That a large portion of one’s customer service call are outsourced to India or other exotic locales is old news. We’ve all been patched through to a phone bank half a planet away to speak with someone trained to suppress their native accent and make references to your local weather and high school sports rivalries. Some fast food chains even outsource your drive-up burger order to countries like India where eating cow is verboten to a substantial portion of population.
Lately, tax breaks and a surfeit of college-educated English speakers have attracted blue chip companies like IBM, Shell and Hershey to the Philippines, creating a customer service economy that, due to the 9-hour time difference from its American customers, operates predominately at night.
Companies like MyCyberTwin, however, are anticipating yet another shift in customer service outsourcing – one that won’t require a legion of nighthawks in Manila, nor pretending to be American – just human. Thanks to advances in artificial intelligence, avatars (or, “virtual agents,” to use industry parlance) can answer complex questions and use rational and logical thinking. Think “Spock in a Box.”
“By combining sophisticated ‘brain’ technology with state-of-the-art animation, MyCyberTwin brings a distinct and advanced virtual specialty to businesses,” explains CEO Liesl Capper. The secret sauce behind “brain technology” is the virtual agents’ ability to learn as they go, “allowing them to consistently get smarter and function at a higher level as time moves on,” Capper adds. NASA has recently implemented the technology suggesting a real-life HAL might not be far behind.
Chat bots have existed in various forms since the mid-60s. MIT’s Joseph Weizenbaum is credited with creating one of the first, ELIZA, a program that used a primitive form of natural language processing to simulate a real conversation with its interlocutor via text-based exchanges. Thousands of so-called “chatterbots” have spawned since with customer service implementations facilitating millions of monthly “conversations” (San Francisco-based VirtuOz claims12 million such interactions a month for clients in the Fortune 1000).
But can a virtual agent pass the Turing Test? Developed in 1950 by researcher Alan Turing, the test was originally devised to answer the question “Can machines think?” and uses natural language conversation with a human as its principle gauge. Though the test has been criticized by such heavy weights as philosopher John Searle for conflating rhetorical manipulation for cognition, the test remains something of a gold standard if only for proving the fallibility of an artificial intelligence’s human interlocutor. The goal of companies like MyCyberTwin isn’t to fool people into thinking their product is human but rather improve the customer experience but interfacing with them in a manner they’re most accustomed – like humans. As online insurer Esurance proclaims in its current ad campaign “People when you want them, Technology when you don’t.”
Ultimately, however, most consumers would prefer not to have to communicate with customer service at all, whether that be in Manila or on with HAL on some customer service odyssey. The virtual agent will surely learn this long before the companies who employ -- but then again, they can't hear you scream in virtual space.
