A new generation of business intelligence applications could provide agencies with deeper insights.
Anyone who attended college in the last decade can tell you, doing research today is not quite the same as it used to be. In this age of the Internet, there's little need to shuffle through index cards or pore over dusty old volumes. Modern technologies provide an astounding amount of the world's collected knowledge with only a few clicks of a mouse.
Included in the mounds of data is information about our daily lives. Federal agencies are beginning to harness the search tools needed to sift through mountains of publicly available data that can help give them deeper insights into how to do their jobs.
Using software tools to analyze and make connections among seemingly unrelated information from raw data is called business intelligence. Agencies depend on business intelligence to support policy decisions, improve operations, and identify potential instances of waste, fraud and abuse. Agencies such as the Health and Human Services, Homeland Security and Justice departments, and the Internal Revenue Service, have found ways to tap search technologies to root out waste and fraud.
Powerful search engines are leading to advances in business intelligence among agencies. Few, however, will talk publicly about how they are using BI, saying they don't discuss operations that include compliance or investigations. But consultants and vendors offer some insight into how the applications can be used.
For example, business intelligence can uncover abuse in the $440 billion Medicare program, says Haywood Talcove, chief executive officer of LexisNexis Special Services. Medicare specialists estimate that up to 20 percent of the claims submitted for payment are not valid. Practitioners, as well as patients, perpetuate fraud, with doctors ordering unnecessary procedures, or billing the government for procedures they did not perform.
Business intelligence searches include scanning the fields in a database, so-called structured data. Most BI tools rely primarily on access to structured data, which is information stored in a database's identifiable fields such as name, address or dollar amount.
Structured data comprises only a small portion of the information that is available to agencies, however. Most data is in unstructured form such as text in documents, images, photos and geospatial data. Ramon C. Barquin, president of the business intelligence consulting firm Barquin International in Washington, estimates that
95 percent of the data created today is unstructured. "Think of the billions of blogs, YouTube postings and pictures," he says. "It's huge."
This information can yield powerful insights, and agencies that ignore the chance to mine this data are doing so "at their own peril," Barquin says.
Proprietary tools that search internal agency databases often can be useful in identifying fraud and waste, according to Andre Etherly, vice president of federal solutions at Keane Inc. in Boston. Implementing a business intelligence software tool that can search through an agency's database and find instances of irregularities can help it cut down on waste and fraud.
Advanced business intelligence applications use artificial intelligence to examine the transaction stored in an agency's database, Etherly says. By comparing valid transactions with fraudulent ones, the software can sense patterns and flag transgressions. "We're talking about the ability of these tools to find otherwise indiscernible patterns," he says. "You don't need to know what the pattern looks like, you just need historical data to tell what a fraud is and find similar patterns."
These tools are becoming more common and integrated with software applications as demand for search capability rises, according to Etherly. Business intelligence applications that allow users to search, report, analyze and present the data in a user-friendly form are in the most demand, he says.
Federal agencies and private sector companies will expect more features and better ways of tracking performance in the next generation of business intelligence applications, Barquin says. The complex problems agencies face require multi-dimensional solutions, which call for more sophisticated software tools that can visualize possible options, he says.
"Search is a blunt tool," Barquin says. "It gives you the first set of hits based on broad concepts. What you need is knowledge exploration." The first step, he says, is a simple search, followed by a more refined search that yields a summary of the results. "That's where you start being able to integrate, establish analytics and metrics," he says. "Now you're using this unstructured data with search capabilities to build a case."
Some agencies already are turning to advanced business intelligence applications to identify potential terrorists, Talcove says. The Treasury Department uses the tools to track suspicious funding sources, Etherly says. With business intelligence technology, the Transportation Security Administration can avoid profiling while identifying travelers going through security who raise red flags.
Business intelligence technologies soon could be applied to a range of practices, from evaluating the productivity of call centers to making sure the government is getting the best deal on its purchases by quickly comparing the price it pays to prices the commercial sector pays.