How to Update Ticket Fields Using AI Trigger Orchestration in BoldDesk
AI Trigger Orchestration can automatically update ticket fields based on structured AI responses. By mapping response properties to ticket actions, you can classify tickets, assign priorities, set categories, and automate other ticket updates without manual intervention.
This article explains how to configure response property mapping, select the appropriate ticket context, test your configuration, and troubleshoot common mapping issues.
Prerequisites
Before configuring AI Trigger Orchestration, ensure that:
- BoldDesk AI features are enabled for your account.
- Your account has sufficient AI credits.
- At least one AI Agent is created and published.
- The AI Agent is configured with instructions appropriate for the intended ticket analysis.
- You have identified the ticket context fields required by the AI Agent.
- You have identified the ticket fields that the AI Agent should update.
- Your role has the Manage Settings permission.
- The target ticket fields contain the values that the AI Agent is expected to use.
Use response property mapping when the AI Agent should automatically update ticket fields. If the AI output is intended only for agent review, retain the default behavior that adds the response as a private note on the ticket.
Understand How Response Property Mapping Works
When an AI Trigger Orchestration rule runs:
- BoldDesk evaluates the rule conditions.
- BoldDesk invokes the selected AI Agent.
- The selected ticket context fields are sent to the AI Agent.
- The AI Agent analyzes the ticket content and instructions.
- The AI Agent returns a structured response.
- BoldDesk extracts the configured response properties.
- BoldDesk maps each property to its corresponding ticket field.
- The mapped ticket fields are updated.
For example, suppose the orchestration rule contains the following mappings:
| Ticket Action | Response Property |
|---|---|
| Set Type | type |
| Set Category | category |
| Set Priority | priority |
The AI Agent should provide values for those properties so BoldDesk can apply them to the corresponding ticket fields.
The response property names used by the AI Agent must match the property names configured in the orchestration rule. For example, priority and ticket_priority are treated as different properties.
Configure the Rule Conditions
Create an AI Trigger Orchestration rule and configure the ticket conditions that should invoke the AI Agent.
The AI Agent is invoked only when a ticket meets the configured conditions. Use conditions that limit the rule to tickets relevant to the intended workflow.
For detailed instructions on creating the rule and configuring its conditions, learn How to Configure AI Trigger Orchestration in BoldDesk.
After configuring the conditions, click Next to configure the AI Agent invocation and response property mapping.
Use specific conditions to prevent the AI Agent from processing tickets outside the intended workflow.
Configure the AI Agent Invocation
In the Invoke AI Agent step:
-
In Choose AI Agent, select the published AI Agent that should process matching tickets.
-
Under Ticket Context Fields, select the ticket information that the AI Agent needs.
-
In Additional Instructions, define the analysis requirements and expected output.
-
Enable Update Ticket Fields from AI Responses.
-
Map each ticket action to the corresponding response property.
-
Click Add or Update to save the orchestration rule.
By default, the AI Agent response is added to the ticket as a private note. When Update Ticket Fields from AI Responses is enabled, the response is used to update the mapped ticket fields instead, and a private note is not created.
Select the Required Ticket Context Fields
The Ticket Context Fields setting determines which ticket information is sent to the AI Agent.
Select only the fields that are relevant to the analysis.
For example, an AI Agent that classifies new tickets may require:
- Subject
- Description Text
- Brand Name
- Category ID
- Priority ID
- Type ID
An AI Agent that analyzes customer replies may also require:
- Last Message Text
- Last Message ID
- Ticket ID
Providing too little context can lead to incomplete classifications. Providing unrelated context can reduce response quality and consume additional AI credits.
Configure AI Response Instructions
The AI Agent should be instructed to analyze ticket information and provide values for the response properties used in the orchestration rule. Include guidance such as:
- Use the configured response property names.
- Return values supported by the mapped ticket fields.
- Do not invent field values.
- Use approved fallback values when information is insufficient.
- Ensure all required properties are included.
Example instruction:
Analyze the ticket using the provided subject, description, latest message, and current ticket properties.
Provide values for:
- type
- category
- priority
Use only values that are valid for the corresponding BoldDesk ticket fields.
Do not invent field values.
Use approved fallback values when sufficient information is unavailable.
Ensure that the property names referenced in the instructions match the property names configured in the orchestration rule.
Apply the Agent’s Response to Ticket Fields
After checking Apply the Agent’s Response to Ticket Fields, configure a mapping for each ticket field that should be updated. To configure property mapping:
- Select a ticket action.
- Enter the response property that contains the corresponding value.
- Add additional mappings as required.
- Remove unused mappings.
- Save or update the rule.
Example mapping:
| Ticket Action | Response Property |
|---|---|
| Set Type | type |
| Set Category | category |
| Set Priority | priority |
Use consistent property names throughout the configuration.
Use short and descriptive property names such as type, category, priority, status, or tags.
Verify the Configuration
After saving the orchestration rule:
- Create a test ticket that matches the rule conditions.
- Confirm that the AI Agent executes successfully.
- Verify that the mapped ticket fields are updated.
- Confirm that a private note is not created when ticket field updates are enabled.
Example to Configure Ticket Classification for Battery-Related Issues
Suppose your support team wants an AI Agent to classify battery-related tickets and automatically update the ticket type, category, and priority.
Configure the Rule Condition
Configure the orchestration rule to match battery-related tickets.
For example:
- Field: Subject
- Operator: Contains
- Values:
- Battery
- Battery drain
- Battery health
- Battery swollen
- Battery overheating
Select the AI Agent
In Choose AI Agent, select the published AI Agent configured for hardware issue analysis.
Select the Ticket Context Fields
Select:
- Brand Name
- Ticket ID
- Subject
- Description Text
- Last Message Text
- Type ID
- Category ID
- Priority ID
Configure Additional Instructions
Analyze the ticket for battery-related issues using the provided ticket details.
Classify the issue based on the reported symptoms.
Provide values for the following properties:
- type
- category
- priority
Use only values that are valid for the corresponding BoldDesk ticket fields.
If sufficient information is unavailable, use the approved fallback values.
Do not invent customer, warranty, cost, or device information.
Configure the Ticket Field Mappings
Enable Update Ticket Fields from AI Responses and configure:
| Ticket Action | Response Property |
|---|---|
| Set Type | type |
| Set Category | category |
| Set Priority | priority |
Expected Classification Result
| Property | Value |
|---|---|
type |
Problem |
category |
Hardware |
priority |
High |
BoldDesk reads the configured properties and updates the mapped ticket fields accordingly.
When Update Ticket Fields from AI Responses is enabled, only the mapped ticket fields are updated. The AI response is not added as a private note.
Handle Missing Information
Define how the AI Agent should respond when there is not enough information to make a reliable classification. Consider the following approaches:
- Use approved fallback values.
- Route tickets to a general category for manual review.
- Apply a default priority based on your organization’s triage policy.
- Configure rule conditions that require sufficient ticket details before processing.
Example instruction:
If the ticket does not contain enough information to determine the appropriate classification, use the approved fallback values for unclassified tickets.
Ensure that all fallback values are supported by the target ticket fields.
Understand Private Note and Ticket Update Behavior
AI Trigger Orchestration supports two response-handling behaviors.
Default Behavior
When Update Ticket Fields from AI Responses is disabled:
- The AI Agent response is added as a private note.
- Agents can review the AI-generated findings.
- Ticket fields remain unchanged.
Ticket Update Behavior
When Update Ticket Fields from AI Responses is enabled:
- BoldDesk reads the configured response properties.
- The mapped ticket fields are updated automatically.
- A private note is not added.
Use the default behavior when agents should review the AI output. Use ticket field updates when the AI response should drive workflow automation.
Security and Data Considerations
Send only the information necessary for the AI Agent’s task. Avoid including unnecessary;
- Customer identifiers
- Contact information
- Internal-only data
- Sensitive custom-field values
- Authentication details
- Confidential business information
Review the target field mappings regularly to ensure updates remain aligned with your workflow requirements.
Do not include passwords, API keys, access tokens, or other credentials in AI instructions or ticket context fields.
Troubleshooting
The AI Response Is Added as a Private Note
This occurs when Update Ticket Fields from AI Responses is disabled. Enable the option and configure at least one response property mapping if the AI output should update ticket fields.
Some Ticket Fields Are Updated but Others Are Not
Verify that:
- Each required property is mapped in the orchestration rule.
- The AI Agent provides values for all mapped properties.
- Returned values exist in the corresponding BoldDesk ticket fields.
The AI Agent Returns an Unsupported Field Value
List the permitted values in the AI Agent instructions. Keep the instructions synchronized with the available values in the ticket field configuration. Example:
For priority, use only Low, Medium, High, or Urgent.
A Required Property Is Missing
Specify that all mapped properties must always be provided and define approved fallback values for situations where information is insufficient.
The AI Output Uses Different Property Names
Ensure that the property names referenced in the AI instructions match the property names configured in the orchestration mappings. For example, if the configured property is category, do not instruct the AI Agent to use ticketCategory.
Frequently Asked Questions
-
Why use response property mapping in AI Trigger Orchestration?
Response property mapping enables BoldDesk to update ticket fields automatically using AI Agent output, supporting automated classification and triage workflows. -
Do AI response property names have to match the orchestration rule mapping?
Yes. BoldDesk treats property names as exact identifiers.priorityandticket_priorityare different properties. -
Can one AI Trigger Orchestration rule update multiple ticket fields?
Yes. A single orchestration rule can map multiple ticket actions to multiple response properties. -
What happens if the AI Agent returns a value that is not available in the ticket field options?
The update may not be applied for that field. Instruct the AI Agent to use only values that exist in the target ticket field configuration. -
Should I keep the default private note behavior or enable ticket field updates?
Use private notes when agents must review AI output before acting. Enable ticket field updates when AI output should drive automated routing or classification. -
What ticket context fields should I send to the AI Agent?
Send only the ticket fields required for the intended analysis (for example, Subject, Description Text, and existing Type/Priority IDs). Avoid unrelated fields to reduce AI credit usage and improve response quality. -
How do I handle tickets that do not have enough information to classify?
Define fallback values in AI Agent instructions and ensure the fallback values exist in the target ticket fields.