When it comes to customer care, the best customer experience is no experience at all because services just work the way the customer expects—the first time and every time. So how can service providers progress toward this nirvana of customer support?
Building better customer care solutions starts with using connected intelligence to understand when and why a customer is experiencing an issue and how these issues can be most efficiently resolved. All of a service provider’s systems and processes must work together to create a great customer experience. Organizational and system silos make it hard to provide a holistic and real-time view of what’s happening with a customer and determine the next best action to take. The ability to connect domains, derive insights, and automate actions in the moments that matter is connected intelligence.
The road to fully autonomous care—where problems are predicted and resolved automatically, before the customer notices anything is wrong—lies in mining the vast trove of data that service providers collect, and learning valuable lessons from it using big data techniques.
As a first step, interactive, algorithm-driven software programs, called “bots,” use machine learning to interact with customers through conversation-based human interfaces. This makes customer care intuitive, always available with no waiting in queue and “zero touch” for both the customer and service provider. These interactive bots are already in service for common problems that drive large volumes of calls to the contact center, such as connecting to Wi-Fi, resetting forgotten passwords, or checking on the status of a scheduled technician appointment. Gartner predicts that by 2020, “twenty-five percent of customer service and support operations will integrate virtual customer assistant (VCA) or chatbot technology across engagement channels, up from less than two percent in 2017.”
These bots can be interfaced with a natural language virtual assistant, such as Amazon’s Alexa or Facebook’s Messenger. The key is to supplement the virtual assistant and its natural language capabilities with strong analytics to correctly identify the customer’s intent using sentiment and topic analysis.
If the task is straightforward enough, bots can be scripted relatively simply, but it is harder when tasks become complex. Fortunately, machine learning is already enabling systems to solve more complex issues. By analyzing customer support logs, self-learning, software-based predictive models have been developed that select the next best action most likely to solve the customer issue in the shortest amount of time. These models leverage the vast experiences of contact centers, field technicians and other personnel. By capturing and submitting the analysis of historical session information and context, it is possible to identify the remediation activities that are the most likely to be effective. These systems continue to learn with every future customer care interaction. Essential to always taking the next best action, workflows can no longer be static, but allow dynamic execution of steps.
The next step on the path to fully autonomous care is to feed this predictive modeling of support workflows into any kind of care scenario. Bots are now able to augment customer self-care, care agents in the contact center, and field agents by taking actions autonomously and only seeking human input when information is required to determine the next best action to troubleshoot the problem more accurately or more quickly.
Finally, this same kind of predictive modeling can be used to create a fully autonomous customer care capability. By analyzing network and device data and correlating it with trouble tickets and resolutions, signatures that predict customer-impacting issues can be created and linked to probable resolutions. These signatures become triggers for bots to take automatic actions to resolve the potential issue using the same next best action capability—all before the customer even notices an issue. For example, device firmware can be automatically updated to improve performance, Wi-Fi channels proactively changed to avoid interference, DSL links optimized before speed and stability suffer. At worst, customer care agents can be notified of potential issues, including remediation actions already taken by the bot, before taking a call for those cases where human interaction is still required.
In each of these stops on the path to fully autonomous customer care, bots help circumvent or alleviate the influx of unnecessary help desk calls while improving the customer experience and perception of the service.
In the customer service space, autonomous care will be a key differentiator. Using a machine-learning approach to preventative analytics and proactive customer care can ultimately lower support costs and sustain customer satisfaction. Preventative and proactive measures will enable service providers to successfully compete in an increasingly competitive digital services market.
For more information about Nokia’s autonomous customer care solution, feel free to visit our webpage.
Image: Nokia 2018