Business intelligence might be about helping everyone in a company see the big picture.
But sometimes technology works best when it helps eliminate seemingly small hitches before they turn into big problems.
The technology behind business rules has its origins in the need to solve large and complex problems.
But it has found its niche in preventing seemingly trivial glitches that may end up costing companies large sums of money.
The first business rules engines developed from computer science research into expert systems. In the 1960s and 1970s, these were designed to solve the big problems of the day, from analysing seismic survey data for oil exploration to monitoring the “health” of spacecraft.
Early expert systems research relied on asking computers to answer a standard set of questions, far more quickly than a human being ever could, for mission-critical applications.
The applications for today’s business rules engines are rather more prosaic, but they share the expert systems’ ability to process multiple rules simultaneously in seconds.
But commercial and open-source business rules engines are also based around natural language, so non-programmers can write, read and adapt the rules as business needs change.
Some applications for business rules are complex, such as assessing whether to allow a derivatives trade in the financial markets or applying underwriting criteria to a mortgage application.
Others are deceptively simple. It might seem obvious that a booking clerk in an airline cargo office should not accept an order to carry a horse, if the aircraft on the route has no space for a horsebox. But the complexity of running an international airline means that sometimes such mistakes are made.
The cost – not to mention reputational damage – that such mistakes could cause prompted KLM Cargo (now part of Air France KLM) to create a business rules system called the Special Cargo Queue Processor.
The system, developed in-house, allows users to define the appropriate business rules for handling items such as perishables, valuables and animals; KLM is one of the largest international carriers for the bloodstock industry.
US brewer Coors uses business rules to ensure that the right number of cases of beer are loaded into its delivery trucks, while the Australian rail operator Railcorp uses a business rules engine to validate its timetables before they are published.
In France, SNCF Fret, the rail operator’s freight arm, has gone further and used a business rules engine as the basis for a new billing system.
The system gives sales people, rather than IT, the ability to modify prices. The system has cut billing costs by 20 per cent and also gives customers more consistent pricing.
Automating repetitive processes – such as booking a horse on to an aeroplane – is one driver for business rules, but such filters could be written in conventional programming languages such as C++, Cobol or Java.
Business rules systems, however, are designed so that anyone could read the rules and understand the business logic that the software relies on.
This makes it easier to check that the rules are correct, but once the initial infrastructure is in place, it also makes it much easier, cheaper and quicker to update or revise the rules.
“Agility is the primary reason for using business rules,” says Mark Layden, vice-president at vendor Fair Isaac, which provides business rules systems mostly to the financial services industry.
It is not so much to get humans out of the equation, but to ensure decisions happen consistently and accountably. Fair Isaac customer Discover Financial Services, for example, uses business rules for real-time checking of credit card transactions.
Touring Assurances, a Belgian insurance company, uses business rules to assess and price insurance quotes. The system consists of two parts: a Siebel CRM system that supports call centre staff and provides an online quotation system, and a business rules engine that encapsulates the company’s underwriting criteria and calculates premiums.
Previously, Touring Assurances used an older version of Siebel coupled with tariff and underwriting rules written in Visual Basic and stored on individual employees’ PCs.
The system was hard to update, prone to errors, and had no support for web-based offers. “If the Visual Basic rules were not deployed correctly, the agent would be quoting out of date tariffs,” explains Nathalie Servranckx, Touring Assurance’s IT manager.
Moving to business rules has made it easier to update pricing and introduce new products, helping the insurer stay competitive, she says.
“About 40 per cent of business rules systems are used in financial services, but it is very generic: it can be used in real-time manufacturing or anywhere else where you want to automate,” explains Pierre Haren, CEO of ILOG, the business rules system vendor used by SNCF Fret and Touring Assurances.
“It is still early days but awareness of the technology is growing rapidly.”
