Business Rule Engine (BREx)

Impactful businesses prosper by implementing strategic decisions through certain guiding business rules. It is the sanctity of the rules which ensure the optimal decisioning, which enhances business outcomes. Decisions may misfire or be counterproductive, if not backed by data and a logical analysis of such data. It is therefore paramount that data is scrubbed, rechecked, and analysed in depth to arrive at strategically aligned business decisions. It is in such an environment that BREx becomes a big differentiator for business driven by data rules. BREx helps a credit policy designer or an underwriter minimize potential NPAs.

Features and Advantages

Some of the key functions of BREx are:

Detailed Features

  • Complex credit decision rules support
  • Built-in eligibility calculators
  • Multi-tier scorecards for custom risk scoring
  • Automated (credit) policy deviations management
  • Simulation for decision analytics
  • Built-in financial object models (bureaus, bank statements, Tax data, Financial Statements) for quick rules setup
  • Workflow processing supported with generation of custom computed fields
  • Intuitive web interface for easy business rules authoring and deployment
  • Version control for appropriate rules & policy roll-outs
  • Granular access control (user & role based) for better security
  • Quick rule & policy deployments without requiring technology team support
  • Configurable integration with just about any application system
  • Low latency and highly scalable execution engine
  • Efficient horizontal scaling to handle variable traffic patterns
  • Constant quick response times (over 200% better than the closest competitor) even at increasing transaction volumes

What Does It Exactly Do?

The Business Rule Engine (BREx) is a decision making tool that helps users evaluate applications. Applications (data inputs) can be categorized and evaluated against rules that can be configured as per the requirement of the business. These categorizations, rules and their weightages can be set by user as per their convenience and as per their own standards. These Rule-settings take only a few seconds to update and can be set-up easily by the user in a highly user friendly interface framework. The rule-setting requires no prior technical knowledge and is highly intuitive and configurable. Once the rules are set, the same can be pushed into the working algorithm with just a click. Apart from the ease in rule settings, BREx also provides built-in analytics that help determine application-wise performance both from a technical as well as functional perspective.

How does it do it?

BREx Workflow diagram

Use Cases

The rules engine comes with a multi-tier scorecard definition module that can be used to generate custom risk scores as well as other decision scores

Untracked deviations from standard underwriting policies without meaningful mitigants, and a lack of proper review and approval, can be big contributors to a loan portfolio’s losses. BREx can scan through application data real-time (including data from other standard sources like the credit bureaus) to swiftly determine, log and raise policy & operational deviations to relevant personnel for review. This reduces turnaround times and errors significantly.

BREx allows complex loan eligibility computations using application data, derived risk scores as well as other derived (computed) fields.

Configurable Portfolio Level Rules which get triggered on a real time basis. The objective is to proactively control the underwriting process instead of doing it with a lag.

Few Examples:

  • 1)Notify the credit manager if the count of applications sourced from the dealer A during the last 15 days (cumulative) of the month cross 30% of the applications sourced from the underlying location. In case, it exceeds 35% mark, then pass the applications of the Dealer A to Hold Queue. (Early Warning Trigger - Risk Control - Exposure Levels to Dealer - Portfolio Sourcing Diversification)
  • 2)If Approval Rate for Bureau Score Band 650-700 during the day accounts for more than 25% of the total approvals, then notify the credit manager X. (Too risky to sustain)

Ability to host, trigger and communicate with the advanced ML AI models in seamless manner for consuming them real time in the underwriting decision process.

Flexibility to configure and run multiple models/ scorecards offline and compare their performance and/or results on a real time basis by leveraging our highly interactive user interface. Further to this, switch to the offline scorecards that emerge as potential challengers to existing live model/s in an effortless manner.
Advantage - Explore multiple scorecards leveraging different models / approaches/ strategies and measure impact real time and perform course correction (risk mitigation / business growth), all at the click of the button, thus saving data extraction, cleansing, analysis and eventual development time delays. ‘You sense it, you see it, you crack it’.

Plug the repayment information, Bureau scrubs data into the engine. Configure rules to validate the scorecard performance and take corrective actions.
Advantage - Shift from the age old mindset where LMS does not communicate back to Rule Engines/ LOS. Why not leverage the feedback loop in today’s machine learning world?

Most credit managers would agree with the following:

  • 1) Many rules never get executed
  • 2) Some rules decay over time
  • 3) Some rules last over time, however their class boundaries may become less effective. Eg: Bureau Score greater than 700 set in Jan 2020 would have to shift to 720 in Mar 2020
  • 4) Some rules are redundant, some parameters are redundant. These lead to operational delays, functional gaps and systems may need sanitization or clean up

Rule optimizer being intelligent avoids users from creating redundant parameters and rules. Further to this, it provides detailed insights (reports) which helps optimize the rule engine from time to time, thus improving the overall efficiency.