OCR Pre-Verification System for Payment Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current optical character recognition systems face inefficiencies in processing payments with varying degrees of legibility and completeness, leading to unnecessary human intervention and resource allocation, as they either require full verification or no processing at all, without a suitable middle tier for mediocre payments.
Innovation Solution
A three-tiered system utilizing optical character recognition software to categorize payments into straight-through processing, pre-verification, and verification subsystems, with a receiver and scanner to transform paper checks and remit stubs into electronic records, determining the necessary resource allocation based on confidence levels for each payment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a two-operator verification system is used for all payments, then processing accuracy is improved, but resource consumption and processing time increase unnecessarily for proper payments
Solution Approach 1:
The verification system is segmented into three distinct subsystems (STP, pre-verification, and verification) that process payments based on their quality characteristics. This segmentation allows proper payments to be handled by automated STP while mediocre payments receive pre-verification, and only illegible payments undergo full two-operator verification, thereby reducing unnecessary resource consumption for high-quality payments while maintaining processing accuracy for problematic ones
Solution Approach 2:
Different verification levels are applied to different payments based on their local characteristics (legibility, completeness, confidence scores). Proper payments with high confidence scores receive minimal or no human verification, while mediocre payments receive intermediate pre-verification, and illegible payments receive full verification. This local quality approach ensures that verification resources are allocated precisely where needed
2Productivity
If straight-through processing is used for proper payments, then processing speed is improved, but the system cannot handle mediocre or illegible payments
Solution Approach 1:
The system dynamically routes payments to different processing subsystems based on real-time analysis of payment quality characteristics. The optical character recognition software continuously evaluates confidence scores and payment legibility, dynamically directing proper payments to STP for high-speed processing while routing mediocre and illegible payments to pre-verification or verification subsystems. This dynamic adaptation allows the system to maintain high processing speeds for quality payments while seamlessly handling problematic payments through appropriate verification channels
Solution Approach 2:
The integrated payment processing system incorporates multiple processing capabilities within a single unified architecture. The STP subsystem handles proper payments automatically, the pre-verification subsystem handles mediocre payments with minimal human intervention, and the verification subsystem handles illegible payments with full human review. This multi-functional design allows a single system to efficiently process all types of payments (proper, mediocre, and illegible) through appropriate subsystems, thereby enhancing both processing speed and payment type coverage
3Reliability
If full verification is applied to all payments, then processing accuracy is improved, but time and human resources are wasted on proper payments that do not require intervention
Solution Approach 1:
The system performs preliminary automated analysis using optical character recognition software to evaluate payment quality characteristics and confidence scores before human verification. This preliminary action identifies proper payments that can be processed automatically through STP without human intervention, thereby preventing unnecessary human review time for high-quality payments while ensuring that mediocre and illegible payments are flagged for appropriate verification levels
Solution Approach 2:
The system applies verification actions proportionally to payment quality - proper payments receive minimal or no human verification (partial action), mediocre payments receive pre-verification with limited human intervention (intermediate action), and illegible payments receive full two-operator verification (excessive action). This graduated approach ensures that verification resources are not excessively applied to proper payments while maintaining sufficient verification for problematic payments, thereby optimizing the balance between verification accuracy and processing time
Data Source
AI summary
A system for optical character recognition pre-verification is provided. The system may review payment documents. The payment documents may include remit stubs and paper checks. The system, utilizing optical character recognition software, may determine dollar amounts on the remit stubs and the paper checks. The optical character recognition software may determine a confidence level of whether the determined amount is the same amount that the writer of the check. If the confidence level is above first predetermined threshold level of confidence and below a second predetermined threshold level of confidence, the system may present an operator with a pre-verification GUI. The pre-verification GUI may include a view of the remit stub, a view of the check, the dollar amount due, a match button and a do not match button. Upon selection of either the match or do not match button, the payment may be processed in another check-processing system.