Automated Pricing Grid Update via ML Rating Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Commercial lenders face inefficiencies in updating performance-based pricing for borrowers, as current methods rely on a manual, offline process to reflect changes in external performance ratings from agencies like Moody, S&P, and Fitch.
Innovation Solution
A system and method utilizing a pricing computer program that automates the process of polling rating agencies, detecting rating changes, and using a machine learning engine to predict and implement recommended changes to a pricing grid, thereby updating loan terms and payments dynamically.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a manual offline process is used to update borrower performance ratings, then human review and control are maintained, but the process efficiency and speed of updating pricing are reduced
Solution Approach 1:
The system segments the pricing update process into automated components (polling rating agencies, detecting changes, calculating pricing adjustments) and human review components. The automated portion handles routine data collection and initial analysis, while human reviewers focus on exceptional cases or final approval, thereby improving overall efficiency while maintaining necessary control.
Solution Approach 2:
The pricing system acts as an intermediary between rating agencies and loan pricing, automatically translating rating changes into pricing grid adjustments. This intermediary function reduces manual intervention by handling the complex calculations and data transformations that would otherwise require manual offline processing.
2Measurement precision
If manual processes are used to track and update borrower ratings from multiple rating agencies, then data accuracy can be verified, but the time and resources required for the process increase
Solution Approach 1:
The system implements self-service by having the pricing computer program automatically poll rating agencies, detect rating changes, and calculate pricing adjustments without requiring manual data collection or verification. The automated system serves itself by continuously monitoring and updating pricing grids based on external rating data.
Solution Approach 2:
The system enables continuous monitoring of borrower ratings by automatically polling rating agencies at regular intervals. This continuous action ensures that rating changes are detected immediately and pricing is updated in real-time, eliminating the discontinuous nature of manual monitoring and significantly reducing the time required.
3Productivity
If automated systems are used to update pricing grids, then processing speed increases, but system complexity and the need for machine learning models increase
Solution Approach 1:
The system performs preliminary action by pre-calculating and storing pricing grids for different rating scenarios using machine learning models. When a rating change occurs, the system simply retrieves and applies the pre-calculated pricing grid rather than performing complex calculations in real-time, thereby reducing the computational complexity required during actual pricing updates.
Solution Approach 2:
The system incorporates feedback mechanisms where machine learning models are trained on historical rating changes and pricing outcomes. The models continuously learn from past performance data, improving their accuracy over time and reducing the need for complex manual adjustment rules, thereby simplifying the overall system architecture while maintaining high productivity.
Data Source
AI summary
Systems and methods for determining performance-based pricing are disclosed. A method may include: creating a borrower entry based on borrower loans for a borrower and a borrower rating; polling a plurality of rating agencies for agency borrower ratings; receiving agency borrower ratings from the plurality of rating agencies; determining that one of the agency borrower ratings has changed from a previous agency borrower rating; predicting, using a machine learning engine that is trained with historic agency rating changes, a recommended change to a pricing grid for the borrower based on the change in the agency borrower rating; updating the pricing grid for the borrower based on the recommendation; and providing the updated pricing grid to a loan platform. The loan platform is configured to implement the pricing grid, and the implementation of the pricing grid changes a payment for at least one of the plurality of borrower loans.