January 31, 2018
In our quarterly blog series, Ask the Experts, we ask COPC Inc. experts questions about specific areas for improving operational performance in call centers, customer experience operations, vendor management organizations and procurement.
In this edition of Ask the Experts, we chatted with Kathleen Jezierski, COPC Inc.’s COO, and Ian Aitchison, CEO of COPC Inc.’s Asia Pacific (APAC) region. Both are well-versed in artificial intelligence (AI) and its practical application within the customer care industry.
For those not familiar, can you provide a quick overview of artificial intelligence (AI), chatbots and machine learning (ML)?
Kathleen:
AI is an area of science that enables machines to do things that would normally require human intelligence. An example of AI is the “intelligence” behind some types of chatbots, the kind that uses more than simple rules. While chatbots enable humans to interact with machines via an interface, behind the interface is a program designed to make complex, autonomous decisions and respond to incoming data.
That leads to Machine Learning which is a specific application of AI where a machine can not only be taught to do a specific task but improve at that task based on new input. The tasks can be anything from analyzing data in context to identifying objects or making predictions based on analysis of previous patterns.
Ian:
That’s right, Kathleen. When we use the term artificial intelligence, we are really talking about the concept of a machine being able to mimic the sort of cognitive functions that are thought of as “human”, such as problem solving and learning.
Machine learning does have a slightly different meaning, as the terminology really refers to a branch of artificial intelligence which is closely related to prediction-making through the use of data analysis, pattern recognition and advanced algorithms.
And as Kathleen mentioned, chatbots are specific programs which enable end-users to communicate with a machine via text or speech — sometimes through messaging services like WeChat or Facebook Messenger. Chatbots have evolved rapidly due to the improvements in AI as well as the growth in delivery channels, such as the messenger services and digital assistants like Siri and Alexa. Because of their interactive functionality, chatbots are of particular interest to those looking to improve the customer experience.
How are these technologies currently being deployed in the contact center?
Kathleen:
In today’s contact center environment, this technology is mostly being used in two different ways: to directly assist customers and to assist agents (AI-augmented customer service).
An example of a customer-facing AI solution might be an advanced online chatbot. More than an interactive FAQ, an advanced chatbot can assign meaning to what it reads or hears and use this incoming information to make decisions or recommendations — even alter its own behavior. Furthermore, some advanced chatbots can even learn from the outcomes of its actions.
AI-augmented customer service is similar, but a live customer service agent is layered in between the customer and the AI. In this case, the AI might provide the agent with suggestions based on the input from the customer. In these sorts of systems, the AI can often observe how agents react to its suggestions and learn to tailor future suggestions based on what does and doesn’t work.
What should brands consider as they explore this technology for their own use?
Ian:
I think companies have to be very careful about implementing chatbots just because it’s the hot, new technology. I have seen too many examples of badly designed chatbots. They are based on limited functionality and are built around too basic of a decision tree. Often times these chatbots are not AI-powered but are instead rules-based. They don’t do a very good job of accounting for the unexpected or solving problems that are the exception to the rule. They aren’t capable of going off script, and they can have a difficult time understanding something if it’s entirely new to them.
That said, AI shows more than promise — it’s already in use today, sometimes with great success. In order to share in that success, companies need to make sure they do two things before wading in. First, they need to truly understand the limitations of the technology, not only in general but also as it applies to their specific application of AI. Then, they need to figure out a way to use this information to temper the expectations of their customers. Too many times companies rush into the latest and greatest solution or technology, and all they manage to do is alienate their customers.
How important is it for brands to be learning, planning and implementing in this area?
Kathleen:
AI isn’t going to replace traditional channels, at least not anytime soon. But AI will enable the automation of ever more complex customer inquiries and sales, and this is of vital importance to the business. As has already happened in recent years with digital self-service, the inquiries reaching assisted channels will become even more complex. This will affect how centers are staffed and managed.
It’s true that some customers will continue to prefer traditional channels, but as automation (including AI-powered solutions) becomes more effective, and as our population ages, the percentage of consumers that strongly prefer live interaction will get smaller and smaller. The application of AI is an opportunity to reduce cost by moving transactions from assisted to unassisted channels. But the application of AI, especially when it’s done well and exceeds expectations, is also an opportunity to further strengthen your relationships with your customers. Every positive touch cements the relationship between brand and customer, even when it’s an unassisted touch. As the technology evolves, especially in areas like emotion detection and empathy, AI will be able to make these unassisted touches feel very personal.
———————————————————-
Do you have a topic you would like to see explored in the next edition of Ask the Experts? Submit your performance improvement questions or topic suggestions by emailing us at info@copc.com.