Although chatbots are just now gaining ground as the next “big thing” in customer service, a small number of them have actually been working behind the scenes for quite some time. SMS banking has been around since 1999, doctors send texts to remind us of our appointment, and Amazon buzzes our phone when our package is delivered. We’ve spent several years assessing chatbots’ ability to handle simple customer service tasks, and they have finally earned a spot in the major leagues. Now we need to truly challenge ourselves — and our chatbots — by integrating them into the epicenter of customer service: the contact center.
Contact centers have evolved quickly over the past 10 years, thanks to the rising popularity of digital communication channels. Phone calls still account for roughly 68 percent of inbound inquiries, while the other 32 percent comes from digital channels such as SMS, live chat, social media, and email. To adapt to the dynamic tendencies of mobile customers, progressive contact center platforms have become all-inclusive, eliminating entangled technology stacks and siloed communication channels. All inquiries, regardless of channel, are routed to a single queue and answered in the order they are received. Contact centers use unifying services to create a central hub from which support agents work. That hub is at its peak when communication channels, customer information, and interaction history live together in a single, uncomplicated interface. Chatbots need to be built directly into the solution in order to be an asset to agents, not a disruption.
The biggest challenge with chatbots has been curating a useful database of information from which they can work. It’s an incredibly tedious task that requires the creation of numerous user flows and continuous data monitoring and upkeep. However, developing a chatbot that lives and learns amongst the dynamic data of a contact center platform makes this challenge obsolete. The chatbot can be programmed to build and maintain its own database using the data within the platform, which is being continuously updated by human agents in real time.
The deep learning methodology is comparable to that of a search engine. The chatbot first develops an initial database from which to work by crawling any and all information available in the contact center platform, including customer files, past interactions, FAQ, product knowledge, and more. Any time the data is updated (for example, after a customer service agent closes out a support ticket), the information is crawled again, and the chatbot’s knowledge base stays updated. This method creates the ideal chatbot — self-sufficient, self-taught, and self-maintaining. It acts as an emissary or concierge to help humans make sense of the volumes of data we couldn’t process previously.
Let’s take a look at how this would function in a real-life situation. Let’s say a contact center agent just ended a live chat session with a customer who had a question regarding a warranty. The conversation would be logged within the contact center platform, which the chatbot crawls periodically, analyzing the initial request, the final answer, and the steps to resolution. The following day, when another customer initiates a live chat conversation and has a warranty question, the chatbot, having already “learned” about the warranty policy from the conversations it has crawled, has the knowledge to handle the inquiry from start to finish without ever involving a human agent. Without the reliance on humans to feed it information, the chatbot is able to mature on its own.
The ultimate goal of a chatbot isn’t to replace human agents, but to serve as an enhancement. Mundane and monotonous tasks should be handled by chatbots to allow human agents the ability to focus on more complex cases. The ultimate goal is efficiency. The inclusion of chatbots in the contact center should result in more first-call resolutions, reduced hold times, and higher customer satisfaction. A byproduct of this greater efficiency is agent satisfaction, which can have a huge impact on the quality of service agents provide. According to recent research, 90 percent of consumers stated than an agent’s perceived happiness affects their overall customer experience.
The Holy Grail scenario for chatbots within contact centers lies in creating a coherent bond between a human agent and a chatbot. When chatbots are deeply integrated into a contact center platform, they can become powerful sidekicks who handle multiple mundane tasks and allow their agent partner to give more complex issues adequate attention. Should a chatbot become stuck, the conversation can be immediately transitioned to their agent partner, along with all the customer account information and chatbot conversation history, thus enabling the agent to pick up right where the chatbot left off. Ideally, the customer wouldn’t even be aware that they were initially speaking with a bot, nor would they be aware that a human agent has taken over the conversation.
Unleashing chatbots into the contact center world is a sure fire way to make chatbots the customer service rock stars we all dream they will be. We will educate them with access to the same tools and resources used by human agents and give them the ability to monitor the humans themselves. With massive numbers of customer files, product information, and actual human interaction at their fingertips, chatbots will become independent learners. At that point, the focus shifts to simply refining the working relationship between chatbots and humans, bringing us one step closer to a perfect solution.
The audio problem: Learn how new cloud-based API solutions are solving imperfect, frustrating audio in video conferences. Access here