Yeeflow AI Agents can be reused across multiple system scenarios, supporting cross-workflow, cross-tool, and cross-platform automation. This guide introduces 8 integration methods for AI Agents, explaining their purpose and how they can be used across systems (steps provided only where applicable).
1. Calling an AI Agent Inside Another AI Agent (Run an Agent)
“Run an Agent” allows one AI Agent to call another, enabling a collaborative, multi-agent architecture.
The primary agent handles the overall logic, while the secondary agent performs domain-specific tasks.
Typical scenarios include:
Breaking down complex tasks (e.g., contract → risk → clause extraction)
Modularizing AI capabilities for better maintainability
Cross-domain collaboration (inventory + procurement + supply chain)
Reusing existing agents for summarization, extraction, or classification
This capability is the foundation of multi-agent systems.
How to Use
Open the primary AI Agent
Go to Tools
Click Add a tool → Run an agent
Select the secondary AI Agent to be called
Configure Input Mapping (AI dynamic fill, variables, or fixed values)
Configure Completion Settings (wait for response or not)
Save and publish
2. Calling an AI Agent in an Approval Form (Form Action)
Approval Forms can invoke AI Agents during form submission, allowing AI to automatically analyze content and generate decisions.
Typical scenarios include:
Auto-generating approval comments
Extracting structured fields (amount, date, project name)
Validating user inputs
Determining subsequent workflow routing
This gives forms “intelligent processing power,” reducing manual review.
How to Use
Open the form designer
Go to Actions
Edit an existing action (e.g., Submit) or create a new one
Click Add Step → AI assistant
Set Event Type to Call AI agent
Select the application and target AI Agent
Configure Input (map form fields to AI variables)
Configure Output (write AI results back to form fields)
Save and publish
Calling an AI Agent in Workflow
AI Agents can be used as intelligent decision-making nodes in workflows.
They can:
Determine whether approval is required
Generate summaries, descriptions, or classifications
Validate or extract data
Control workflow paths based on AI results
This allows AI to participate directly in BPM logic.
How to Use
Open the workflow designer
Drag the AI assistant node into the canvas
Set Event Type to Call AI agent
Select the appropriate AI Agent
Configure Input (map workflow variables to AI inputs)
Configure Output (map AI outputs to workflow variables)
Save and publish
3. Calling an AI Agent in Copilot (Copilot Management)
Yeeflow Copilot serves as the natural language interface for users.
When integrated with an AI Agent, users can trigger complex business actions simply by typing a request, such as:
“Show me products that are low in stock.”
“Summarize this contract.”
“Generate a justification for a purchase request.”
Copilot becomes the unified AI entry point across the organization.
How to Use
Go to Application Settings → AI Agent → Copilot management
Select an existing Copilot or create a new one
Open the Tools tab
Click Add a tool → Run an agent
Select the target AI Agent
Configure input and output
Publish the Copilot
Users can now trigger the AI Agent through natural language.
4. Calling an AI Agent in Microsoft Power Automate
AI Agents can be used as intelligent processing steps in Power Automate, enabling cross-system automation with Outlook, Teams, SharePoint, and more.
Typical scenarios include:
Email summarization
Teams message interpretation
AI-based data extraction or validation
Returning AI-generated results into downstream actions
5. Calling an AI Agent in Microsoft Copilot Agent Flow
AI Agents can serve as enterprise intelligence sources for Microsoft Copilot, enabling Copilot to:
Query internal business data
Generate business-specific outputs
Trigger Yeeflow logic and return contextual results
This is key to building enterprise-grade, Copilot-powered workflows.
6. Calling an AI Agent in Zapier Automations
Zapier connects thousands of SaaS tools, enabling AI Agents to act as intelligent processing steps in multi-system workflows.
Example scenarios:
Google Forms → AI extraction → Write to Yeeflow
HubSpot lead creation → AI scoring → Sales follow-up
Shopify order changes → AI risk assessment → Notify team
7. Calling an AI Agent via REST API
REST API provides the highest flexibility, enabling AI Agents to be embedded into any system, including:
ERP / CRM / HR systems
Internal custom platforms
Enterprise portals
AI orchestration platforms
This is ideal for deep system integration and enterprise AI architectures.
(Instructions not included.)
Summary of AI Agent Integration Methods
Integration Method | Primary Purpose | Ideal Use Cases |
Run an Agent | Multi-agent collaboration | Complex task decomposition, modular AI |
Approval Form | Smart form submission | Approval, application, registration |
Workflow | Intelligent workflow logic | BPM automation and branching |
Copilot | Natural language interface | Enterprise AI assistant |
Power Automate | Microsoft ecosystem automation | Emails, Teams, SharePoint |
Copilot Agent Flow | Copilot enterprise intelligence | Conversational business workflows |
Zapier | SaaS automation ecosystem | CRM, e-commerce, form tools |
REST API | Deep system integration | ERP, custom systems, AI platforms |