IBM Platform Promises Easy Predictive Modeling

Lots of business people are captivated by the idea of using predictive analytics, but not every organization has analytics gurus on staff. IBM plans to bridge that gap with IBM SPSS Decision Management software.

"It used to be that if you're trying to do what-if modeling, you had to turn to statisticians, but there are only so many analytics Jedi Knights," said Jeff Jonas, chief scientist of the IBM Entity Analytics group. "This new platform enables millions of business analysts to do what used to require statisticians, so it's going to open up a whole new range of things that organization can do for themselves."

The Decision Management platform taps predictive modeling, business rule and optimization technologies to automatically deliver accurate decisions in high volumes and at fast speeds. The platform integrates available real-time and historical information from structured and unstructured data sources and delivers recommendations that can be embedded into transaction systems and human interfaces such as Web sites and contact center operator screens. Non Jedi Knights use a simple, plain-language interface to tweak thresholds and rules so that the engine's automated decisions are consistent with the business outcomes desired.

"The interface lets the business user specify the source of the data, define what the business is trying to predict and then push a button to deliver a model while hiding the complexity of the mathematics in the background," said Erick Brethenoux, IBM's predictive analytics strategist.

What kinds of decisions will the software handle? The platform will ultimately be applied in many high-volume, high-value scenarios, according to IBM, but it's launching with two available applications: Customer Interaction and Claims.

The Customer Interaction application can be applied to inbound and outbound retention, cross-sell and up-sell scenarios via call center, web site, point-of-sale or e-mail channels. Integrated into these environments, the platform can trigger recommendations predicted to have the best fit for each specific customer.

The Claims application is designed to spot potentially fraudulent insurance claims through automated risk assessment. Claims adjustors or business analysts will use the application's simple interface to adjust how risk is defined as business conditions and objectives change. The engine can then spots the claims that can be quickly approved and those that seem suspicious and require investigation. This automated assessment speeds the entire claims process and increases overall customer satisfaction, according to IBM.

IBM SPSS Decision Management for Customer Interaction and IBM SPSS Decision Management for Claims are available immediately worldwide. Costs were not disclosed.