Automated Invoice Verification via OCR and Purchase Order Matching

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Manual verification of invoices in accounting systems is labor-intensive and prone to errors, as it involves comparing printed invoice data with purchase order data, often leading to discrepancies and requiring special processing.

Innovation Solution

An automated system retrieves image data from invoices, extracts text fields, and compares them to data from a purchase order system, designating mismatches for special processing while allowing users to add or correct data, thereby reducing manual labor and errors.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If manual verification of invoices is performed, then data accuracy can be maintained through human review, but labor intensity and processing time increase significantly

Engineering Contradiction:

Improvedata accuracyVSAvoidprocessing efficiency

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The patent replaces the mechanical manual verification process with an automated optical character recognition (OCR) system that captures invoice data, extracts text fields, and compares them against purchase order data electronically. This substitution eliminates manual labor while maintaining verification accuracy through systematic automated comparison of invoice data with purchase order data.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system enables self-service verification by automatically performing data extraction, comparison, and mismatch identification without requiring human intervention for routine verification tasks. The automated system serves itself by systematically processing invoices, identifying discrepancies, and flagging items requiring special processing, thereby freeing human workers from repetitive manual verification work.

Inventive Principle:

Principle #25Self-service

2Productivity

If automated data extraction is implemented, then processing speed increases, but error rates may increase due to misreading or misinterpretation of text fields

Engineering Contradiction:

Improveprocessing speedVSAvoiddata accuracy

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The system incorporates feedback mechanisms by automatically comparing extracted invoice data against purchase order data in the external system, identifying mismatches, and flagging them for special processing. This feedback loop ensures that automated extraction errors can be detected and corrected by routing suspicious items to human review, thereby maintaining data accuracy while preserving processing speed benefits.

Inventive Principle:

Principle #23Feedback

3Productivity

If complete automated verification is performed, then labor costs decrease, but system complexity and initial implementation costs increase

Engineering Contradiction:

Improvelabor efficiencyVSAvoidsystem complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The verification system is segmented into distinct functional modules: data capture module for OCR processing, data extraction module for identifying text fields, comparison module for matching against purchase order data, and special processing module for handling mismatches. This segmentation allows the complex verification task to be divided into manageable components, reducing implementation complexity while maintaining comprehensive automated verification capabilities.

Inventive Principle:

Principle #1Segmentation

4Loss of time

If all invoice data is automatically processed, then processing time is reduced, but special cases requiring human judgment may be overlooked

Engineering Contradiction:

Improveverification timeVSAvoidhandling of special cases

Core Design Contradiction:

Loss of timeVSReliability

Solution Approach 1:

The system performs preliminary automated verification of all invoice data, quickly identifying and flagging items with mismatches before they require human review. This preliminary action filters out routine items that can be processed automatically, allowing human workers to focus only on flagged special cases that require judgment, thereby reducing overall processing time while ensuring special cases are not overlooked.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS10176512B2Processing electronic data across network devices

Publication Date: 2019.01.08 RICOH CO LTD

AI summary

An approach is provided for processing electronic data across network devices. One or more text fields represented in image data that represents an electronic document are compared to data from a first external system. If the one or more text fields represented in the image data do not match the data from the first external system, then the one or more text fields represented in the image data are designated for special processing. If the one or more text fields represented in the image data match the data from the first external system, then the electronic document is designated as verified. The approach also includes the ability for a user to supplement image data with additional data, such as codes used by business organizations, and for automatic correction of errors in text fields using data maintained by the first external system.