For many people, discussing business-process-management is an academic exercise that doesn't lead to business results.

But Rosemary Baczewski, director of process and performance improvement at Horizon Blue Cross Blue Shield of N. J., summed up what it's all about at the BPM Conference in San Francisco this week when she said her team had identified a health care claims process that usually took 7.6 days to complete. In an experiment, two managers hand-carrying the same claim managed to push it past the right parties in only 45 minutes.

"This was a wakeup call to the company. The employees can't necessarily work harder. The process has to work harder," she told a group of conference attendees in a session on using Six Sigma, an international methodology for eliminating defects and improving quality to improve business processes.

Six Sigma methods seek to root out causes of defects in products and services, eliminating the need to correct defects and deal with customer complaints and thereby improving the efficiency of the business. Baczewski said her business-process improvement team at Horizon tries to teach business-process managers to avoid excessive studying and defining of a business-process problem. "When you have 60% of the data you need, evaluate it and go with your gut" on how best to improve it, she said.

The Horizon example was one of many cited during the first two days of the conference of how business processes can be improved even as day-to-day business goes on.

"Our challenge is to be able to change the business while running it. We have to change it while bettering last year's results," said Roger Burlton in an opening address Tuesday. He is founder of the Process Renewal Group, a business-process consulting firm, and conference chairman.

Many speakers acknowledged that the field of business-process management, despite keen interest among managers inside companies, got something of a bad name during the 1990s when it was presented as business-process reengineering. Trying to reengineer many company processes at the same time proved to be disruptive and, in many cases, counterproductive. Now the goal is to improve business processes and capitalize on potential new efficiencies as a business evolves.

"CIOs are in their third year of relentless cost pressure to deliver more results, while meeting new requirements like those of the Sarbanes-Oxley Act. Improving business processes has to include collecting feedback on changes and using that feedback to further improve the overall business," said Robert Farrell, president of Metastorm Inc., a business-process-management software supplier, in a keynote address.

"We need round-trip business-process management," he said, repeating a phrase coined by Gartner analyst Jim Sinur. Such an approach monitors and measures the results of business-process change, evaluates it, and tries to further improve the process with that information. In other words, the job isn't over just because the process was changed, he said.

Peter O'Donoghue, senior consultant at Splash Consulting Inc., said in a session called "Giving Your Processes a Brain" that rules engines should be tied into business-process management systems. With a rules engine, business managers can change the rules that guide a process rather than tinkering with the process itself.

IT staffs are used to abiding by restrictions on changing systems and they engage in extensive testing before implementing changes. An audience member asked O'Donoghue whether business analysts and managers will have the same discipline if they have the freedom to repeatedly change the rules that govern a business process.

Rashid Khan, CEO of BPM software supplier Ultimus Inc., proposed to the conference that business process reengineering took too much time in convening multiple business managers, defining the process, and then trying to cover every eventuality in an improved process. "Fifty percent of the project time would be spent in defining the process. Most companies don't have their processes defined," he noted.

He urged process managers to engage in "adaptive discovery," where process managers quickly model a process and deploy a process based on that model. If the process stalls, a process manager identifies the problem, fixes it, and adds that information to the process model. It then runs until it meets a new roadblock.

Kahn said adaptive discovery has been built into the Ultimus BPM Suite 7.0, where a process model can "learn" from the rules and new exceptions added to the model and guide the flow of a business process more successfully. A business process might run successfully for six months before encountering an exception that causes it to stall. Instead of trying to foresee every exception, Kahn says, process managers should say, "Come to me when it no longer works."