Trends in Analytics: The Potential of Autonomous Analytical Agents
Independent agents such as event-driven architectures and self-learning machines consume information about us and formulate their own hypotheses. This evolution of technology has already begun.
by Christopher Sowa, January 2013
Much has been written about how we can tame big data with big data appliances that make it easier to sequence and analyze data. However, little has been written about the potential for the convergence of big data technologies and other technologies to create autonomous analytical agents such as event-driven architectures and self-learning machines. Not only could these agents help us to consume data—one day, they could learn to consume information about us and formulate their own hypotheses.
If this concept seems far off in the future, think again. This evolution of technologies has already started. It is now possible to leverage information discovery tools to detect relationships between seemingly unrelated events. For example, government agencies are using these discovery tools to combine unstructured data from social sites on the web with structured intelligence data to predict future terrorist attacks. Financial services companies are combining structured and unstructured data using analytical discovery tools to perform root cause analyses on customer satisfaction—a process that used to take days and now only takes seconds. Retailers are rapidly analyzing vast categories of unstructured data to determine such things as the right sequence of products to present and the most fruitful sources of retail traffic. The potential for these smart discovery tools is also being seen in healthcare, where better understanding of patient treatment can mean life or death. The technology can not only look for unexpected trends in data—it will also send alerts to users when interesting information is discovered. What used to take teams of PhD students years can now be discovered in just seconds.
In 2013, as business users become savvier to the potential of these smart information exploration and intelligence analytics, we will see more organizations finally unlocking the value of information that is currently available but unorganized and too vast to be easily utilized. For the IT department, the ability to monetize the value of big data from unexpected big insights will have profound impacts. Companies will need to start to reevaluate what type of data they collect and how long it is stored, prioritize data sets in new ways, and based on new uses of data, change which data is deemed mission critical.
Likewise, business users and researchers will no longer be taking months to dream up new hypotheses around a handful of potential significant variables. Instead, with the aid of analytical discovery tools, they will be able to explore vast unstructured data rapidly and will be tasked with prioritizing data sets and testing hypotheses on ways to change relationships within the data. For some companies, processes such as customer satisfaction analysis, competitive threats, and service performance analysis could be condensed from months down to days. Additionally, business decision-makers will have new, near-time information to make key business decisions in ways that could scarcely be imagined just a few years ago.
Technologies related to artificial intelligence coming out of universities such as MIT could one day take these devices from being amazing analytical tools to being self-seeing analytical agents. These agents will sift through information and will likely find trends and learn about new relationships between the data. They will also start to infer which relationships and trends are likely to come next. While some in the data/research community may be frightened of this vision of intelligent agents replacing tasks that today are done by researchers and data scientists, I believe that working in collaboration with these new autonomous agents is likely to create exciting new frontiers in information creation and learning.
Christopher J. Sowa is vice president of Oracle Insight and co-head of Oracle’s Global Business Intelligence and Exalytics Strategy Pillar.