The IT world at any large company is a complex, heterogeneous collection of hardware and software obtained from a variety of vendors as a result of years of technology investment. Creating actionable intelligence from myriad information sources is a challenging exercise. In the past, business-intelligence software, designed to do such things as identify patterns in large data fields, create cost and profitability models, predict future results, optimize a set of variables to a desired outcome and monitor performance metrics, ran primarily as stand-alone applications, often with access to only geographic or functional silos of data. Operating in isolation, decisions based on this intelligence, while appropriate for a particular department, function or locale, might not be best for the enterprise as a whole. Also, the power of business-intelligence applications rested more in the hands of specialized programmers than the general employee population. That all changed with the March 2004, release of SAS 9 Intelligence Platform, from SAS Institute Inc., Cary, N.C. Three years in the making, the platform is a combination of servers, data storage and distribution applications and statistical programs that provides a unified foundation for business-intelligence investigations. The system integrates multiple data sources and connects heretofore disparate applications from SAS, other vendors, customers and suppliers into a coordinated hub-and-spoke architecture with the platform at its center. Not only does SAS 9 provide the gate to information across applications such as ERP, CRM, warehousing and accounts payable, it cleanses data from the distributed sources and compiles it in the way best suited for the analysis undertaken. Companies gain new insights from an enterprise-wide view into the behavior of customers, suppliers and internal operations via existing infrastructure. In the past, employees wishing to apply business analytics were dependent on specialists to create the statistical models that are the basis for these calculations, leading to bottlenecks and delays. But with SAS 9, algorithms created for one type of analysis are stored in the central server and are easily shared throughout the organization, via a common user environment. In addition, access to business-intelligence tools is catered to a user's skill level, with what the company calls interfaces "fit to task." For instance, a marketing executive reviewing sales figures in a certain region in an Excel spreadsheet can seamlessly call up a program to do a price/volume "what if" evaluation, or predict future sales based on past history. The executive also can tap other data sources, such as warranty claims, returns and technical support requirements to understand cost-of-sales and thus customer profitability, with an enterprise-wide view. The analytic tools are retrieved while still in the Excel environment, and there is no need to open a new application and navigate back and forth from the spreadsheet data. With its skill-based interfaces, non-SAS experts also can more easily write and execute their own highly customized business intelligence reports. SAS estimates that companies implementing the new platform will give up to 80% of their employees the ability to apply analytics in decision-making. The statistical programs running in the background of SAS 9 also contribute to the platform's power. Enhanced mathematics "smooth" the impact of data anomalies, while at the same time helping to uncover reasons for them. SAS 9 also brings new capability to marry unstructured text data and structured text or numeric data in creating intelligence. This can be especially valuable in combining structured data with free-text comments from customer service and technical support communications. For instance, a computer company is using this capability to minimize customer attrition. Knowing a customer's purchasing background and hearing specific key words in a complaint situation, the call-center representative is directed to specific response patterns and recovery strategies to avoid loss of the customer, based on analytics running in the background as data/text is logged in. As "negative" key-word trends build, manufacturing can be alerted to potential product issues.