and all those nonverbal cues that are so important to understanding meaning are completely removed.
A group of scientists conducted a study to see how well humans can infer meaning and emotion in email.
What they found was that we get it completely wrong as much as half the time.
What’s more, 90% of the time, the person reading the email believes they correctly understood the meaning and intended tone. (3)
And that’s just in email - what about Live Chat? If it’s been proven difficult to correctly know the emotion and meaning in a single email, how about two consecutive live chats with separate individuals? It’s no wonder we sometimes feel overwhelmed.
In fact, in another study on empathy, scientists found that when we have to make particularly quick decisions, our ability for empathy is dramatically reduced. (4)
So here we are... at a rift between the expectation of what the best customer service should look like and the realities for the customer service agent. While here come from a place of wanting to help our customers, the fact is that..
...while some amount of empathy is always expected in our interactions, it simply is not always possible. And so the results can be dishonest.
Dishonesty. It’s certainly not something we like to talk about, but I would say that - in some ways - it is part of what we do.
But then, when I say dishonesty, I think of a certain kind of dishonesty: the dishonesty of theater. When you see a play, you realize that the actors on stage are, well, actors, but you willingly suspend your disbelief during the experience. And aren’t we all actors here, playing the part of patient and positive product experts who are always happy to help?
Customer service is a kind of theater - when you think about it. The server at the diner calling you darling, the airline hostess laughing at a passenger’s bad joke, all actors playing a part. And if customer service interactions are a kind theater, then they can be scripted, right?
Now for those of you saying to yourself “no! Don’t script replies to our customers!” - consider for a moment the findings of those research studies we looked at earlier:
With that in mind, let’s go back to that picture of the opposing realities of customer service, because I say there is a third option here - let's automate.
I’m not the only one dreaming of a way to apply the efficiency and consistency of software to human emotions. There are many, many laboratories, companies, and individuals striving to teach computers to do such a thing.
For instance, AT&T Laboratories, who developed an algorithm to identify and label emotionally charged emails into categories and automatically assign them to the appropriate agent. They also looked at more general negative words and phrases to assign those emails a higher priority.
We are already tagging and categorizing emails, but wouldn’t it be neat to see the average time it takes to craft a reply to an angry customer? Or to have a warning in place for an agent that the customer may be angry?
Companies are already using software to explore and handle human emotion. We already know how much time our Typinators save us, so why not teach our Typinators to feel?
Against the advice of some of my peers and supervisors, I began writing emotional Typinators using some of the best practices for customer service routed in applied psychology. These include positive positioning and advocacy.
The Typinators act like a shield, allowing me to reply to what might be a heated comment from a customer quickly, without becoming invested, so I can focus on what I need to do to resolve their issue. They also offer a kind catharsis - since customers never see the abbreviation I use, I get to type what I might like to say in the moment but can't. The customer simply sees a lovely, empathetic response that is quickly and easily customized to their unique situation.
So, why not use the tools we have and the research available to feel smarter? By allowing ourselves to occasionally step back from the task of emotional labor, we can save time, save our energy, feel great about our jobs, and focus on the problems that need fixing, leaving the emotions to therapists.
- Hochschild, A. (2012). The managed heart commercialization of human feeling. Berkeley: University of California Press.
- *Grandey, Alicia (2003) Surface Acting and Deep Acting As Determined
of Emotional Exhaustion and Peer-Related Service Delivery
- Kruger, Justin & Epley, Nicholas. Egocentrism Over E-Mail: Can We Communicate as Well as We Think? Journal of Personality and Social Psychology 2005, Vol. 89, No. 6, 925–936
- Giorgia Silani, Claus Lamm, Christian C. Ruff, Tania Singer. Right Supramarginal Gyrus Is Crucial to Overcome Emotional Egocentricity Bias in Social Judgements The Journal of Neuroscience, Online 25 September 2013
- Narendra Gupta, Mazin Gilbert, and Giuseppe Di Fabbrizio Emotion Detection in Email Customer Care, Computational Intelligence, Volume 59, Number 000, 2010