Machine Learning Payment Extraction from Emails and Documents

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

The current manual process for extracting payment information from electronic mails and documents is labor-intensive, error-prone, and inefficient, leading to delays and inaccuracies in accounts receivables management.

Innovation Solution

A machine learning-based computing system and method that automatically extracts payment information from electronic mails and documents by tokenizing text, analyzing contexts, and using a machine learning model to determine payment amounts and identifiers.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If manual process is used for extracting payment information, then human judgment and flexibility are maintained, but labor intensity and error rate increase significantly

Engineering Contradiction:

Improveaccuracy of payment information extractionVSAvoidlabor intensity of extraction process

Core Design Contradiction:

ReliabilityVSEase of operation

Solution Approach 1:

The system enables self-service automation where the machine learning model automatically extracts payment information from emails and attachments without human intervention. The model processes documents, identifies payment details, and updates databases autonomously, eliminating manual labor while maintaining high accuracy through trained algorithms.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces the mechanical manual process of reading and extracting information with an automated machine learning system. The ML model substitutes human cognitive operations with computational processes, using natural language processing and pattern recognition to extract payment information accurately without human effort.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Productivity

If manual extraction and recording is performed, then data accuracy can be verified by human judgment, but processing time and operational delays increase

Engineering Contradiction:

Improvespeed of payment information processingVSAvoidtime delay in updating accounts receivables database

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The system implements continuous automated processing where the machine learning model operates without interruption to extract payment information from incoming emails and updates the accounts receivables database in real-time. This continuous automated workflow eliminates the batch processing delays inherent in manual operations, maintaining constant productivity without time losses.

Inventive Principle:

Principle #20Continuity of useful action

3Productivity

If automated machine learning system is implemented, then productivity and speed are significantly improved, but system complexity increases

Engineering Contradiction:

Improveefficiency of payment information determinationVSAvoidcomplexity of machine learning based system

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The machine learning system is designed as a universal multi-functional platform that handles various document types (emails, PDFs, images), extracts multiple payment information fields (amounts, identifiers, dates), and integrates with different database systems. This universality consolidates multiple functions into a single system, managing complexity through integration rather than multiplication of components.

Inventive Principle:

Principle #6Universality (Multi-functionality)

4Reliability

If manual process is used, then system simplicity is maintained, but error proneness and labor costs increase

Engineering Contradiction:

Improveerror rate in payment information extractionVSAvoidcomplexity of automated extraction system

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The system incorporates feedback mechanisms where the machine learning model continuously learns from extracted payment information and validation results. The feedback loop allows the system to refine its extraction accuracy over time, reducing errors while managing complexity through iterative improvement rather than requiring perfectly complex initial design.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS20250217799A1Machine learning based systems and methods for identification of payment information from electronic mails

Publication Date: 2025.07.03 HIGHRADIUS CORP

AI summary

A machine learning based computing method for determining payment information from electronic mails, is disclosed. The machine learning based computing method includes steps of: receiving data from databases; extracting information tokens from the data associated with electronic mails and electronic documents; determining payment information features for the information tokens by analyzing contexts of the information tokens extracted from the data associated with electronic mails and electronic documents; selecting optimum information tokens by analyzing the payment information features by parameter-driven pre-configured rules; determining first payment information including payment amounts and payment identifiers within the electronic mails and the electronic documents, for the optimum information tokens by a machine learning model; and providing an output of the first payment information including the payment amounts and the payment identifiers to users on a user interface associated with electronic devices.