Many years ago when I was early in my career, I worked for one of the leading market research companies in the San Francisco Bay Area. Shortly after I joined the company it was acquired by DRI, a small subsidiary of McGraw-Hill. After the acquisition, a DRI employee was transferred into our offices to serve as a liaison between corporate and their new acquisition. He was a very bright guy with an exceptionally quirky sense of humor. Among the various things I learned from him, the most notable was a comment he once made: “You don’t have to know everything, but you do have to know how to find everything”.
The first iteration of what our liaison was describing, in a sense, became Google and other search engines. In theory, at least, there is a collection of all relevant information somewhere in the cloud and all you have to do is type in a few words to gain access to it. In short, you don’t have to know everything, but you at least have the opportunity to find everything. There continue to be shortcomings in modern search engines driven by incomplete information, intentional biases from prioritizing some information sources over others, the desire of search engine companies to generate revenue from search, and user limitations in not being to adequately describe exactly what they’re looking seeking. Moreover, as useful as search engines are, they don’t give us the answer we’re looking for – instead, they give us an enormous number of answers that might – or might not – be right. This results in search engines being extraordinarily useful tools, but not really a panacea to finding what we need to know.
What would be useful, then, is a way to do two things: a) have access to every bit of knowledge and information that is possible to have on a subject, and b) get the answer that is most likely to be right in as short a time as possible. That, in essence, is what IBM Watson does. Using natural language processing, “reasoning” capabilities, and voluminous amounts of data, Watson sifts through enormous amounts of data in a manner somewhat akin to a search engine, but it does so using natural language inputs. More importantly, Watson is focused on delivering the answer that is most likely to be correct. It’s not always right, of course, but it has demonstrated the ability to be mostly right – for example, on Jeopardy and in a contest with various members of the US House of Representatives.
Why is Watson important? Simply because it can receive inputs using natural language and process vast quantities of information to come up with an answer in a way that humans might if they had the capacity to sift through hundreds or thousands of terabytes of data in a very short amount of time. There are numerous potential applications in a wide variety of fields like medicine and law, among others. In the communications and collaboration realm, Watson could be used for things like analyzing who in a company is most likely to commit fraud by asking who is being abused verbally by their managers and correlating this with employee sentiment expressed in social media, email, text messages and the like.
In short, Watson could be enormously useful in providing direction for a wide variety of business activities like investigations, early case assessments, eDiscovery, fraud detection and mediation, and host of related types of efforts. While we will never know everything, Watson will help us get closer to being able to find everything.