There are two ways to use analytic techniques to improve business rules.
Firstly you can use data mining techniques. This involves using data driven analysis to identify more profitable rules and strategies. With data mining you can use historical data to see what segments of a population respond best to what actions or what values predict members of which group and turn this into business rules. With optimization techniques, multiple data variables, predictive models, business objectives and constraints to produce optimized decision strategies or “decision models”. With optimization, decision models can determine the most successful rule sets, models and thresholds to reach explicit business goals and meets constraints.
Secondly you can embed more sophisticated predictive analytic models into rules-based decisioning. Predictive analytic models improve the precision and targeting of decisions based on forcasts of future customer behavior and other metrics. These models work within business rulesets to segment customers for more targeted and relevant actions. Thus I can write rules that use these predictions to improve the actions I take such as routing a customer with a strong prediction of churn to a better offer or more knowledgeable CSR.