Hitmetrix - User behavior analytics & recording

Blaze Advisor: Shared Rule Management

Nearly every modern customer contact system can execute rules during an interaction. In most cases these rules are created and maintained within the system itself.

This is fine as long as the system runs independently, but is a problem if you want to ensure each customer is treated consistently from one system to another. Without a way to share rules, you must somehow recreate the same rules independently in each customer contact system. At best, this is a maintenance headache. At worst, it’s impossible – if the rule engines in some systems are unable to do what you want.

One approach is to avoid the problem by using a single product, with a single set of rules, for contacts in all different channels. But most organizations have multiple systems and lack the time or resources to replace these with a single solution. Even firms willing to commit to one system are likely to find themselves using different products on occasion: perhaps an old operational system is too hard to replace, or perhaps you need a function that the preferred vendor does not yet support. Nor is there a guarantee that the rules engine in the preferred system will be powerful enough to meet all your requirements. After all, a multichannel customer management system must meet many different criteria – and it’s likely that functions other than rule sophistication were given top priority.

Blaze Advisor (HNC Software, www.hnc.com, 877/464-4849) is designed to address these issues by providing powerful rule management functions that can be shared throughout an organization. Because power and sharing are separate issues, let’s look at each in turn.

In terms of power, Blaze is hard to beat. It’s one of those systems that solves problems you didn’t know you had. Users can write rules in three languages, ranging from English to geek. The rules themselves extend well beyond simple if/then logic to include complex calculations, procedural constructs such as program loops, decision tables that list the outcome for each specified combination of inputs, and question sets that can gather several answers before moving to the next decision.

Rules can be grouped into rule sets, which in turn can be arranged in rule flows on a graphical flow chart. This lets users control the sequence in which rules are executed and define multi-step, branching processes within a single flow. Special technology ensures that rules run efficiently even when there are thousands of cases and when one rule refers to the outcome of another. Other technology lets Blaze gather data from nearly any source to use as input during rule execution. A data dictionary lets users reference this data in business rather than technical terms.

Blaze’s administrative functions are similarly robust. Rule sets help organize rules into manageable groups, with users assigned different rights for different groups. The system can generate Web-based templates that let end-users make some rule changes – say, updating the volume discounts in a price list – but not others. It also provides debugging tools to ensure the user understands exactly how a given rule will operate, tracks rule changes and deployment, provides reports on which rules are firing in production and how quickly they are running, and lets users store an audit trail of the process associated with each business decision.

Blaze is less impressive when it comes to sharing. Basically, it’s up to the user to modify existing systems to call Blaze and to process Blaze’s results. Arguably, Blaze does all it can to make this easy, by supporting inputs and outputs in a variety of formats. But some vendors go a step further by providing prebuilt adapters for third-party products, while Blaze does not. As a result, users are left with the need to perform custom integrations. Avoiding the pain of such integrations is a primary reason companies look for single-system solutions in the first place.

It’s also important to realize that Blaze is not – and does not claim to be – a true interaction management system. These are products designed to coordinate treatments across separate customer contact systems. Examples include Elity, Yellow Brick and Harte-Hanks Allink Agent. These systems include specialized features such as predictive modeling, response analysis, behavior pattern tracking and customer profile maintenance. As a general-purpose rules engine, Blaze should not be expected to provide these functions. Still, companies considering Blaze as part of a customer management architecture should recognize they will need to find other ways to meet those requirements.

As the sophistication of its rule functions suggests, Blaze has been around for some time. The product traces its origin back 15 years, while the current Java-based version was initially released in 1998.

Blaze has been sold to 300 companies and is used as a rule engine within several major customer management products. The system is highly scalable, supporting parallel processing and running on nearly any platform including mainframe, Unix, Linux and major application servers. Wizards automatically generate configuration and installation files to help speed deployment.

Blaze was originally developed by an independent software company. The system has changed owners twice in recent years, and the current owner, HNC, announced a merger with Fair, Isaac in May 2002. HNC and Fair, Isaac have announced plans to incorporate Blaze as part of a larger set of offerings aimed at integrating and optimizing enterprise business decisions.

Pricing of Blaze is based on the number of rule development seats and the size of the servers it will be deployed on. Cost can start as low as $50,000 while a typical enterprise customer might expect to pay $250,000 for licensing, support, basic training and start-up services.

Total
0
Shares
Related Posts