The first race of the 2014 Formula One season, in Melbourne, Australia, marked the debut of a new hybrid engine. This revolutionary engine uses an advanced energy capture technology that provides the same horsepower and performance of nonhybrid race cars at two-thirds the rate of fuel consumption. While this technology is currently only available in F1 race cars, experts predict that this hybrid engine—and other advanced technologies introduced by F1—will be a standard feature in luxury cars within three years and most average cars within about five years or slightly longer.
This three-to-five year time frame is well understood by technology analysts and futurists. A cursory glance at some of the technologies from the Consumer Electronics Show (CES) in 2010 reveals the first drones, hybrid tablet/laptops, and 3-D and smart televisions, many of which are now widely available. At CES 2014, “wearables” was one of the biggest technology themes, along with 3-D printers (food, metal, body parts) and virtual reality gear—in particular, the Oculus Rift, which was recently purchased by Facebook. Experts expect many of these technologies to become ubiquitous by 2019.
Innovations like those we see in F1 race cars and at CES can give us a pretty good idea of where technology will be in three to five years, or what analysts and futurists call the “useful future.” Of course, the introduction or mere availability of a technology doesn’t mean it will be successful in the future; there are many other factors at play (societal, political, organizational), and most of these technologies will have no impact on customers or technical support. But some will, and, in fact, already are. In this article, we will explore the useful future of customer service and technical support, focusing on a number of key topics: electronic virtual assistants, Big Data, analytics, and cognitive computing.
EVA: A Conduit to the Future
What are electronic virtual assistants (EVAs)? In essence, an EVA is a virtual interface that serves as a data conduit. It’s a port through which Big Data and cognitive computing can be applied to the real world. And it’s much more common in today’s business and consumer marketplaces than most people realize. According to a recent report by Opus Research:
Dozens of companies have already entered the EVA world by supporting “virtual chat” and automated resources that provide human-like interactions through the chat window on an e-commerce website.
Whether they are embedded in mobile devices (like Siri or Nuance’s Nina Mobile), integrated into e-commerce websites as automated chat, or instantiated on an IVR platform to greet inbound callers, these human-like resources complement or augment efforts of live agents while providing highly personalized service to individual browsers, shoppers, or customers. They are playing an important role in defining the future of self-service and assisted self-service.
Apple and Google have been important players in the advancement of EVAs, through Siri and Google Now and Voice Actions, respectively. (Google will likely be a huge part of any future in this arena, as it has the technological and financial resources to “own the future.”) Less well known companies like Anboto Group, Artificial Solutions, and Creative Virtual are among many of the most important players supplying the building blocks of the EVA universe: natural language processing, virtual support agents, and “intelligent personalized customer experiences.”
However, beyond the conversational experience of making a technology/human interface feel like a human/human interface is the role EVAs can play in providing deeper customer experiences. Human communication is context dependent, and until recently, computers weren’t capable of reading a situation, identifying the contextual cues, and adapting the response. But thanks to improvements in machine learning, domain recognition, and rules-engines, it’s becoming much easier for computers to respond and interact like humans.
The potential benefits for customer service and technical support organizations are huge: imagine being able to divert customers away from human agents while still providing them with fulfilling experiences (i.e., the feeling that they are being listened to and understood). Given the innovations coming in the useful future, the need for customer service and technical support will continue to grow. This kind of technology/human interaction could, among other things, significantly reduce labor expenditures for frontline support staff. It could also change the way we collect and analyze information and use that data to provide better support for our customers.
The Customer: The Center of the Information Universe
The story of the father who found out his daughter was pregnant from Target is legendary. Target assigns every shopper an ID, which enables it to track debit and credit card purchases by date, time, and store location (and a number of other data points Target refuses to disclose).Based on your purchasing history and behaviors, Target can then respond with targeted marketing strategies.
In this young lady’s case, she was able to hide her pregnancy from her father, but not from Target: she had been purchasing vitamin and dietary supplements that are recommended for the first trimester. Target determined that this customer was probably pregnant and responded by sending the young lady circulars featuring products for pre- and postnatal child care. The father wasn’t happy, and he called his local store and took the manager to task. Sometime later, however, the father found out his daughter was indeed pregnant, and he called the store and apologized.
This is just one example of how we can use new and emerging technologies in sophisticated ways to harvest data and gain a deeper understanding of our customers’ needs and wants. This is what’s known as predictive analytics. It’s no surprise that the retail industry was an early leader in predictive analytics, but another industry at the forefront of predictive analytics is insurance.
Insurance companies, especially those that sell health and life insurance policies, have long used data on weight, age, and smoker/nonsmoker status to predict risk and set premiums. Now they’re leveraging more sophisticated techniques to gain a deeper understanding of their customers. In 2010, for example, Deloitte began exploring ways in which the life insurance industry could use the huge amounts of personal data being generated on the Internet to develop more-accurate risk profiles for potential and existing customers. Deloitte’s model assumes that lifestyle plays a key role in determining future health risks (which science has shown to be a fairly safe assumption). If a potential customer is active on social media, it doesn’t require much effort to find out what her hobbies are. Is she a member of a Facebook group for skydivers? Bungee jumpers? If so, that raises her level of risk. (Although, insurance companies insist that no final decisions on whether or not to provide coverage are based on analyses like this.)
Complex data mining is fast becoming a routine operation among cutting-edge companies. Acxiom, for example, collects data from across the web and uses it to determine people’s favorite networks, their political leanings, how socially active they are, what sports teams they support, etc. Alliance Data Systems and Experian are two other companies using some of the newest methods to build enhanced customer profiles that go far beyond traditional measurements. Emerging technologies will allow companies to automate and streamline the process, and any organization that isn’t looking at ways to leverage advanced technology to learn more about their customers is already behind the times.
Predictive analytics has great potential for customer service and technical support, particularly in the areas of organizational risk assessment, opportunity identification, and customer profiling. With more sophisticated data and analysis, support organizations could reduce cost per incident and increase customer satisfaction. This will require a much more sophisticated understanding of Big Data, and it will require support organizations to train or hire more data professionals.
The Rise of Big Data
Every few years, a new management or business trend comes along that experts predict will “change everything.” Sometimes it makes a big splash, but often it’s just a fad. It comes and goes without even disturbing the surface. But this is not the case with Big Data.
Big Data is all over the news, but it’s not a fad: it’s a social and technological revolution that will qualitatively change the world in a way that is as profound and far-reaching as the changes brought about by the Internet itself. One problem with the term “Big Data,” however, is that there’s no universally accepted and agreed-upon definition; data professionals are still debating what it actually is. But a reasonably safe definition is “any type of data that is not easily handled by traditional relational databases.” Relational databases contain structured data that is easy to record, easy to categorize, easy to analyze, and easy to use (if only in limited ways): customer name, date of a support incident, nature of an incident, etc.
Big Data, on the other hand, is largely unstructured data: videos, audio recordings, pictures, emails, chat room transcripts, support forum texts, etc. While there are no clear numbers as to how much unstructured data is being currently created, estimates put it at 80–90 percent of all data generated on the Internet. Companies are increasing their focus on finding ways to leverage this vast amount of unstructured data. How, for example, would you categorize a call that occurred during the resolution of an incident? You could log the call according to incident number, date, the time the call was received, the customer’s name, etc., but, in order to be useful, this data must be reviewed each time that customer reports a new incident. More importantly, a lot of beneficial information is being left unused, or at least that’s the contention of Big Data proponents.
The promise of Big Data is that we will be able to use heretofore unused (and unusable) information. Going back to our support example, imagine that something happened during that call that you later realize was critically important. Even with the most advanced EVA, finding that information will be very difficult without a great deal of effort (and probably expense). In the end, if you want or need it badly enough, you’ll probably have to sit down and listen to the entire call.
That recording is part of the Big Data universe, but it’s of relatively limited value as long as there’s no way to interpret it. The way this will be done in the future is with data analytics: the tools needed to make heads and tails out of the Big Data (i.e., unstructured data) universe.
The Tools of Big Data: Data Analytics
There are many tools for processing and deriving value from unstructured data, including A/B testing, data visualization, machine learning and intelligence, RFID, NFC (near field communications), natural language processing, and text analytics.
These last two—natural language processing and text analytics—are part of a larger group of tools that support something known as sentiment analysis. Sentiment analysis is defined as:
A linguistic analysis technique where a body of text is examined to characterize the tonality of the document. Though the method predates modern technological tools, the use of sentiment analysis has accelerated in recent years with the development of large-scale computational infrastructure that can analyze large unstructured textual datasets.
In the next three to five years, sentiment analysis could change everything in customer service and technical support.
The potential of sentiment analysis is staggering. Imagine being able to apply an automated semantic analytics tool to a chat transcript or a call recording and immediately know how happy that customer was (on a scale of 1–10) by the end of that interaction, using text and tone analysis—no customer satisfaction survey needed here! After only a few exchanges with a customer, you could generate a personality profile that will tell you the best way to deal with that individual in the future. Such tools could also be used in the opposite direction: your technicians’ past interactions could be used to create profiles that would allow you to match an individual customer with the technician most likely to have the most positive interaction with that customer.
Imagine if the system could warn a technician not to mention the Seahawks in a discussion with a Broncos fan because, based on that customer’s history, mentioning the Super Bowl loss had a negative effect on the customer’s mood. Or imagine if the system could provide you with a metaphor or explanation guaranteed to work for a given customer based on successful applications in past interactions. And although these examples have to do with text and speech, you can apply the same analytics to visual records. Consider, for example, an automated process that scans a video and determines the mood of each participant using advanced recognition tools that can analyze facial cues. This is the kind of qualitative understanding that Big Data promises to bring to the customer service and technical support industry.
Watson and the Emergence of Cognitive Computing
If the tools of data analytics don’t seem revolutionary enough, consider cognitive computing.
In February 2011, Ken Jennings and Brad Rutter—who hold the longest and second longest winning streaks in Jeopardy! history, respectively —went up against a special challenger in a two-game match. That challenger was IBM’s Watson supercomputer.
During those games, Watson did something that no computer had ever done before: it replicated the human thought process. If you presented Watson with a question about a medieval warrior who serves kings, it knew you were talking about a knight; ask him what you call the period after dark, and it knew you were asking for night. Parsing homophones based on context and deciphering plays on words are skills unique to the human brain—or they were. Presented with classic Jeopardy! questions, Watson was able to evaluate context and arrive at the correct answers—and quite effectively, too, since he beat both opponents handily.
The ultimate promise of cognitive computing, represented here by Watson, won’t be fulfilled in the useful future, but in the meantime, IBM is pursuing short-term applications of Watson’s AI capabilities: IBM recently announced the Ask Watson initiative, which is a much more sophisticated incarnation of an EVA than you’ll find anywhere else. It’s essentially an EVA that combines cognitive computing, Big Data, and data analytics.
In January 2014, IBM announced the creation of a new Watson business unit dedicated to the development and commercialization of cognitive computing innovations. This business unit includes Watson Discovery Advisor (to create advanced data-driven research tools), Watson Analytics (to explore Big Data insights through visual representations), and Watson Explorer (to help users uncover and share data-driven insights more easily). At the same time, IBM is investing heavily in initiatives to move Watson to the cloud (Watson currently runs on a massive server complex in a single location. This could be the first move toward something I often refer to as the Evernet: an all-encompassing (i.e., no standalone devices) all expansive (i.e., all types of technologies) web of interconnected global data that itself functions as a global, virtual cognitive computer. It’s the stuff of sci-fi, without the “fi.”
Many people are skeptical about the ultimate success of Watson and the Ask Watson initiative. After all, we’re still waiting for predicted and promised breakthroughs in artificial intelligence. However, IBM is pouring tremendous resources into its AI research, and if it’s at all possible to make this technology useful and ubiquitous, it will be an important part of the future.
In the realm of cognitive computing, IBM is joined by another juggernaut: Google, which recently purchased an AI startup called DeepMind. Google has a history of investing in “far out” companies that are high risk/high reward, with a low probability of success in the short term. But the fact that Google is investing in AI illustrates the fact that tremendous resources are being brought to bear in the arena of cognitive computing.
If cognitive computing succeeds at the level envisioned by IBM and Google, what might that mean for the customer service and technical support industry? Will there even be a customer service and technical support industry, at least in anything resembling its current incarnation, in the useful future? What about the distant future?
These questions are why its so important to consider not only the technology being implemented but also how individuals, organizations, and larger society feel about and respond to the technology. This, more than the capabilities of the technologies themselves, will likely determine the success of this brave new future.
But it’s also important to remember that the future isn’t something that happens to you. The future is shaped by determination and vision. Those who created our present (think Jobs, Gates, and Zuckerberg) didn’t wait for the future to happen to them. They created it.
In the next issue of SupportWorld, we will explore several scenarios that illustrate the future of customer service and technical support. Some of it will happen, some of it won’t—creating the narrative itself is an important step toward creating a successful future.
Keith Orndoff is a futurist who delivers high-impact workshops and keynotes on the near future (3-5 years). Keith has been a consulting futurist since 1997. He has written internal reports for, consulted with, and spoken to hundreds of clients, including NASA, General Motors, the American Society of Interior Designers, the American Bankers Association, the Kellogg Corporation, and many others. Keith has a MS in Studies of the Future from the University of Houston – Clear Lake, and he can be contacted by phone at 832.335.2031 or by email at KeithOrndoff@FutureImpactEducation.com.