Overcoming the Limitations of Legacy Rules-Based Models

Disruptive IT Trends2 minutes readSep 12th, 2019

We live in a fast-moving, globally-connected digital world. Manual systems and limited personnel can no longer keep up with the growing volume of transactions and data processing required to make timely, informed decisions. Legacy rules-based analytical models cannot scale to meet today’s requirements. Fortunately, technology innovation is fueling significant improvements in analytics to meet the needs of business in our data-rich world. The application of artificial intelligence (AI) and machine learning is ushering in the next generation of analytics modeling using data-driven decision making.


Where We’ve Come From

Where We’re Going

Limited Data. Legacy models use subsampling and analyze cases as groups, ignoring important data sources that could impact outcomes.

Holistic Data. Advanced analytics that leverage increased transaction volumes and across any useful data sources.

Group Decisions. Hardcoded rules and simple linear models can restrict decision-making to the group level.

Individual Decisions. More accurate, granular models enable a shift from decisions for larger groups to understanding individual situations.

Limited Scope. Legacy models are limited by the perspective of those creating the rules, making it hard to incorporate parameters that are not anticipated or impose human bias. As new rules are added, it becomes more difficult to see gaps amid the overlapping rules within hundreds of rulesets.

Unlimited scope. Advanced algorithms in AI and machine learning infer this automatically from data. They examine data quickly and update models frequently, identifying newly-emerging patterns and correlations to continually improve results.

Generalization Error. Since the number of samples and variables is restricted in legacy models, the best that automated decisioning solutions can do is to define groups or segments which share some high-level characteristics. This over-generalization increases the chances of false positives and negatives.

Increased Accuracy. By collecting and analyzing more information with intelligent automation, “normal” patterns common in different dimensions of the data are learned. The machine is then able to pick up “abnormal” outliers at the individual level.


To learn more about how to realize the full benefits of advanced analytics for your business, read the paper Changing The Risk Evaluation Process Through Advanced Analytics today.

Tags-   AI Analytics Machine Learning