Generative Workflow
An AI execution paradigm in which a system dynamically generates and executes workflow structures based on user intent. Core capabilities include data retrieval, step orchestration, state management, and tool invocation. It serves as the underlying execution engine for generative UI, enabling flexible, intent-driven process construction rather than reliance on predefined workflow templates.
Timeline (last 90 days)
The difference between RPA (Robotic Process Automation) and generative workflows is that RPA processes are manually recorded or scripted in advance, whereas the execution graph of a generative workflow is inferred in real time by AI
The article divides agent systems into two layers: Generative UI (the presentation layer, addressing how users interact) and generative workflows (the execution layer, addressing how applications complete tasks)
The article proposes that the interaction paradigm should evolve from prompt→answer to a feedback loop of intent→UI+workflow→execution→feedback→workflow update→better execution
All Facts (3)
The difference between RPA (Robotic Process Automation) and generative workflows is that RPA processes are manually recorded or scripted in advance, whereas the execution graph of a generative workflow is inferred in real time by AI
50%UnverifiedThe article divides agent systems into two layers: Generative UI (the presentation layer, addressing how users interact) and generative workflows (the execution layer, addressing how applications complete tasks)
50%UnverifiedThe article proposes that the interaction paradigm should evolve from prompt→answer to a feedback loop of intent→UI+workflow→execution→feedback→workflow update→better execution
50%