Automated Business Process Model Generation from Session Recordings
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for generating business process models are time-consuming and require manual intervention, often resulting in incomplete or inaccurately represented process flows due to the lack of detailed recording during design sessions, necessitating additional effort to fill in missing information and identify correct shapes.
Innovation Solution
A system that automatically generates business process models by analyzing video and audio recordings of design sessions to identify tasks, determine missing shapes, and establish task dependencies, using deep-learning classifiers and natural language processing to accurately represent the process flow with correct geometrical shapes and temporal ordering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual methods are used to generate business process models, then the models can be created with human expertise and judgment, but the process is time-consuming and requires significant manual intervention
Solution Approach 1:
The system captures video and audio recordings of the design session as copies of the original process design activity. These recordings are then analyzed through AI models to extract process flow information, task dependencies, and geometrical shape meanings, automatically generating the business process model without requiring manual recreation of the process documentation.
Solution Approach 2:
The patent replaces manual mechanical processes (human analysts watching videos and manually creating models) with automated AI-based systems. Deep learning classifiers analyze video frames to identify geometrical shapes and their meanings, while NLP models analyze audio transcripts to extract task information and dependencies, substituting human cognitive and manual work with automated computational processes.
2Loss of information
If detailed recording of design sessions is not captured, then the recording devices are simpler and less intrusive, but the process flow information is incomplete or inaccurately represented
Solution Approach 1:
The system uses a multi-functional approach where a single recording setup captures both video and audio simultaneously. The AI processing system performs multiple functions: identifying geometrical shapes in video frames, determining their intended meanings through classification, extracting task information from audio transcripts, and establishing task dependencies. This universal system handles all aspects of process flow capture without requiring separate specialized devices for each function.
Solution Approach 2:
The patent introduces AI models as intermediaries between the raw video/audio recordings and the final business process model. The deep learning classifiers and NLP models act as mediators that translate unstructured multimedia data into structured process flow information, automatically inferring missing details and resolving ambiguities without requiring additional manual documentation or simpler recording setups.
3Manufacturing precision
If manual creation of business process models is required, then the models can be accurately constructed with proper notation, but additional effort is needed to fill in missing information and identify correct shapes
Solution Approach 1:
The system enables self-service automatic generation of business process models by using AI models to autonomously analyze design session recordings and create accurate process models without human intervention. The deep learning classifiers automatically identify geometrical shapes and their meanings, while NLP models extract task information and establish dependencies, allowing the system to generate complete and accurate models independently.
Solution Approach 2:
The system employs feedback mechanisms where the AI models continuously refine their analysis by cross-referencing video frame data with audio transcript information. The classifiers use feedback from multiple video frames and audio segments to accurately determine the meaning of geometrical shapes and their relationships, ensuring high manufacturing precision in the generated process models while maintaining ease of creation through automation.
4Measurement precision
If expertise in both process flow and business process model notation is required, then the models can be created with high accuracy, but the operational complexity and skill requirements increase
Solution Approach 1:
The patent replaces the need for human expert knowledge with automated AI systems. Deep learning classifiers are trained to recognize geometrical shapes and their meanings in process flow diagrams, while NLP models are trained to extract task information and dependencies from design session transcripts. These automated systems substitute human expertise, eliminating the need for operators to possess specialized knowledge of business process model notation while maintaining high accuracy in model generation.
Data Source
AI summary
One embodiment provides a method, including: obtaining at least one video capturing images of a writing capture device used during a business process design session, wherein the images comprise portions of the process flow; obtaining at least one audio recording corresponding to the business process design session; identifying an intended business process model shape; determining at least one business process model shape missing from the process flow provided on the writing capture device; identifying a task dependency for pairs of business process model shapes; and generating a business process model from (i) the intended business process model shapes, (ii) the at least one business process model shape missing from the process flow, and (iii) the identified task dependencies.