FICO’s Blaze Advisor Brings SOA Precepts to Rules-Based “Decision Services” 
FICO’s Blaze Advisor business rules management system (BRMS) sports a new visualization feature to improve business visibility and agility. IDN speaks with FICO about Blaze Advisor’s Decision Graph and other SOA-like approaches to help IT create a rules-based “decision services” architecture.
FICO’s Blaze Advisor 7.0 business rules management system (BRMS) sports a new visualization feature to improve business visibility and agility. Decision Graph makes it easier for IT and business to work with complex decision trees, which can span 10,000 or more pieces of business logic and other elements.
The Blaze Advisor Decision Graph is among the latest updates to FICO’s Blaze BRMS family that look to help IT deliver to business users a powerful rules-based “decision service” architecture that can improve the speed, predictability and profitability of business decisions, Don Griest, FICO’s senior director of product management told IDN.
FICO’s “decision services” approach maximizes the value of BPM, business rules, events, predictive analytics and data feeds and components to aid in real-time intelligence, he added.
The Blaze Advisor Decision Graph presents to business users a visual metaphor for grappling with complex decision trees, Griest said, and makes such complex rule sets much easier users to follow, manage and optimize. "As strategies get more sophisticated, business logic can expand exponentially. FICO Blaze Advisor 7 helps users understand the big picture and easily 'double click' into any segment for investigation,” Griest added. This feature will help any business manage complex strategies, he said.
Don Griest
Senior director of product management, FICO
Specifically, while Decision Graph views are easier to follow for all stakeholders, the views provided are not dumbed-down representations, Griest added. In fact, under the covers, Decision Graphs are powered by FICO-developed algorithms, graph theory and other technologies that can shrink the size and complexity of the original decision trees – all without taking away any of the decision logic of the original view.
“With Decision Graphs, we can present to users the minimum number of decision nodes needed to represent the exact same logic used in the more complicated view,” Griest noted. “This less noisy view of decisions allows users to make more rapid, effective and safe updates.”
Decision Graphs also enable better “what if” strategy development through dynamic analysis to assess the impact of changes by seeing how sample data flows through the logic of the tree, he added. A companion add-on product allows users to test their strategies with simulations.
Another key technology powering Decision Graphs is FICO’s SOA-inspired “decision services” architecture. Notably for IT architects, this approach to infusing intelligence leverages many of SOA’s most core and widely-embraced principals – loose-coupling, reuse, easy integration with outside resources and proper “granularity” of services.
This last point, granularity, is especially important, Griest said – especially when comparing BPM platforms with the capabilities of business rules engine.
“When you look at the drivers for BPM, many of them are all about business agility,” Griest told IDN. “But real business agility only comes when you go beyond the [business] process level because processes don’t really change that often – not nearly as often as business decisions do.” Or, as frequently as events and rules that trigger those decisions, he added.
“We feel the business architect and SOA integration professionals are looking both for better ways to help their business users make better business decisions, and make them smarter and with more agility or ability to update and change them based on business conditions,” Griest said. “For us, one key way to do that is with the development and deployment of ‘decision services,’ which use SOA for decisions in much the same way you would use SOA for integration and web services.”
In FICO’s approach, these “decision services” can be created from virtually any application, website or data set – even legacy applications without apparent APIs. This information can be converted into items that feed a certain decision when put together with processes, rules and analytics and best practice templates, he added.
As a result of this “decision services” approach, decisions can be quickly and easily aligned with key business goals, such as profitability or customer satisfaction, Griest said. It also means business users can make more accurate decision much faster, because the decisions are backed by so much more information and insight, including macro trends from historical data, predictive analysis and quantifiable business events such as a customer’s past activities.
“In today’s business environment, we feel you’ll need to track and model your customer reactions to your offer and your decisions, because different actions will cause different reactions from the same customer profile,” Griest said. Decision Graph empowers business users to get a better handle on the impact of such decisions, and provides IT a roadmap to the types of data and events business users need to make better decisions, he added.
To facilitate easy data imports and integration, Blaze Advisor is based on an open architecture that allows easy integration with a wide variety of databases, XML documents, Java objects, .NET/COM objects, and COBOL copybooks.
Other notable Blaze Advisor 7 features include:
Rich IDE for development, authoring and testing: Multiple methods can be used for rule creation and management (decision trees, scorecards, decision tables, formula builder, graphical decision flows and customized templates).
Rules repository for better collaboration: The Blaze enterprise rule repository enables devs to work in teams, as well as share and reuse rules, rule sets, rule flows and object models. Teams can also compare projects (with the same repository or different ones).
User-enabled ability to add, update, and manage unlimited rules: Rules and groups of rules can be defined and maintained in easy-to-use constructs (decision tables, decision trees, etc.).
Ability to monitor “decision performance” for KPIs, business outcomes: User-defined events will monitor the business performance of rules / rule sets. This approach provides a SOA-like building block approach to strategy orchestration and champion/challenger strategies.
Simulation capabilities to test for results, outcomes: The add-on Blaze Decision Simulator module can leverage historical data to analyze the potential business impact of rules – prior to moving them into production. It uses wizards, pre-packaged templates and reports to make it easy for business users to flexibly configure and run simulations for their specific needs.
One customer, Samsung Card, used Blaze Advisor to better control risks in its credit card business while increasing card loan volume. It also was able to offer better customized offers to creditworthy customers, improving profitability without taking on undue default risks, Griest said.
One analyst said FICO’s Decision Graph approach to simplifying rules and business logic will make it easier and more useful for business to employ them effectively. “Better visualization of business logic can provide a huge uplift for companies that are looking for ways to improve business decisions,” said Jim Sinur, a vice president at Gartner Research specializing in business rules management systems.
The FICO approach also sets a tone for how SOA strategies will drive the direction for decision management solutions and architectures in this decade, according to Griest. “Our approach delivers SOA for decisions and allow business users to tap into data and events and immediately know more about the decisions they make, how they impact the business and the customers,” he said. “To make the same decision in every situation or for every customer is just not the way to be competitive today.”
