How to Enable and Configure AI Field Prediction in BoldDesk

AI Field Prediction in BoldDesk automatically classifies tickets by predicting and setting supported picklist fields based on ticket content. Predictions are applied through Automation Rules and operate strictly within configured brands and sources, helping teams improve routing accuracy and reduce manual updates.
AI Field Prediction analyzes a ticket’s Subject, Description, and new replies to predict values for supported fields such as Category, Type, and eligible custom dropdown fields. Predictions are applied only through Create or Update (new message) automation triggers and never overwrite existing field values.

This feature is part of AI Copilot and integrates directly with BoldDesk’s Ticket Automation framework.

Prerequisites

This feature is available exclusively on the Enterprise plan.
AI Copilot is available in your account.
You can access Admin → AI Copilot and Admin → Automation → Rules.

Below are the scope & supported fields

System fields: Category, Type
Custom fields: Single‑select dropdown, Multi‑select dropdown
Events covered: Create Ticket Triggers, Update (with a new message) Ticket Triggers
Not supported: Text fields, numeric fields, date fields, checkbox fields, tags, and priority (unless your environment explicitly supports these via custom dropdowns).

AI predictions are utilized only when the target field is empty during evaluation, and the AI Copilot is enabled. For further information, explore How to Enable AI Copilot Features in BoldDesk.
The Create Ticket trigger populates AI-assisted Category and Type values only when the AI prediction confidence score is greater than 50%. If the confidence score falls below this threshold, the prediction is not stored and the trigger will not assign a Category or Type value. Low confidence scores typically occur when ticket content contains unclear classification keywords or when multiple categories have similar relevance (for example, Email and Admin).

How to Enable AI Field Prediction

Navigate to AI Center → AI Copilot
Enable AI Copilot Features
Turn on Field Prediction

Once enabled, AI Predict becomes available as an action in Automation rules.

Manage AI Field Prediction Settings

BoldDesk provides centralized controls to manage how and where AI Field Prediction runs.

Navigation Path
Field Prediction → Manage

The available controls that can be customized include brand and source as discussed below.

Brand Restrictions

Select the brands where AI Field Prediction should apply.
Predictions run only within the selected brands.
Cross‑brand prediction is not supported.

Source Restrictions

Specify which ticket sources are eligible.
Predictions run only when the ticket source matches the configuration.

Use brand and source restrictions to control rollout and minimize unintended automation.

Configuring Automation to Apply Predicted Values

AI Field Prediction is applied through Automation Rules, enabling predicted values to be automatically set on tickets based on defined conditions.
Navigation Path: Admin → Automation → Rules

Choose a Trigger
Select the appropriate trigger based on when you want the prediction to run:
- Create Ticket Trigger
  Runs the prediction automatically when a ticket is created, allowing fields to be populated at the time of submission.
- Update Ticket Trigger
  Runs the prediction when a customer or agent edits ticket properties, ensuring field values stay updated based on the latest interaction.
Configure Rule Actions
- Define required conditions (Brand, Source, Group, etc.)
- In Set Action, choose one of the following:
  - Set Category → AI Predict
  - Set Type → AI Predict
  - Set Custom Dropdown → AI Predict
  - {{Select.All}} to apply prediction across all supported fields
- Add additional actions if required
- Click Save or Update

Use one AI Predict action per field unless using {{Select.All}} to keep Automation History easy to audit.

Recommended rule design

Use one action per field to keep outcomes auditable.
Scope with conditions (e.g., Brand equals…, Group equals…) if you want controlled rollout.
Avoid conflicting actions that set the same field in the same rule.

Understanding Field Dependency with AI Predict and {{Select.All}}

Ensuring accurate data entry in BoldDesk is essential, especially when using AI-powered field prediction. The way field dependencies are managed depends on how fields are configured within your automation rules. Mastering this distinction helps maintain valid, consistent ticket information.

1. Using {{Select.All}} for Seamless AI Prediction (Recommended)

When you choose the {{Select.All}} option in an automation rule:

AI Predict evaluates and fills all supported fields collectively.
Field dependencies are automatically validated, guaranteeing data consistency.
Relationships between fields (such as Category → Type) are resolved before values are applied.

Result:
You get accurate, reliable ticket data that adheres to dependency requirements, streamlining ticket management and improving support workflows.

2. Manual Field Selection (Without {{Select.All}})

When fields are selected individually:

AI Predict processes each field separately, without considering inter-field dependencies.
Field dependency validation is not enforced, which may lead to mismatched or invalid combinations.

Result:
Manual selection can violate dependency rules, resulting in inconsistent or erroneous ticket information.

Example of Field Dependency in Action

Dependency Configuration:

Category: Hardware / Software
Type (dependent field):
- If Category = Hardware → Options: Laptop, Monitor
- If Category = Software → Options: Bug, Feature Request

Case 1. Using {{Select.All}}

Scenario:
Ticket Content: “My laptop screen is flickering.”

AI Prediction:

Evaluates all fields in context
Sets:
- Category = Hardware
- Type = Laptop (valid dependent value)
  Result:
  Both fields are set correctly, respecting dependency rules.

Case 2. Manual Field Selection (Without {{Select.All}})

Automation Rule:

Set Category → AI Predict
Set Type → AI Predict

AI Prediction:

Processes fields independently:
- Category = Hardware
- Type = Bug (triggered by keywords like “flickering issue”)

Result:
This creates an invalid combination (Type = Bug is not valid for Category = Hardware), causing a dependency mismatch.

Always use {{Select.All}} when working with dependent fields to ensure accurate validation and reliable data population.
Use manual field selection only if fields are independent and do not rely on other field values.

How and When AI Predictions Run in BoldDesk

AI predictions in BoldDesk are triggered only when a ticket receives a new message, such as when a ticket is created, or when a requester or agent replies. Updates that only change ticket properties (like status, priority, tags, or custom fields) without a new message do not trigger predictions.
At runtime, here’s how predictions work:

When a new message is added (e.g., Ticket Created, Requester Reply, or Agent Reply), automation rules are evaluated.
If your automation rule includes AI Predict for a supported field, BoldDesk sends the ticket’s subject, description, and latest replies to the AI.
The AI returns predicted value(s) with an internal confidence score and language context.
If the target field is empty and the prediction’s confidence score meets the required threshold (e.g., ≥ 0.60), BoldDesk updates the field automatically.
The change is recorded in Automation History as an AI Predict action.

The confidence threshold of predicted value(s) is system‑defined. Use Automation History to verify whether a prediction was applied or skipped.

Support Matrix (Quick Reference)

Use the following table to confirm which fields are eligible for AI prediction and how automation applies them at runtime:

Item	Behavior
Fields	Category, Type, dropdown fields
Triggers	Create, Update (new message only)
Brand scope	Restricted
Source scope	Restricted
Overwrite	Applies only if empty
Select All	Supported (`{{Select.All}}`)
Audit	Automation History

Frequently Asked Questions

Which events trigger ticket field prediction in BoldDesk?
Ticket Created and new Reply events. Property‑only updates (e.g., status change) do not trigger predictions.
Can I use Field Prediction with custom fields in BoldDesk?
Yes. Single‑select and multi‑select dropdown custom fields are supported.
Where do I enable Field Prediction in BoldDesk?
Turn it on in Admin → AI Copilot. Then add automation actions Set → AI Predict.
How do I confirm a field prediction was applied to a ticket?
Open the ticket’s Automation History. You’ll see an entry marked AI Predict (or Applied via AI Prediction).
How can I automatically add values into ticket fields?
- Enable Field Prediction in Admin → AI Copilot.
- Go to Admin → Automation → Triggers (Create or Update) → choose Add Rule or edit an existing rule.
- On Set Action, choose Set → AI Predict for each picklist. The field selected must be either Category, Type or supported custom dropdowns.
- Runs across email/portal/chat‑to‑ticket/integrations; applies only if the target field is empty; brand‑scoped; logged in Automation History.
How to classify the Category and Type of a ticket automatically?
After enabling Field Prediction in Admin → AI Copilot, add two separate actions in your automation rule, either within the Create Trigger or Update Trigger:
- On Set Action: Set Category → AI Predict and Set Type → AI Predict.
- Uses Subject, Description, latest Replies; no overwrite if values exist; runs on Create or Update (new message); brand‑scoped and audited in Automation History.

Was this article useful?

Dislike

Comments (0)