Respond to AI is a dedicated action in Yeeflow Workflows that allows you to return the current workflow context or results to AI (such as an AI Agent or Copilot) when AI is involved in the process.
It is important to note that Respond to AI does not have to be placed at the end of a workflow.
In real-world scenarios, workflows often include time-consuming steps such as task assignments, approval actions, or email notifications. In many cases, AI does not need to wait for the entire workflow to complete. Instead, you can place a Respond to AI action at a key point in the workflow to immediately return the available results to AI, while allowing the remaining steps to continue executing in the background.
This design enables a workflow to function not only as a full process engine, but also as an instant calculation and decision service for AI.
Core Purpose of Respond to AI
The Respond to AI action is used to:
Return variables or results produced at a specific stage of the workflow to AI
Allow AI to obtain required information without waiting for the entire workflow to finish
Enable asynchronous collaboration between AI and workflows
Improve AI response speed and overall user experience
By placing Respond to AI appropriately, workflows can support a
“fast front-end response with continued background execution” pattern.
Adding a Respond to AI Action to a Workflow
Steps
Open the form or process Workflow editor
In the left component panel, expand Logic and flow
Select Respond to AI
Drag the action onto the workflow canvas at an appropriate position
Select the action and configure it in the right-hand settings panel
Action Configuration
Action name (Required)
Defines the name of the Respond to AI action within the workflow.
This name is used only for internal identification on the workflow canvas and does not affect execution or the response sent to AI.
Description (Optional)
Provides a brief explanation of what the action returns to AI or which stage of the workflow it represents, such as “Return validation result” or “Respond before approval steps.”
This field is for documentation purposes only and does not affect workflow or AI behavior.
Execution Logic
When the workflow reaches a Respond to AI action:
The current workflow context data is immediately returned to AI
AI uses this data as the response for the current invocation
The workflow is not forced to stop by the Respond to AI action
If additional actions follow (such as task assignments, approvals, or email actions), the workflow continues running in the background
A single workflow may contain multiple Respond to AI actions, typically used in different branches or at different stages.
Common Use Scenarios (Brief)
Returning Intermediate Results
After completing validation, calculations, or queries in the early stages of a workflow, Respond to AI can return the key results to AI without waiting for later manual or asynchronous steps.
Workflows with Time-Consuming Steps
When workflows include approvals, assignments, or notifications, Respond to AI can be placed before these steps to allow AI to respond immediately while the workflow continues.
Asynchronous AI and Workflow Collaboration
AI interprets user intent and triggers the workflow, while the workflow executes business logic. Respond to AI returns stage-based results to AI at key points, enabling more flexible interactions.
Best Practices
Do not treat Respond to AI only as a “workflow end action” — it can also serve as a stage-based response point
Place the action based on AI’s actual data needs to avoid unnecessary waiting
Ensure returned data has clear business meaning for AI
Avoid returning multiple responses without a clear purpose along the same execution path
Summary
The Respond to AI action transforms Yeeflow Workflows from simple end-to-end process executors into intelligent execution engines that AI can interact with at any time.
By returning stage-based results at the right moment, you can achieve:
Faster AI responses
Continued background workflow execution
More efficient and flexible AI + Workflow collaboration
This capability is a key building block for enterprise-grade intelligent automation.