Rounding up business rules

Do you know where all of your company's business rules are?

Most enterprise users are surprised to discover how many important -- and not-so-important -- rules, regulations, policies and procedures are scattered all around the organization. For example, last year's marketing manual has guidelines for creating advertising campaigns; equations for calculating employees' health and retirement benefits are embedded in Cobol code; and best practices for writing software code reside only in the minds of senior developers, since no one has been asked to write them down.

In older, slower eras, this diffusion of policies and rules wasn't such a big problem. But business and IT executives find themselves under greater pressure than ever to adapt to rapid changes in the market and in government regulations -- as well as to operate at maximum efficiency. As a result, they are looking to round up these renegade rules and put them someplace they can be easily accessed, updated and applied to business processes. To do that, they're turning to business rules engines -- execution environments and repositories for business rules -- and management systems.

Catching ERRANT Claims

A case in point: The District of Columbia provides financial assistance to needy residents, some of whom also qualify for Medicaid or other federal programs. Recently, managers working for the district discovered that the local aid program was often getting the bill for services that should have been covered by federal programs. If an employee failed to catch such errors, it would be a costly misapplication of the rules.

To catch more of the bad claims and more quickly process legitimate ones, the district began developing its Automated Client Eligibility Determination System. The new system relies on ILOG Rules business rules engine to determine eligibility for D.C. and federal programs. It asks applicants a series of questions -- much like a TurboTax automated tax program does -- and then prints out completed applications for the programs for which they are qualified.

The ILOG engine, which is accessible to anyone with a Web browser, has a very high accuracy rating -- 99 percent, according to Donna Ramos-Johnson, associate director at Washington's Office of the Chief Technology Officer. That delivers better performance than the legacy system, which is an IBM mainframe running an Adabas/Natural database that was used internally for claims processing and financial transactions.

Ramos-Johnson says more federal programs will be added to the rules repository, which will eventually be used by the legacy system as well. "We expect to have the major federal programs online by September," she says.

Who Needs Them

Rules engines have been around since the early 1990s when companies such as Pegasystems, Fair Isaac and ILOG, sold them. They were typically used in rules-heavy industries such as finance and insurance. Over the past few years, however, many vendors have entered the market, and more companies are looking at rules engines as a way to gain greater flexibility in business operations.

"What's driving new interest in business rules is the need for business agility," says David Kelly, president of Upside Research. "Companies need to be able to create applications and business processes that can adapt rapidly to marketplace demands."

Rules engines provide this kind of flexibility by making it possible to edit the steps, or rules, of a business process. Traditionally, those steps have been coded into the application. But with a rules engine, they can be written in a natural-language authoring language and stored separately in a managed repository. Applications are then instructed to access the rules engine, and the rules themselves can be updated quickly by semitechnical users rather than programmers.

Also, notes Kelly, business rules systems can help companies prove compliance with government regulations by providing an audit trail of procedures and changes to those procedures.

Living with Legacy Apps

Legacy applications are one major reason organizations are turning to rules engines. When companies have many rules embedded in legacy code, moving them to a rules engine enables users to make changes without having to constantly rewrite code.

First American Field Services, which provides property inspection and maintenance services to banks, turned to rules management after it reached an impasse with its legacy system.

"It was so spider-webbed, there was custom code for each of our clients, and it was just so difficult to change," says Mark Davis, development manager for MIS at First American.

Three years ago, First American began developing a property inspection and maintenance system using Fair Isaac's Blaze Advisor rules engine. That application is linked to a DB2 database and Visual Basic .Net workflow engines that consult the rules engine to determine a course of action, such as what service to order. Rules are edited via an English-based authoring language and Fair Isaac's Visual Ruleflow Editor, with drag-and-drop icons for graphically creating business processes.

"It's very easy to make changes now," says Davis.

Brian Stucky, the "enterprise rule steward" at New York-based Freddie Mac, also credits business rules management with simplifying the process of changing rules. Managing policies became much easier after the federally chartered mortgage lender replaced a legacy system with an application tied to an ILOG JRules engine.

"We have a huge number of business rules. Before, to make a change, we'd have to get a mainframe guy to find the rule, make the change, retest the system, put it back into service," Stucky says. "It was such a lengthy procedure that we often waited until we had several changes to make. Now we can support rapid change in rules as needed."

Other companies are also using rules engines to improve operating efficiency. AMR, a medical transportation company, uses a rules engine to manage its fleet of vehicles more cost-effectively.

"Before, if someone needed transport to get an X-ray, we might send out the most expensive rig -- an advanced life-support system -- and transport them to the hospital at a high cost," explains Mark Kalevik, a software engineering manager at AMR. Now the company relies on CleverPath Aion Business Rules Expert from Computer Associates International to determine which type of vehicle to authorize and how quickly it must respond.

Driven by BPM and SOA

Interest in business process management (BPM) is also driving interest in business rules.

"Business rules engines are becoming an important part of other solutions, such as business process management," says Kelly, noting that it's common for BPM vendors to partner with rules engine providers.

Another complementary trend is the increasing use of Web services and service-oriented architectures. When building an SOA framework, organizations are adding a business rules layer to go along with the business logic, workflow and data layers.

Chicago-based Promissor, a provider of educational testing and licensing services, is developing just such an SOA. The company created a registration system that could be used remotely by on-site registrars with laptops or handheld devices for screening and registering test applicants. To make the system more accessible by handhelds in remote locations, Promissor built the application using Web services.

"We've rearchitected, with the rules engine as the cornerstone," says Robert Crouch, vice president of IT at Promissor. The company selected Haley Systems' HaleyRules engine and HaleyAuthority rules-authoring tool to create and manage the registration rules. "The Haley engine is light enough to load on a PDA, so we do not need Internet connectivity to operate," Crouch says.

Promissor preferred Haley's natural-language interface, which enabled business users to easily edit rules. It also liked Haley's small footprint, says Crouch.

Options for viewing and editing rules can be important. Users may want to work with rules via a decision table, a decision tree or some other format that they're familiar with.

Cesar Gomez, manager of systems operations and application development at Horizon Casualty Services, especially likes the visual features of the RulesPower product from RulesPower, which Horizon installed as part of a new bill-processing program last year.

"What impressed us was the visual diagramming of the workflows," Gomez says. "It's like an interactive Visio screen. It gave the business people the ability to visualize how the business rules flowed within the program."

Horizon's RulesPower-based bill-processing application has enabled the firm to reassign three of its six bill processors to handling exceptions -- nonstandard claims that require human scrutiny -- and to substantially reduce its backlog of claims. The use of a rules engine has even cut the cost of processing a claim by 30 percent, according to Gomez.

User-Friendliness

What matters in a rules management system, says Barnes, isn't the list of features; it's how user-friendly it is to nontechnical people. Most organizations buying rules engines today want their business managers to be able to create and edit their own rules.

"The real differences, and the real areas for improvement, have to do with usability," says Barnes. He suggests that businesses begin by evaluating how easy it is for users to formulate business rules with the product.

"The value proposition of a rules engine is the ability to manage business rules, and those rules should be defined by business people," Barnes says. "Unfortunately, many products are still too immature and too technical at this point."

Advise or Control

James Sinur, an analyst at Gartner, explains that rules engines are used either to control transactions and processes or to provide advice and analysis.

"About 50 percent of business rules engines are used in an advisory role: 'Should I do this or that?' The other 50 percent are used in business processes," he says.

According to Sinur, there are three categories of rules systems.

1. Simple rules externalization. This system allows an organization to express its rules in a standard format, house them in a repository, view them in decision trees or tables, and edit them as needed.

2. Inference engine. If the questions you need to put to a rules engine tend to be more sophisticated than simple "yes" or "no" equations, then you may need an inference engine, which uses probabilities and backward chaining through the rules to discover multiple possible solutions to the same end.

3. Behavioral learning. These advanced systems use case-based reasoning and are "trained" to recognize a variety of scenarios.

What You Need in a BRMS

Michael Barnes, vice president of technology research services at Gartner, says business rules management systems typically have the following features:

Business Rules Development

- Guidelines for identifying and documenting rules

- A graphical user interface for authoring and editing the rules

- Visualization of the process flows created by multiple rules

- Rule testing and debugging

- Integration with other development applications

- Rules-mining capability to harvest rules from legacy systems

Business Rules management

- A rules repository

- Support for role-based access by different users

- Reporting and querying capabilities

- Support for versioning

- Rule consistency checks to ensure accuracy and enable rules reuse

Start with Good Rules

A business rules engine is only as good as what's in it. And the first step of any business rules project should be to identify all of the rules in your organization, according to Ladd Bethune, senior technical consultant at Lambert Technical Services.

Once you've identified and extracted your existing rules, and before you transfer them into a business rules engine, you need to evaluate the quality of the rules, says Bethune. They may need to be edited or rewritten in order to make them sustainable for the long term.

According to Bethune, sustainable business rules have at least 12 characteristics. They should be adaptable, auditable, easily implemented, extensible, manageable, queryable, reusable, securable, testable, traceable, understandable and verifiable.

Join the newsletter!

Error: Please check your email address.

More about CA TechnologiesCornerstoneGartnerIBM AustraliaIlogInferencePegasystemsTechnology ResearchTransportationVIAVisio

Show Comments

Market Place