AI-Driven Process Metadata Classification

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing process design tools are inefficient and error-prone in classifying process metadata, requiring manual review and selection, which can lead to inconsistencies and compliance issues with data privacy regulations.

Innovation Solution

The use of machine learning models, such as generative models, to automatically classify process metadata, flag personal data attributes, and categorize attributes based on predetermined schemes, thereby facilitating automated compliance with data privacy regulations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual classification of attributes is performed, then users can review and classify each attribute, but the process becomes tedious, error-prone, and inefficient

Engineering Contradiction:
Improveclassification accuracyVSAvoidclassification efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent replaces the manual mechanical classification process with an automated machine learning-based classification system. The system uses trained models to automatically classify attributes into categories such as personal data, financial data, and operational data, eliminating the need for manual review while maintaining or improving classification accuracy and significantly increasing efficiency.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Adaptability or versatility

If manual classification is used, then users can specify attribute categories, but inconsistencies between users and process instances occur

Engineering Contradiction:
Improveclassification flexibilityVSAvoidclassification consistency
Core Design Contradiction:
Adaptability or versatilityVSStability of the object's composition

Solution Approach 1:

The system enables self-service automated classification where the machine learning model independently and consistently classifies attributes across all process instances without human intervention. This ensures uniform classification standards are applied universally, eliminating inter-user inconsistencies while maintaining the ability to adapt to different data types through trained models.

Inventive Principle:
Principle #25Self-service

3Productivity

If automated machine learning classification is implemented, then efficiency and accuracy are enhanced, but system complexity increases

Engineering Contradiction:
Improveclassification efficiencyVSAvoidsystem complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by pre-training machine learning models with labeled training data before deployment. The models are prepared in advance with learned classification capabilities, allowing them to automatically and accurately classify attributes when deployed in the process design system, thereby achieving high efficiency without requiring complex real-time processing.

Inventive Principle:
Principle #10Preliminary action

4Ease of manufacture

If personal data attributes are not automatically flagged, then compliance review is simpler, but compliance issues with data privacy regulations may occur

Engineering Contradiction:
Improvecompliance review simplicityVSAvoidcompliance assurance
Core Design Contradiction:
Ease of manufactureVSReliability

Solution Approach 1:

The system implements feedback by automatically flagging personal data attributes and providing this information back to users during the process design phase. This feedback mechanism ensures compliance requirements are met by highlighting attributes that require special handling, allowing users to review and adjust classifications while maintaining compliance assurance without overly complicating the review process.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP4564190A1Artificial intelligence-driven data classification
Publication Date: 2025.06.04 SAP SE
  • EP4564190A1 patent drawingFigure 1
  • EP4564190A1 patent drawingFigure 2
  • EP4564190A1 patent drawingFigure 3

AI summary

Systems and methods described herein relate to artificial intelligence-driven data classification. Process metadata of an automated process associated with a process design application is accessed. The process metadata includes a plurality of attributes. Prompt data is generated by adding an instruction to the process metadata. The instruction includes at least one request to classify the plurality of attributes according to a predetermined classification scheme. The prompt data is provided to a machine learning model to obtain output. The output includes a classification result for each of the plurality of attributes that is based on the predetermined classification scheme. The classification results are stored in association with the automated process. One or more of the classification results are presented in a user interface.