Next Generation Rules and the Business Brain
Shaku Selvakumar 060001XT47 email@example.com | | Tags:  ibmpartners esb prolifics business_rules ibmimpact impact_comes_to_you expertintsys bpm analytics soa puresystems
0 Comments | 17,978 Visits
Guest post by Ryan Trollip
Ryan Trollip is Prolifics’ Decision Management Practice Director. Ryan is an experienced solutions architect and implementation lead with a strong background in business driven and improvement focused solutions with an emphasis on Decision Management. Follow Ryan on Twitter @ryan_trollip
I’ve learned some important lessons from my misadventures in buying an old sail boat. Firstly that you’ll spend the vast majority of your sailing time fixing the old boat and secondly, that you should use the right tool for the job, that is if you can find it. As usual I’ve learned this the hard way, for example trying to fix a broken rudder while playing chicken with a cruise ship. Don’t panic! Find those spare bolts and no, you can’t just tie it together in rough seas.
Over the years that I have spent consulting in business rules systems, a challenge that continually presents itself is the general lack of understanding of where rules and events management systems belong in the myriad of specialized technologies. The challenge is getting the client to use the right tool for the job at hand. There is certainly enough confusion created by the various overlapping technologies and vendor specific flavors of rules systems that cloud the picture for those new to the application of rules to a business problem. I will attempt here to separate the operational use of rules and events systems from these complementary technologies to allow for a better understanding of its application. In doing so we’ll look at rules in conjunction or versus expert systems, analytics, PBM, ESB, SOA and optimization.
Let’s step back and get a handle on what a rules system is. An analogy I like to use in describing the typical application of a rules management system is how the human brain operationalizes repetitive tasks. Let’s use driving as an example. When you first learned to drive it took a heavy amount of concentration, your conscience mind was working in overdrive creating and implementing all the rules on the fly. But once you have enough experience these rules start to become subconscious (operationalized). Ever caught yourself driving half way to work when you were supposed to get milk from the store on a Sunday? The implementation of rules is similar in concept in a business where rules management systems operationalize predetermined decisions made by the business, whether the source of these rules is predictive or prescriptive in nature. The primary value of a rules system is simply in the externalization, management and rapid change of these operationalized rules.
Rules, Rules Everywhere
Adding to the confusion is the presence of rules engines and tools in other complementary systems such as BPM, ESB and analytics applications. Again it’s a case of selecting the right tool for the job. Yes you can operationalize business rules in many of these tools but you lose a good amount of the benefits attributed to a rules management system. For example you may implement general data mapping rules in an ESB and that’s perfectly acceptable but if you were to embed some business rules in ESB, some in your BPM and some in an analytics application, you then have the problem of being tied to each of these development cycles and they do not offer the best of breed tools to manage your business rules. In addition, you would not be able to manage your business rules centrally and independently using a single comprehensive set of management tools designed specifically for the job.
Rules and BPMLooking more closely at BPM and its typical interaction with business rules, it is helpful to think of BPM as a way to implement long running process level flows visually. The tasks in a loan applications flow may call a business rules system to answer specific questions based on information gathered. For example: “Does this borrower meet the minimum qualifications for a loan product given credit and other supporting information?” The process may go on to have a loan officer request additional information or recommend an alternative product. SOA or use of an ESB allows for the centralization of rule sets or decision points that can be called from business process management applications or other systems.
Rules, an Expert System?A business rules management system is not an expert system. Although an algorithm, specifically the RETE algorithm, is typically present in a rules engine this does not make the system intelligent. Yes this algorithm allows for forward chaining, and more efficient execution of rules but this does not make it a goal driven, artificial intelligence or expert system. The conditions and actions for rules are largely pre-determined by the business and executed in a predictable order. It is not suited to optimization problems or designed to handle constraint based problems or a mathematical solver for linear programming, etc. Yes, this is not as sexy as AI, imperfect reasoning, fuzzy logic etc. but certainly practical and useful in a business context.
Rules and Analytics
Understanding the source of rules helps to form an idea of what decision points may be implemented and how analytics would play a part in feeding a rules system. I like to put rules into two categories, predictive and prescriptive. Prescriptive rules come from sources such as regulatory requirements imposed on the business, general business guidelines, product specific design outcomes, marketing campaign design etc. An example; “If the age of the customer is less than 18 then classify the customer as minor”. Predictive rules, in this categorization would typically result from analytics. I like to think of the predictive analytics source as the R&D department for a rules system. Segmentation is a common use of analytics in understanding for example the customer base and predicting the propensity for certain behaviors. For example “if the net worth of a customer is more than 500,000 then designate the customer as level 5”. In this example the level of the customer may have been determined by analytics based on past behaviors or based on surveys or other information gathering exercises. This classification may well be refined using other factors for example geographic location to arrive at certain final outcomes like offering a certain level of service, determining an APR, product eligibility or targeting a certain mix of products to the customer.
In keeping with the driving analogy, the traffic laws we obey and general courtesy toward other drivers would be the operationalized prescriptive rules. The environmental learning, adjustments and predictive assumptions made by the conscience mind would be analytics, for example adjusting to driving conditions like rain or noticing that someone is driving erratically and making the assumption that they are inebriated and putting into play avoidance rules. These predictive rules can be operationalized in a business rules system as well as hooked up to an analytics system for real time updates of variables based on changes in the population or environment in question.
The point, if I have to get one across, is that understanding and applying the correct tool will save you down the road and may save you from getting run over by a cruise ship. Rules allow for operationalizing business policies, regardless of where they are derived, the business rules management system allows for centralizing, externalizing, and managing rapid change of business policies independent of the typical IT development lifecycle, which in many cases is typically way too lengthy for most business change needs.
Later this year, I will be speaking at the Business Rules Forum in Ft. Lauderdale, FL. During my session, “Beyond Business Rules – An Agile Enterprise,” I will focus on enterprise decision management and take attendees through strategic and tactical approaches to decisioning that allow for sustained agility and growth. I will discuss what can be done in the organization at an enterprise-level, as well as at a solution-level, to better leverage decisioning capabilities and drive flexibility.
For more information about Prolifics BPM solutions, visit www.prolifics.com/bpm.htm.
Prolifics will be partnering with IBM across the United States at Impact Comes to You events.
Follow @ibmimpact for real time news and join the conversation #ibmimpact