Document Workflow Automation via OCR and Neural Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current workflow initiation methods require users to manually identify and navigate to specific web pages or applications, leading to redundant effort and inconsistencies due to the heterogeneity of service providers, and prior machine learning models struggle with matching entities across different contexts.
Innovation Solution
A system that uses optical character recognition (OCR) and deep neural networks to extract data from documents, identify relevant workflows, and prefill forms by learning relationships and patterns from historical data, allowing users to submit documents to automatically select and execute workflows without manual data re-entry.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If users manually identify and navigate to specific web pages or applications to initiate workflows, then users can execute workflows, but users experience redundant effort and inconsistencies due to heterogeneity of service providers
Solution Approach 1:
The system performs automatic workflow identification and initiation without requiring user navigation. The processor extracts data from submitted documents, automatically identifies relevant workflows, and initiates them, allowing the system to serve itself rather than requiring manual user intervention for each step.
Solution Approach 2:
The system pre-extracts data from documents and pre-identifies workflows before user submission. By having the processor analyze documents and prepare workflow options in advance, the system eliminates the need for users to manually navigate through service provider interfaces at the moment of initiation.
2Productivity
If users manually enter data for each workflow, then workflows can be executed, but redundant effort increases and inconsistencies arise
Solution Approach 1:
The system merges data extraction from multiple documents into a single unified process. The processor extracts relevant data from all submitted documents simultaneously and uses this consolidated information to pre-fill workflow forms, eliminating the need for users to re-enter the same data multiple times across different workflows.
Solution Approach 2:
The system creates copies of extracted document data and reuses them across multiple workflow initializations. Once data is extracted from a document, it is stored and automatically copied to pre-fill forms for relevant workflows, preventing redundant manual data entry.
3Measurement precision
If prior machine learning models are used for entity matching, then some workflow identification is possible, but they struggle with matching entities across different contexts
Solution Approach 1:
The system changes the parameters of entity matching by using document context, document type, and extracted data relationships as additional matching criteria. Rather than relying on fixed entity matching parameters, the processor dynamically adjusts matching based on the specific document context and workflow requirements, enabling accurate matching across diverse contexts.
Data Source
AI summary
In an example embodiment, a solution is provided that allows a user to submit a document. Information can be obtained from the document using optical character recognition (OCR) or other techniques. This information can then be used to identify one or more workflows that pertain to the document. The one or more workflows may be ranked using machine learning techniques and presented to the user. Once the user selects a desired workflow, the information obtained from the document can then be used to automatically complete at least a portion of the workflow, for example by prefilling one or more fields in a form.