Skip to main content

Access application resources

Enables AI to securely access selected application resources, such as forms, data lists, reports, document libraries, and AI Agents, with configurable permissions and safety controls.

The Access application resources tool allows AI Agents and Copilots to securely access resources within selected Yeeflow applications. Administrators can configure which application resources the AI is allowed to access, along with the permitted operations for each resource type.

Using this tool, AI can interact with approval forms, data lists, document libraries, reports, and even other AI Agents, while maintaining strict permission control and operational safety.


Add the Access application resources Tool

In the Tools section of an AI Agent or Copilot:

  1. Click Add a tool

  2. Select Access application resources

A configuration window will appear where you can:

  • Select a Workspace

  • Select an Application

  • Enter a Tool name

After configuration, click:

  • Add

  • or Add and configure


Detail

In the Detail section, you can:

  • Modify the tool name

  • Enter a tool description

Providing a clear description helps the AI better understand the purpose of the tool and improves response accuracy.


Define resource access scope

Under Define resource access scope, you can configure AI access permissions for the following application resources:

  • Approval Forms

  • Data Lists

  • Document Libraries

  • Form Reports

  • Data Reports

  • AI Agents

Each resource type can be configured independently with different access scopes and operation permissions.


Approval Forms

Expand the Approval Forms section to configure approval form access permissions.

Access Scope

Three access scope options are available:

Access Scope

Description

No access

The AI cannot access any approval forms

All approval forms

The AI can access all approval forms in the selected application

Selected items

The AI can access only selected approval forms

The default setting is No access.

Available Operations

Approval forms support the following operations:

  • Submit

  • Read

  • Read tasks

  • Process tasks

All approval forms

Select the All approval forms tab to configure permissions for all approval forms in the application.

Selected items

Select the Selected items tab to configure access for specific published approval forms only.

To add approval forms:

  1. Click Add approval form

  2. Select the required published approval form

  3. Configure the allowed operations for the form


Data Lists / Document Libraries

Both Data Lists and Document Libraries support the same operation permissions:

  • Read

  • Add

  • Edit

  • Delete

Under the Selected items tab:

  1. Select the required data lists or document libraries

  2. Configure the allowed operations

These permissions allow the AI to read, create, modify, or delete records and documents within the selected resources.


Form Reports / Data Reports

Both Form Reports and Data Reports support the same operation permission:

  • Read

This allows the AI to retrieve and summarize report data for analysis, statistics, and information queries.


AI Agent

The AI Agent resource type allows one AI tool to call another AI Agent.

Available Operation

  • Call

This capability enables collaboration between multiple AI Agents for more advanced automation scenarios.


Safety & control

The Safety & control section is available only in Copilot.

Administrators can define how user confirmations should be handled based on the risk level of operations.

Risk Levels

Low-risk actions

Examples include:

  • Viewing records

  • Searching data

  • Reading information

  • Summarizing records

  • Viewing tasks

  • Checking workflow status

Standard actions

Examples include:

  • Creating records

  • Updating records

  • Submitting requests

  • Processing tasks

High-risk actions

Examples include:

  • Deleting records

Confirmation Methods

Confirmation Method

Description

No confirmation required

No user confirmation is needed. Default for low-risk actions

Require confirmation

Users must click a confirmation button during the conversation. Default for standard actions

Require strong confirmation

Users must click a highly visible confirmation button during the conversation. Default for high-risk actions

These confirmation mechanisms help reduce operational risks and prevent unintended actions.


Completion

In the Completion section, you can configure the behavior and output after the tool execution is completed.

The system provides the following fixed output variables:

  • Data

  • Status code

These output variables can be used for subsequent AI processing and workflow automation.

Did this answer your question?