Liberty Regional Agency Markets

Liberty Regional Agency Markets is a business unit of Boston-based Liberty Mutual Group.  Regional Agency Markets (RAM) is a group of regional property and casualty companies that distribute their products and services exclusively through independent agents and brokers.  Concentrating on small commercial and personal lines markets, RAM is represented throughout the United States by a variety of locally branded names.^[1]  Each RAM company is led by its own local management team, and focuses on fulfilling the product and service needs of the independent agents in its respective region.  RAM companies have a combined direct written premium of more than $3 billion.

Figure 1. Liberty Regional Agency Markets.

One way in which RAM grows as an organization is through external growth strategies.  With its diverse insurance operations requiring regional support by RAM IT, coupled with increased reliance on the web and internet based technologies for our independent agents and internal staff, this external growth activity creates a strong need for a service-oriented architecture (SOA).

Meanwhile, RAM's need to reduce cost created a number of opportunities to use rules technology to automate key features of its insurance operations.  In order to leverage the integration capabilities of a SOA and the business intelligence strengths of rules technology, Liberty RAM created a framework to support the deployment of rules-based applications as standards-based, reusable, independent services.  Services that are built on this framework can be used by service consumers on any platform, opening a myriad of opportunities to improve insurance operations regardless of the platform on which the activities are initiated.

Enterprise Rules Services

By combining a rules application framework with a service-based architecture, Liberty Regional Agency Markets is able to create and deploy rules-based applications as standards-based, reusable, independent services.  To the consumers of an enterprise rules service, these services don't appear any different from the other services provided through the service-oriented architecture.

Since the SOA creates an abstraction layer between the service consumers and service providers, enterprise rules services are technology agnostic and platform independent.

Figure 2. Service-Oriented Architecture (SOA).

Why Use A Rules Application Framework?

To support the development of rules-based services, RAM Enterprise Architecture has created the ERS Framework.  The ERS Framework provides tools, libraries, and design patterns to support integration with the SOA, error handling, rule management, and business domain model development.

All Liberty RAM-built rules applications are created using this common application framework.  By building an application framework for applications that use rules technology, RAM hoped to achieve the following key goals:

• Facilitate a repeatable approach to creating rules applications.
• Reduce the start up costs associated with new projects using rules technology.
• Improve application architecture by providing a best of breed approach.
• Standardize the integration approach across a diverse technology landscape.

Achieving these goals would have a significant positive impact on rules project development team productivity by allowing them to focus their efforts on implementing well-designed, carefully-planned business rule services.

Features of Liberty RAM's Rules Application Framework

The ERS Framework is a full-featured skeleton for rules-based applications.  By providing the technical structure needed by rules application development projects, project teams can minimize the time that is spent building technical infrastructure to support the execution of codified business rules.  Instead, rules project development teams can focus on defining the business domain objects and rule hierarchies required to effectively codify business knowledge.

The ERS framework provides application teams with a standardized, repeatable approach to building rule-based services that reduces the technology learning curve and application development start-up effort by providing pre-built solutions in several areas, including the following:

• Integration with the SOA
• Support for XML-based messaging
• Standardized interface for enterprise rules services
• Application-customizable control logic
• High-level rule management
• Error handling

By providing these standard solutions in the ERS framework, RAM application teams that build rule-based applications have more resources that can focus on the most important aspect of any rule-based application:  the business rules themselves.  Encoded business rules are intended to contain the logic needed to support or automate a decision-making process and should be as readable as possible.

Business rules created as part of an ERS application are not limited by the ERS framework.  They can be simple or complex, as required by the business, and any of the inferencing techniques provided by the rules engine may be used as needed.

Simplified Application Development

The ERS framework provides a standardized application architecture that reduces the need for new project teams to create custom solutions for common application tasks.  This allows project teams to focus on the critical business objectives of the project:  automating a business decision-making process using rules technology.

ERS includes a number of features that simplify the design and development of rules.  One of these features is called high-level rule management (HLRM), which creates a simple system for organizing business rules based on key data values specific to the rules application being developed.  For example, most business rules in insurance are time-bound, meaning that they are only applicable during a certain period of time (often specified by legislation).  Other rules may only be applicable in certain states or for certain lines of business.  As application teams encounter situation like these, they are able to define high-level rule management parameters used by the ERS framework to manage which rules are considered during any given interaction with the rules application.  This greatly simplifies the readability of the programmed business rule.  The following example illustrates the simplicity that this feature brings to rules development:

Additionally, ERS includes a collection of design patterns created specifically for rules-based applications, as well as template implementations of each design pattern that application teams can use as the basis for their rules-based application.

Components of a Liberty RAM Rules Application

Every rules application built using the ERS framework is divided into four distinct domains that are held together by the inferencing capabilities of the rules technology:

• The Rule Agent
• The Business Domain Object Model
• The Rules Domain Model
• The Reference Data Model

Figure 3. Components of a Liberty RAM Rules Application.

The rule agent encapsulates the technical infrastructure required to support a rules-based application, including:

• Control logic
• Error handling
• Population of business models and reference data from service XML
• High level rule management
• Integration interface

Rules developers configure the rule agent provided by the ERS framework to meet the needs of their project.

The business domain object model encapsulates all of the objects that the rules operate over and the objects that are created as part of the actions of the rules.  Rules developers rely heavily on business experts to create a business domain model that is an accurate logical view of the business data.

Rules developers use the rules domain model to create a hierarchy for organizing rules into logical groups based on the business objectives for rule reusability and maintainability.  The hierarchies created in the rules domain control the circumstances under which a group of rules will be applied to the business domain and enable rules management features of the ERS Framework.

Rules developers create a reference data model to store values that are used by the rules for comparisons against established standards (i.e., an age limit used to define a 'youthful driver').  Typically, business teams require the ability to change these values independently of IT.  These values are stored in a database outside of the rules application and passed into the rules application as parameters of the service.

Typical Rules Service Project

Projects that use the ERS framework to build integrated rules applications are very similar to other IT development projects in terms of project management.  The primary difference lies in the way that rules are collected, analyzed, approved, and modified by the business stakeholders.  It is important that the business project team and technical project team collaborate in defining, executing, monitoring, and correcting this process, and should work together to create templates that define the information that must be collected about each rule.  Business rules documents become the contract between IT and the business sponsor and deserve a formalized change management process with an appropriate amount of rigor and ceremony.

Figure 4. Timelines of a typical project that uses the ERS Framework.

Figure 4 illustrates the timelines of a typical project that uses the ERS Framework.  The ERS framework reduces the overall complexity of getting a new rules project up and running because it limits the technical knowledge required by the development team responsible for encoding the business knowledge into rules.

Case Study: Automated Underwriting

RAM's framework for rules applications was recently used to support a service that provides real-time online underwriting for insurance applications.  This service relies heavily on rules technology for evaluating risk and determining the appropriate pricing tiers.

The rules team performed an analysis of the benefits of the ERS framework to their project.  This report concluded that the ERS framework significantly improved application quality and effectively reduced the overall cost of the project by an entire man-year.

The conclusions of this report are summarized below:

• Only 300 distinct rules needed to provide two services.
• Rule management features of the ERS framework reduced the number of rules required by at least 70%.
• Reference data features allow the business owners to manage certain aspects of the rules independent of IT.
• Reduced project startup time by 16 weeks.
• Reduced staff learning curve by 8 weeks per person (32 weeks total).
• ERS Framework significantly improved application quality by providing best-of-breed design patterns and application templates.
• ERS Framework design allows for framework evolution without disrupting project continuity.

Most importantly, the project team was able to concentrate its energies on fulfilling the business objectives of turning business knowledge into reusable, independent services.  The following is a statement from an internal memorandum produced by the rules project team:  "The overall result [is that ERS] has allowed the project to let business drivers be the primary focus while minimal time has been required focusing on technical architecture."

Enterprise Rules Services Going Forward

As the insurance industry continues to experience consolidation through mergers and acquisitions, and pressure to reduce insurance operations-related expenses, Service Oriented Architecture and rules technology will continue to be critical technologies for facilitating improved time-to-market of critical applications, reduced scope and cost of integration-related development, and the manageable encoding of business knowledge.  These technologies will be cornerstone technologies for providing the business agility required by the complex integration and operational requirements of insurance providers.

---

References

[1]  Golden Eagle Insurance, Pacific Region; Montgomery Insurance, Southeast Region; Colorado Casualty, Mountain Region; Peerless Insurance, Northeast Region; Indiana Insurance, Midwest Region; America First Insurance, Gulf Region; and Hawkeye-Security Insurance, North Central Region.  The group also is represented by two specialty insurers:  Summit Holding Southeast, offering workers' compensation products and services; and GoAmerica Auto Insurance, offering specialty auto insurance. 

# # #