AI is the new BI: Hello, Analytics. Meet Watson.
Let’s face it. Modern business intelligence (BI) is not very intelligent. At least not of its own accord.
There have been many BI and analytics “solutions” over the years that have turned out to not be as successful as it was foretold in sales presentations. There can be many reasons why BI falls short. Most often, it can be traced to the fact that BI fails to address business problems with clarity. Clarity is when a clear path, answer or solution is made plainly obvious and enough people are on the same page to see the solution carried out. Unfortunately, BI, as we know it today, does not often provide enough clarity without help.
However, it is not BI’s fault. The reason for this lack of clarity is first and foremost a failure to communicate in terms that converges humans to share a view of the clearest, best solution. Humans, by nature, do not communicate with charts, graphs and data displays. Regardless of how aesthetic or well-designed a visual display of data is, human interpretation is still required. Some people will derive their own conclusions by creating a narrative in their mind. Others, for the sake of time or cognition, need the data explained to them.
Regardless, the understanding we arrive at (or are still confused about) requires natural language to bring clarity in our own minds. To bring complete clarity among a group of people requires more communicating, group interpreting, possible solution ranking and consensus building. BI cannot achieve this on its own. It is easy to notice when data interpretation breaks down, because the language jumps from specific data (such as numbers and figure) to abstract and ambiguous terms (such as “culture” and “customer sentiment”). The gap in between the data and human abstraction is a chasm that BI cannot possibly span on its own. A new level of business intelligence is required.
For those of you aware of popular culture may remember in 2011 a famous episode of the television game show, Jeopardy!, between two of the show’s winningest human champions and the IBM supercomputer, Watson. Watson not only won, but won big, compiling a winning total three times as large as its competitors. The intriguing thing about the match was Watson had to comply completely with the human rules of the game. It had to “listen” to Alex Trebek read the answer, “think” of a question, “ring” its buzzer and speak its response before the human contestants could chime in. It is interesting to note that Watson did not totally dominate the game until the end. It got several key questions wrong. However, the outcome was it won beating two of the game’s best human beings.
How would you like to have a member on your team smart enough to beat two Jeopardy! grand champions? Well, last week, in fact, IBM announced that Watson is now available for businesses to use as an analytics platform.
Before you get too far ahead conjuring memories of 1960s sci-fi, like Hal 9000 from 2001: Space Odyssey, in your mind, we should learn more about Watson and how it works. Watson is an artificially intelligent supercomputer. Yet even with its nearly 3,000 processors and 16 terabytes of RAM, you won’t even find it on the list of the world’s top 500 fastest supercomputers.
What makes Watson unique is its question answering capability that combines natural language processing, hypothesis generation and machine learning. These are the three features that differentiate Watson from the contemporary BI solution space. BI communicates with charts, graphs and data displays, while Watson communicates in natural human language. BI can be used to address a hypothesis, while Watson formulates and tests its own hypotheses. BI can be smart, but it does not learn. Watson never forgets how it answered a prior question and reuses that knowledge to answer new ones.
The bottom-line is Watson addresses the biggest hurdle that modern BI has to yet to decidedly surmount: coming up with the best solutions to complex questions on its own. Watson synthesizes the best possible solutions and ranks them, just like a group of humans would. It also considers the question and communicates the answer in natural human language, just like we would. Watson wraps its arms around complexity by boiling a tough question down to its key components, formulating a hypothesis, researching possible solutions from known solutions and prior experiments, and ranking the answers most likely to be the best—just like we humans would answer a multiple choice question we weren’t sure about.
It will be interesting to watch and see if the business world is ready to adopt and utilize the power of Watson. I predict the innovators and frontrunners in their spaces will see this as an opportunity for a competitive edge—to take their analytic efforts beyond where conventional BI could take them.
Disclaimer: This post was brought to you by IBM for Midsize Business and opinions are my own. To read more on this topic, visit IBM’s Midsize Insider. Dedicated to providing businesses with expertise, solutions and tools that are specific to small and midsized companies, the Midsize Business program provides businesses with the materials and knowledge they need to become engines of a smarter planet.