The irony is not lost on us that on this Valentine's Day we highlight an aspect of banking that has no heart. This afternoon kicks off an epic showdown between man vs. machine. For the next 3 days, the iconic show Jeopardy! will be hosting a $1mm special addition where it pits the show's biggest human winners, Ken Jennings and Brad Rutter, against Watson, a computing system from IBM. While computers have beaten man handily at computation and even strategy for years, Jeopardy! represents another level of challenge, as the computer now must understand popular culture, irony, puns and the subtle nuances of human language. While a human might figure out that the clue "Amen! After 43 years, our prayers are answered" refers to the New Orleans Saint's Super Bowl win, Watson flubbed the answer. However, after reprogramming, Watson is back. It is armed with 10 racks of Linux servers, 15 terabytes of RAM, 2,880 processors and 200mm pages of content (it is not allowed to access the Internet). Watson is now ready to hit the buzzer in less than 3 seconds.
When Deep Blue beat then reigning World Chess Champion, Garry Kasparov at chess back in 1997, it did so by brute computational force, analyzing hundreds of thousands of individual moves to determine the optimal path. Watson, on the other hand, operates closer to human intelligence, by processing different aspects of the question separately and then making a guess based on probabilities. Knowing that there is no "optimal" answer, Watson understands that the World is broken down into a spectrum of answers, some that are more "correct" than others. Watson is the next step of human evolution, as we combine both man and machine in a symbiotic relationship. We predict Watson will trounce its human competitors and in so doing, usher in the rise of machines (perhaps even similar to Terminator) and that expert systems will be the norm in our hospitals, law firms, factories and banks.
Banking for one is already starting to be more quantitative. Our BIGProfit relationship profitability system, for instance, holds more than a terabyte of data on banks that allows it to draw inferences from the industry and then use individual bank's data to determine the likelihood a customer will leave over a certain period of time. While current banking models are far from the power of Watson, it is important to note that unlike Deep Blue, Watson is comprised of off-the-shelf software that can be purchased today, so the future is closer than we think. We predict that it will be less than 10Ys before management routinely consults expert systems for a wide variety of questions, including deposit pricing, marketing, compliance and general management. Answering the question "What deposit product am I missing to best increase profitability?" is actually easier than the question Watson successfully tackled in practice rounds. Under the heading "Money Not From A Mint" for $1,000 - Watson correctly came up with "Line of Credit" to the answer - "An agreement with a bank that it will loan you a certain amount whenever you ask is called this." While lacking a heart, expert system's quantitative-based answers combined with management's current passion for customers, may be something shareholders grow to love in the near future.