In my last article, I discussed knowledge management as being a critical component of business intelligence. In this article, I advocate another discipline which should be a central ingredient of business intelligence—namely, artificial intelligence (AI) [and a subset of AI called “discovery informatics” (DI)]. AI deals with supplementing human brain power with intelligent computer power, through the use of intelligent systems technologies like knowledge-based systems, neural networks, intelligent agents, genetic algorithms, case-based reasoning, and the like. Discovery informatics (DI) is a subset of artificial intelligence dealing with discovering relationships and patterns in large masses of data and text. DI involves such technologies as data mining, text mining, rule induction, self-organizing maps, and other related techniques.

Much of what is being espoused about business intelligence reminds one of the earlier conversations involving artificial intelligence. Some people say that AI should have been called IA—intelligence amplification. In this sense, AI should have been designed to assist the human and support the decision-maker versus trying to create a computer program to eventually replace the human. In the same way, there hasn’t been much intelligence behind business intelligence— business intelligence techniques have been used to primarily support the decision maker.

Some organizations, like CSIRO Australia (the Commonwealth Scientific & Industrial Research Organization), define business intelligence to include five key stages: data sourcing, data analysis, situation analysis, risk assessment, and decision support. This definition takes on an AI flavor in terms of extracting, synthesizing, filtering, and discovering information from multiple sources of data. However, most people haven’t defined business intelligence in the same manner. However, CSIRO has a refreshing view on business intelligence which includes AI.

How could AI techniques help business intelligence? Intelligent agents, for example, could be used in a number of ways to facilitate business intelligence. They could be used to help “push” lessons learned and best practices to the decision maker via integration with lessons learned systems. They could assist the business intelligence user by developing a dynamic profile of the user’s patterns and interests for better targeting information to the user. They could also be used as searching and filtering tools, as well as user profiling and classification aids. Expert systems technology could be applied to business intelligence in applying knowledge elicitation techniques to acquire lessons learned. They could also be used as on-line pools of expertise in rule or case-based systems. Through knowledge representation techniques used in the expert systems field, knowledge taxonomies and ontologies could also be better defined and developed for business intelligence application.