Despite being buzzed about for decades, artificial intelligence (AI) has finally hit the mainstream, with organizations incorporating it into everything from life-saving tools like IBM Watson for Oncology to Laundroid, the world's first automated laundry-folding robot.
Because of its wide range of uses and long-time depiction as a technology that could only exist in science fiction, most are unclear on what exactly AI is and how it can be used to benefit their day-to-day activities. This is especially true in call centers, where AI integrated into chatbots has taken on the role of the villain who is going to steal the jobs of customer service agents by solving customer queries—no matter how complex—quickly and autonomously.
But contrary to popular opinion, AI is not a silver bullet, but rather a means to an end. In fact, in many cases, it can be used to the advantage of call center agents, lessening workloads, improving efficiency, and helping identify trends to improve customer experiences.
There are two main types of call center AI: the kind that is customer-facing and the kind that is agent-facing. Customer-facing AI is what's most commonly associated with chatbots, characterized by a bot that responds to customer queries without help from the human agent. For most organizations, the goal of incorporating AI into the chatbot is not to replace human agents, but rather to help them become more efficient. According to a 2017 report by Aspect Software, 72 percent of customer experience experts and professionals agree that human agents have a stronger impact when AI chatbots handle routine work, and 79 percent say chatbots taking over complex requests sharpens the agents' skills.
To identify which queries can be solved via chatbot and which require a human touch, many organizations use chatbots at the beginning of every query, programming them to gather preliminary information, ask fact-finding questions, and alert agents when necessary, whether that's because a customer is frustrated and requires human intervention or simply because the query is too complicated for the bot.
Natural language processing (NLP) technology is a key enabler of this capability. With NLP, chatbots can understand customer intent. If, for example, the customer asks, "What does this laptop cost?" the chatbot understands that the customer is asking about price. NLP also enables chatbots to field clarifying questions, such as asking the color or size of a specific article of clothing a customer is looking to purchase, and execute transactions, such as booking a reservation or accepting payment. It also enables chatbots to respond to questions with images, videos, and URLs, which can be useful both in sharing marketing collateral or product instructions and for adding personality to the chatbots' responses by allowing them to respond with jokes or quips.
While customer-facing AI focuses on solving customer queries, either by managing the problem without human interaction or providing human agents with the information they need to solve it quickly, agent-facing AI focuses on diving deeper into the customer experience to enable agents and organizations to improve customer experience as a whole, rather than on a case-by-case basis. For many organizations, agent-facing AI involves capturing data from customer interactions and feeding it into analytics software that can identify trends, such as specific areas of dissatisfaction, allowing live agents to know that they always need to handle specific problems. These tools can also be used to identify market trends, tracking new customer expectations and providing customer experience managers with the information they need to train agents to anticipate those needs and increase customer retention. In specific query situations, agent-facing AI can access its database of queries and let an agent know if a customer has reached out before (especially if he's asking about the same issue), providing information on the person, past transactions, and history with the company. Conversely, the AI can also alert an agent if a customer is reaching out for the first time, marking the query as high priority to ensure the customer is provided with the best possible service to ensure satisfaction and loyalty. Organizations can also implement AI systems that can reference these call profiles and route queries to an agent in the correct department to cut down on transfer times.
But implementation of AI is not enough. For an AI system to be effective, it must be fully integrated with all call center software and fed sufficient data to allow it to be well-trained enough to respond to queries appropriately. Incorporating a knowledge management system creates a backbone for the AI, increasing the bot's ability to understand and learn and results in better customer service from both the bot itself and the agents using the system.
As AI gains popularity, organizations that fail to implement it will fall behind their competitors. Those that do implement it—and implement it correctly by incorporating the data required—will enable their agents to perform better, anticipating customer needs and solving queries in record time.
Jeff Epstein is vice president of product marketing and communications at Comm100. He's a B2B marketer with more than 20 years' experience creating messaging and content for sales enablement and demand generation. Prior to Comm100, he held roles in sales, product marketing, business development, and partner marketing at IBM, General Motors, Sophos, QuickMobile, and Allocadia.