AI for customer experience: overhyped but ready for some roles
The hype surrounding artificial intelligence (AI) in today’s marketplace cannot be overstated. Most of the excitement is just that—hype and hot air, plus a whole lot of confusion about what AI actually encompasses.
Whether it’s 1) voice-driven assistants such as Alexa and Siri, 2) chatbots that engage with customers, 3) robotic process automation (RPA) that automates mundane tasks like “reading” e-mails and integrating applications, 4) recommendation engines that personalize your shopping experience, or 5) powerful, enigmatic data tools such as IBM’s Watson or Salesforce’s Einstein, the tantalizing possibility of using AI to drive great customer experiences is sucking up all the oxygen in technology, business and marketing discussions. (See When Machines Overtake Us, RTInsights .) Have we really reached a magical tipping point—like that futuristic world in Blade Runner–where the difference between what humans and robots can do is no longer discernible? The simple answer is: NO.
AI covers a big spectrum of many diverse technologies and product categories–not just one thing. It runs the gamut from self-learning tools already in use today to emerging technologies that will unfold between now and the next 20-30 years. What’s demonstrable in the lab is significantly more advanced than the AI tools a business should rationally consider for a customer-facing system. AI’s readiness for prime time depends on upon which specific technology we’re talking about. For sure, some AI tools (plus older technologies being reclassified and described by vendors as AI) can become key differentiators in creating great customer experiences and automating customer service processes. For example, chatbots, RPA, predictive analytics, natural language processing and facial recognition/biometrics are being combined today with BPM software, business rules and CRM to provide much greater customer intimacy and personalized experiences on websites, and also significantly improve operational efficiency in contact centers and customer support centers.
At Pegaworld this week, Dr. Rob Walker, VP of Decision Management and Analytics at Pegasystems, gave a fascinating presentation on the state of AI as it relates to customer experience. He kicked off with a classical music selection written by a robot and continued to tease the audience by showing four paintings (two by Rembrandt, one painting of Rembrandt and one painted by a robot in the style of Rembrandt.) Dr. Walker asked the audience to select the robot’s painting; after he eliminated two of the four options Dr. Walker concluded that maybe approximately 50% of the audience selected the robot’s painting. I was in the audience and it wasn’t easy to discern the difference except that, to me, the real Rembrandt had a certain glow to it that the robot’s painting lacked.
Dr. Walker then described the power of opaque AI to make decisions. (Opaque AI refers to the non-transparent use of machine-learning algorithms for decision-making.) Using Facebook likes and other data, Dr. Walker said opaque AI has the power, with almost 100% accuracy, to discern customer details, such as: gender, age, smoker, alcohol consumption, if parents were together at age 21, lesbian/gay, density of friendship network, extraversion, agreeableness, emotional stability, intelligence, satisfaction with life, and several other factors. Frankly, it sounds like a marketer’s nirvana and a private person’s nightmare. However, despite unbelievable accuracy with opaque AI, Dr. Walker said the risk of opaque AI tools getting something wrong in business operations is still far too high to consider using it. For the immediate term, he said that businesses should focus instead on transparent AI tools.
Here’s a great explanation of the need for transparent AI from Kris Hammond, an advisor to Computerworld:
“ . . . there is a growing realization that we cannot start deploying and using intelligent systems, machine learning solutions or cognitive computing platforms if their reasoning is opaque. We need to know what they are thinking. Without explanation, we are blindly trusting the output of systems that many of us don’t understand at an algorithmic level. While somewhat tolerable for systems that are identifying faces on Facebook, it is unacceptable for systems that are integrating highly sensitive and highly valuable business logic, goals and priorities into their reasoning.”
Alan Trefler, CEO and founder of Pegasystems, also spent time at Pegaworld debunking much of the hype about AI, while also stating that certain transparent AI tools are ready for—and are being deployed in—customer service operations in many businesses, such as banks, insurance companies, retail, telecoms and transportation. Specific technologies that Mr. Trefler said are ready to use include predictive analytics and next best action (software which recommends the best option to propose to the customer based on data inputs and learning) and chatbots (software that engages with customers by successfully imitating a human, while simultaneously keeping track of what is happening in each channel.) His recommendations were pragmatic and right on the money; if you want to successfully implement AI for customer service and customer facing activities, seek to:
- Integrate AI into end-to-end processes. Too many firms have created large numbers of disconnected, non-strategic, citizen-developed chat bots or RPA bots. These bots are unmanaged and unintegrated, bringing very little value to the business. Without a more disciplined approach that integrates bots into horizontal business processes, the organization has no control, no visibility and no way to track what is happening.
- Implement AI within context. Context is essential for AI tools, such as natural language processing, predictive analytics and bots. Context for customer service and customer-centric processes can be derived from many different data points, but to deliver business value, it ultimately must involve connecting a customer to a business outcome.
Undoubtedly, the AI space is exciting, more than a little bit scary, while also firing our imagination for a new way of living and engaging in human activities. It’s important to track the AI space now and determine which opportunities to capitalize on. However, it’s equally important to know what is achievable and what is further out in the dim, distant future– and not get the two confused.