When IBM's Watson computer project stole headlines two years ago by beating the best available human at Jeopardy, experts wondered if IBM had a long-term plan in mind, or if putting the computer on a television quiz show was merely a marketing gambit designed to show off the company's technological prowess. The answer came recently with the announcement of the IBM Watson Engagement Advisor, a system designed to provide customer service responses through near-instantaneous big data analysis.

According to IBM, "270 billion customer service calls are handled annually, with roughly 50 percent unresolved, which means an increase in cost-per-escalated-call by three times. 61 percent of those calls could have been resolved with better access to information."

Descriptions of the new service are fairly vague, but maybe that is the point -- Watson is much more adept than most computers at finding specific answers to vague questions. According to IBM, Watson will "help companies make their interactions count by knowing, delivering, and learning what each customer wants -- in the context of their preferences and actions -- sometimes before even the customer knows it themselves." The service appears to fall into the general category of "Big Data," however, rather than analyzing the data in advanced for a finished report, Watson will organize the information into an intermediate, indexed state, then analyze on the fly based on its interpretation of the customer's question. Watson's formidable natural language capabilities will contribute to what IBM hopes will be a seamless and efficient customer interaction.

Since winning a Jeopardy, Watson has gotten smaller and faster, with a 240 percent improvement in system performance and a 75 percent reduction in physical size. IBM says the system can now run on a single Linux-based Power 750 server.

At least for now, the Watson Engagement Advisor seems intended to provide services for clients that support customers who have a need for drawing highly specific information from very large data sets, such as information services for the banking and consumer marketing industries. The service will be delivered in a variety of formats, including via HTML, online chat, and mobile device.