Skip to main content

AI Assistant step (Form Action)

Use AI Assistant in form actions to automate document review, content analysis, and intelligent decision support.

The AI Assistant step allows you to trigger AI-powered logic directly within a form action.
By adding this step to a form action flow, you can send form data, user inputs, and uploaded files to the AI assistant, receive intelligent responses, and store the results for further use in the form or subsequent actions.

This step supports multiple AI interaction modes and is commonly used for document review, image analysis, content summarization, and calling predefined AI Agents during form processing.

Where to add this step

To add an AI Assistant step to a form action:

  1. Open your form in Form Designer

  2. Go to Form Actions

  3. Click Add a step

  4. Select AI assistant

The AI Assistant step will be added to the current form action flow and executed based on the configured conditions.

Common configuration fields

The AI Assistant step follows the same execution logic as other form action steps and includes the following common fields:

Execution Condition

Defines whether the AI Assistant step should be executed.
If the condition evaluates to false, the step will be skipped unless Continue next step when condition is not met is enabled.

Continue next step when condition is not met

When enabled, the form action flow will continue to the next step even if the execution condition is not satisfied.

Save response to

Specifies the variable used to store the AI-generated result.
The saved output can be referenced by later steps, workflows, or displayed in the form.

Event Type overview

The AI Assistant step supports four event types, each designed for different scenarios:

  • Conversation with Assistant – Generate AI responses based on text input and optional attachments

  • Analyze image content – Extract and analyze information from image files

  • Analyze document content – Read and summarize document files such as PDFs

  • Call AI agent – Invoke a predefined AI Agent with structured inputs and outputs

Event Type: Conversation with Assistant

This event type generates AI responses based on text input and optional attached files.
It is suitable for summarization, content generation, validation logic, and decision support during form actions.

Configuration

  • User message
    Defines the request or task sent to the AI assistant.
    This field typically describes what the AI should review or generate and can reference form variables.

  • Assistant instructions
    Defines the role, scope, and output requirements of the AI assistant.
    It is recommended to clearly specify that the AI should only use information provided in the form or attached files and avoid assumptions.

  • Select input file(s)
    Specifies the files that will be provided to the AI assistant, such as uploaded certificates or supporting documents.
    You can also switch to the expression editor to dynamically reference file variables.

  • Including online research result
    Controls whether the AI assistant is allowed to reference external information.
    For internal or compliance-related scenarios, this option is usually disabled.

  • Save response to
    Defines where the generated response will be stored for later use in the form action flow.

Event Type: Analyze image content

This event type analyzes image files submitted through the form and extracts visible information from the image.

Configuration

  • Image
    Select the image variable to be analyzed, typically an uploaded certificate, license, or scanned document.

  • User message
    Describes what information should be identified or extracted from the image, such as supplier details or validity dates.

  • Assistant instructions
    Controls how the AI interprets and presents image analysis results.
    It is recommended to instruct the AI to only extract information that is clearly visible and avoid inference.

  • Save response to
    Specifies the variable used to store the image analysis result.

  • This event type is suitable for processing scanned documents or photos uploaded through the form.

Event Type: Analyze document content

This event type allows the AI assistant to read and analyze document files such as PDF or Word documents.

Configuration

  • User message
    Defines the analysis objective for the document, such as summarizing key information or reviewing specific sections.

  • Assistant instructions
    Specifies how the AI assistant should interpret the document and format the output.
    Clear instructions help ensure concise and consistent summaries.

  • Select input file(s)
    Select the document files to be analyzed, typically uploaded attachments from the form.
    You can also switch to the expression editor to dynamically reference files or combine multiple attachment variables.

  • Save response to
    Defines where the document analysis result will be stored.

  • This event type is suitable for extracting structured information from formal documents such as supplier certificates, contracts, or compliance materials.

Event Type: Call AI agent

This event type invokes a published AI Agent to perform intelligent processing within a form action.

Unlike general AI conversations, an AI Agent already contains its own persona, instructions, knowledge, tools, and response format. The AI Assistant action is responsible for selecting the AI Agent, providing the required form context, and saving the generated results for later use.

The configuration includes the following sections.

Application

Select the application where the target AI Agent is defined.

Only AI Agents published within the selected application are available for selection.

Select an AI agent

Choose the AI Agent that will process the current form action.

For example, an AI Agent can be used to:

  • Summarize submitted forms

  • Analyze support tickets

  • Classify requests

  • Generate business recommendations

  • Extract structured information

  • Produce reusable business outputs

The AI Agent's behavior is fully determined by its own configuration, including its persona, instructions, knowledge, tools, and output format.

Include current context

When Call AI agent is selected, you can enable Include current context to automatically send form-related context to the selected AI Agent.

Instead of relying only on the input prompt, the AI Agent can generate responses based on the current form and its associated business data.

This is especially useful for scenarios such as ticket analysis, form summarization, request classification, intelligent routing, and business recommendations.

Available context

Current application context

Provides information about the current application and related business resources.

Current user context

Provides information about the current user performing the action, such as the user identity and profile information.

This allows the AI Agent to personalize responses based on the current user when required.

Template and filter variables

Passes template variables and filter values that are currently available in the form.

This enables the AI Agent to use dynamically generated values when processing the request.

List Fields

Provides all field values from the current record or form submission.

These values become part of the context sent to the AI Agent, allowing it to analyze the submitted business data without manually mapping every field.

Tip

Enable only the context required for your scenario. Providing relevant context helps the AI Agent generate more accurate and useful responses.

Input

Use the User message field to describe what you want the AI Agent to accomplish.

The User message defines the objective of the current action, while the selected context provides the business information needed to complete that task.

You can also switch to the Expression Editor to build dynamic prompts using variables, expressions, and field values.

Output

Select one or more form variables to receive the AI Agent's response.

For example:

  • Summary → var_Summary

  • Category ID → var_CategoryID

  • Support Team ID → var_SupportTeamID

Once the response is stored in form variables, it can be reused throughout the application, including:

  • Form fields

  • Workflow actions

  • Approval processes

  • Notification emails

  • Subsequent AI Assistant actions

Example

A common scenario is automatically analyzing a support ticket immediately after it is submitted.

The AI Assistant calls an AI Agent using the submitted ticket information and generates structured outputs such as:

  • Ticket summary

  • Suggested category

  • Recommended support team

These results are saved to form variables and can be used by subsequent workflow actions, approval logic, notifications, or routing rules.

This approach allows organizations to centralize AI logic inside reusable AI Agents while automatically supplying the current form context whenever the action runs.

Best practices & notes

  • Clearly define user messages and assistant instructions to ensure consistent AI output.

  • Use execution conditions to control when AI logic should be triggered.

  • Choose image analysis or document analysis based on the file type being processed.

  • Avoid enabling online research for internal business or compliance scenarios.

  • Test AI Assistant steps in a controlled environment before applying them to production workflows.

Related Articles
Form Actions
Document Recognition Action
AI Agent
Respond to AI Action(Workflow)
AI Assistant action(Workflow Action)
Did this answer your question?