No-Code Workflow UI for Structured Data Capture From AI Dialog
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Capturing structured information from free-form human conversation is challenging due to the ambiguity and variability of human responses, requiring complex software engineering efforts and deterministic workflows that are difficult to maintain.
Innovation Solution
A system utilizing a Large Language Model (LLM) with prompt engineering to extract structured data points from unstructured conversations, facilitated by a no-code user interface that allows defining workflows through a flexible, single-step process, switching between structured and unstructured logic to capture required parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If deterministic workflows with complex software engineering are used to capture structured information, then measurement precision is improved, but device complexity increases and ease of operation deteriorates
Solution Approach 1:
The patent introduces a natural language processing intermediary layer that translates unstructured human conversation into structured data. This mediator handles the complexity of human language variability, allowing the underlying workflow system to remain simple and deterministic while still achieving high precision in capturing structured information.
Solution Approach 2:
The patent replaces traditional mechanical/deterministic workflow systems with an AI-based generative system that can dynamically adapt to unstructured human input. This substitution allows the system to maintain simplicity while handling the complexity of natural language through machine learning models rather than rigid programming.
2Reliability
If deterministic workflows are used to ensure comprehensive coverage of human responses, then reliability is improved, but adaptability deteriorates
Solution Approach 1:
The patent implements a dynamic workflow system that can switch between structured and unstructured processing modes based on the nature of human input. This dynamic adaptation allows the system to maintain reliability for predictable scenarios while gaining versatility to handle unexpected or varied human responses through generative AI capabilities.
Solution Approach 2:
The patent creates a universal workflow system that can handle both structured and unstructured data through a single platform. The system provides multi-functionality by combining deterministic workflow execution with generative AI capabilities, allowing it to adapt to different types of human input while maintaining a consistent reliable framework.
3Adaptability or versatility
If generative AI is used to handle unstructured conversation, then adaptability is improved, but manufacturing precision deteriorates
Solution Approach 1:
The patent introduces a structured output framework as an intermediary that constrains the generative AI's output to follow predetermined schemas. This mediator ensures that while the AI can adapt to various unstructured inputs, the resulting output maintains consistent structure and precision required for business process automation.
Solution Approach 2:
The patent changes the parameters of the generative AI system by introducing structured constraints and schemas that guide the output format. This parameter adjustment allows the system to maintain adaptability in processing unstructured conversation while ensuring manufacturing precision in the resulting structured data through controlled output parameters.
4Reliability
If complex deterministic workflows are built to capture all human input permutations, then completeness is improved, but loss of time increases
Solution Approach 1:
The patent replaces time-consuming manual workflow creation with generative AI that can automatically generate and adapt workflows. This substitution dramatically reduces development and maintenance time while maintaining comprehensive coverage of human input scenarios through the AI's ability to learn and adapt to various response patterns.
Solution Approach 2:
The patent enables the workflow system to serve itself by using generative AI to automatically create, update, and maintain workflows based on observed human interactions. This self-service capability eliminates the need for continuous manual programming and maintenance, reducing time loss while ensuring comprehensive coverage through automated learning.
Data Source
AI summary
In an embodiment, the following are repeated during a session with a user compute device. A message is received from the user compute device. A determination is made as to whether a workflow portion from a plurality of workflow portions is structured or unstructured. The workflow portion is executed as structured based on the message when the workflow portion is determined to be structured, to produce a structured response. The workflow portion is executed as unstructured based on the message, a task description and a list of structured parameters to be captured, when the workflow portion is determined to be unstructured, to produce an unstructured response via a generative artificial intelligence model.