AI can be magic. AI can be controversial. AI can be way off the mark in its answers to human queries.
But in contact centers, artificial intelligence isn't entertaining. It is essential. It would be natural for contact center agents to view AI as a threat that could automate them right out of a paycheck.
In live customer service, AI can handle many tasks. Those include agent assist, which plumbs knowledgebases for answers to customer questions via human prompts or automatically when listening to the conversation via speech recognition.
Another tool is the chatbot, which can solve some customer problems on their own but are most effective in teeing up a customer's problem for human agents and gathering case details while customers wait in the queue.
Empathy vs. efficiency
Contact centers can be stressful environments, where success is measured in metrics such as time-to-answer and minimal hold times. Efficiency is rewarded. AI that doesn't waste the customer or agent's time is the order of the day.
Is AI up to the task? This was the topic at an ICMI Contact Center Expo session late last October, led by Leslie O'Flahavan, founder of E-Write, a consultancy that specializes in helping customer service agents compose empathetic emails, chat messages and Tweets. She and other contact center leaders discussed the value of AI in a workplace where an agent's empathy is the key to customer satisfaction yet efficiency is what the bottom-line demands.
ChatGPTLanguage model, OpenAI
Bridging the gap between empathy and efficiency -- especially when we're positive AI is unable to generate genuine empathy -- would seem an impossible problem to solve, like a modern version of Schrödinger's cat. Vendors expect their AI tools to delight customers, but the customer relationship may die in the AI black box.
The answer, O'Flahavan said, possibly lies in the idea that empathy can be scripted by humans and served up to human agents by AI bots. The AI can locate empathetic responses for agents in the moment of a phone call or written customer conversation.
Such empathy prompts may not be necessary for veteran agents who are steeped in their company's culture, have years of practice and understand the nuances of each incoming caller's predicament. But with contact agent turnover pushing 45% annually, the less-experienced agents need all the help they can get to thrive in their jobs and preserve customer relationships.
Don't worry about your job -- yet
For the last five or so years, an open question has recurred at conferences and meetings, especially for contact centers: When will AI take my job?
Two trends make me believe it won't be soon. First, pandemic layoffs, the Great Resignation and the general switch to remote work has downsized contact centers. Contact center leaders are begging for better self-service chatbots and automations to take manual, low-level tasks -- such as package tracking, password changes and unlocking accounts -- that overwhelm the human agents who remain.
Secondly, AI always has and always will have its limits. One of the true sages in the customer experience industry is Rich Dorfman, vice president of customer experience for the 120-location Eastern Bank. When I asked him if cloud services such as AI sentiment analysis, deep analytics and workflow automation -- working in concert -- would ever replace the work he does to improve the bank's digital experience, he said no without hesitation, shaking his head.
In his mind, AI is good at finding and measuring things. But it takes humans to figure out what to do with the trends that AI uncovers. It also can't be trained to feel empathy. AI applications will change, improve, and do more work. But there will never be a substitute for that last mile of humanity.
"A digital-only bank is only going to get so many customers, because when I want to do a mortgage -- or I have a problem with fraud, and I'm angry, I'm scared, and I'm frustrated -- I want to talk to somebody who can talk me off the ledge, have a sense of helpfulness, and create a sense of caring and sharing," Dorfman said. "A machine can't do that."
When asked if bots would ever replace humans in customer service, ChatGPT itself is skeptical.
"It's possible that ChatGPT or other language models could be used in customer service. But it's unlikely that they will completely replace humans," the bot said. "While language models can understand and generate human-like text, they do not have the same level of understanding and empathy as a human customer service representative. Additionally, there are many tasks that a human customer service representative can do that a language model would not be able to, such as making decisions and solving complex problems."
Humans will harness the power of customer service AI tools to make their work easier, more accurate, and efficient. AI makes it so they don't have to reinvent the wheel every time a new call comes in. Even when an agent doesn't know the solution to a customer problem, AI can avail to them the entire corpus of answers their fellow agents have given, and zero in on the ones that works best.
In a thoughtful defense of the human agents that work the front lines of contact centers, O'Flahavan said any tool, including AI, that helps create the appearance of empathy -- genuine or not -- is what they need. It's a difficult job that often requires talking to angry people all day, every day -- or people who could become angry a few sentences into any given conversation.
"For what we're paying them hourly, I'm not sure we're entitled to our representatives' human empathy," O'Flahavan said. "They don't have to bring [to work] the part of them that loves their family, the part of them that they express in dance or painting. We don't deserve that. That is not what we're entitled to."
That, in the contact center, is how the next wave of smarter and faster AI tools will improve the lives of workers and not replace them. AI might be getting better -- frighteningly better, in the case of ChatGPT -- but it is nothing to fear. Instead, this technology will usher us into the next generation of customer service.
Don Fluckinger covers enterprise content management, CRM, marketing automation, e-commerce, customer service and enabling technologies for TechTarget Editorial.