Category

How to Update Resolution Due Based on Customer Commitments Using AI in BoldDesk

Updated:

Support agents often make customer-facing commitments such as providing an update within a specific timeframe, resolving an issue by a certain date, or following up at an agreed time. To help ensure that ticket due dates reflect these commitments, you can use the Commitment Timeline Validation AI Agent together with AI Trigger Orchestration.

When a human support agent posts a public response, the AI Agent analyzes the latest customer-facing message, identifies the next valid commitment, and automatically updates the ticket’s Resolution Due field through AI Trigger Orchestration response mapping. The workflow can also add a private note that records the detected commitment, the previous Resolution Due value, the new Resolution Due value, and the validation result.

This approach helps align ticket due dates with commitments communicated to customers, improving visibility and accountability for promised follow-ups and resolutions without requiring API Actions or MCP tools.

  • In BoldDesk, Resolution Due typically represents the SLA target for resolving or closing a ticket.
  • Use this workflow only when customer-facing commitments are intended to supersede the SLA-calculated Resolution Due value.
  • If you want to track promised follow-ups separately from SLA commitments, map the AI response to a custom DateTime field such as Next Promised Update instead of Resolution Due.

How the Workflow Works

  1. A support agent publishes a customer-facing response that contains a commitment.
  2. The ticket satisfies the configured AI Trigger Orchestration conditions.
  3. BoldDesk invokes the published Commitment Timeline Validation AI Agent.
  4. The AI Agent evaluates the content of the Last Message Text field.
  5. The AI Agent returns:
    • resolutionDue
    • private_note
  6. BoldDesk updates the ticket’s Resolution Due field and adds the validation note.
  7. The execution details and field updates are recorded in the ticket history.

Example

Assume the current Resolution Due value is:

2026-11-07T10:32:00Z

The support agent posts the following public response:

We will provide the next update on 2026-11-06T10:32:00Z.

The expected result is:

Item Result
Previous Resolution Due 2026-11-07T10:32:00Z
Detected Commitment 2026-11-06T10:32:00Z
New Resolution Due 2026-11-06T10:32:00Z
Validation Result Updated

In this example, the AI Agent identifies the commitment date in the agent’s response and updates the ticket’s Resolution Due field to match the promised customer update timeline.

Before You Begin

Prerequisites

Before configuring this workflow, ensure that the following requirements are met:

  • BoldDesk AI 2.0 is enabled for your account.
  • Sufficient AI Credits are available to process AI Agent executions.
  • You have AI Agent Builder Access: Full Access permissions.
  • You have permission to create and manage AI Trigger Orchestration workflows.
  • A test ticket is available for validation and troubleshooting.
  • Your organization has established guidelines for interpreting commitments, time zones, business hours, relative dates, and ambiguous timelines.

Define Commitment Interpretation Rules

Before enabling the workflow, establish clear and consistent rules for how customer-facing commitments should be interpreted. This helps ensure that the AI Agent produces predictable and reliable results.

The following examples illustrate common commitment expressions and recommended handling approaches:

Commitment Expression Recommended Handling
Within two hours Add two hours to the timestamp of the public response.
Within two to three hours Use the latest boundary or another consistently defined rule.
By Friday Use the next applicable Friday and a predefined cutoff time.
Tomorrow Use the next calendar day and a predefined time.
Within two business days Calculate the date using the organization’s business calendar.
Next week Map the phrase to a specific weekday and time.
Soon, shortly, or later Treat as ambiguous and require manual review.
After migration completes Require a confirmed completion timestamp or route for manual review.

In addition, define the following organizational policies:

  • The default time zone used for commitment calculations.
  • Whether durations are based on calendar time or business hours.
  • How weekends and public holidays are handled.
  • How date and time ranges are interpreted.
  • How commitments that resolve to a past date or time are handled.
  • Whether the Resolution Due value can move earlier, later, or both.

The examples in this article assume a policy where every valid customer-facing commitment replaces the existing Resolution Due value, regardless of whether the new commitment is earlier or later than the current due date.

Configure the Commitment Validation Workflow

Step 1: Create the AI Agent

Create an AI Agent named Commitment Timeline Validation.

Configure the AI Agent to:

  • Analyze the latest public response posted by a human support agent.
  • Identify the next valid customer-facing commitment.
  • Convert the commitment into an ISO 8601 UTC timestamp.
  • Generate a private validation note that explains the outcome.
  • Avoid modifying unrelated ticket fields.

For detailed instructions, see How to Create an AI 2.0 Agent in BoldDesk.

Step 2: Configure the AI Agent Instructions

Add the following instructions to the AI Agent:

Role

You are a Commitment Timeline Validation AI Agent.

Analyze the latest public response added by a human support agent, identify
the next reliable customer-facing commitment, and return the commitment date
and time for the ticket's Resolution Due field.

Scope

Evaluate future commitments to provide an:

- Update
- Follow-up
- Result
- Decision
- Fix
- Resolution
- Status report
- Confirmation

Message Selection

1. Analyze Last Message Text only.
2. Last Message Text contains the latest public response added by a human
   support agent.
3. Use Resolution Due Date only as the previous Resolution Due value.
4. Ignore customer messages, private notes, AI-generated responses, system
   activities, signatures, forwarded content, and quoted earlier messages.

Commitment Detection

1. A valid commitment must promise a future customer-facing action.
2. If multiple reliable commitments are present, select the earliest next
   customer-facing commitment.
3. Ignore historical dates, customer-provided dates, quoted dates, log
   timestamps, and release dates that are not support commitments.
4. Treat expressions such as "soon," "shortly," "later," and "as soon as
   possible" as Ambiguous.
5. Do not invent a date for an ambiguous commitment.

Date and Time Calculation

1. Use an ISO 8601 timestamp exactly as supplied.
2. For a specific date and time, use the stated time zone.
3. For a relative duration, calculate from the latest public human-agent
   response timestamp.
4. For a time range, use the latest boundary.
5. If no recognized time zone is stated, use the portal time zone.
6. Convert the result to ISO 8601 UTC format.

Resolution Due Policy

1. If a reliable commitment is detected, set resolutionDue to the calculated
   commitment timestamp.
2. Replace the current Resolution Due whether the commitment is earlier or
   later than the existing value.
3. Do not retain the existing Resolution Due merely because it is earlier or
   later than the commitment.
4. Return No Change only when both timestamps are exactly equal.
5. If no reliable commitment is detected, return the current Resolution Due.
6. If the commitment is ambiguous, return the current Resolution Due.
7. Do not change unrelated ticket fields.

Private Note

Set private_note to include:

- Validation Result: Updated, No Change, Ambiguous, or No Commitment
- Commitment Detected
- Previous Resolution Due
- New Resolution Due
- Reason

Use Validation Result: Updated whenever a reliable commitment differs from
the current Resolution Due.

Output Requirements

1. Use only these exact properties:
   - resolutionDue
   - private_note
2. Return resolutionDue as an ISO 8601 UTC timestamp.
3. Return private_note as plain text.
4. Return one flat JSON object.
5. Do not return Markdown, code blocks, nested JSON, or additional properties.
  • The response property names must match the orchestration mappings exactly.
  • Do not use alternative property names such as new_resolution_due, resolution_due, ResolutionDue, or newResolutionDue.

Step 3: Test and Publish the AI Agent

Before publishing the AI Agent, validate it against representative scenarios, including:

  • A commitment containing a specific date and time.
  • A commitment expressed as a relative duration.
  • A commitment containing a time range.
  • Multiple commitments in a single response.
  • An ambiguous commitment.
  • A response that contains no commitment.

Verify that:

  • resolutionDue returns a valid ISO 8601 UTC timestamp.
  • private_note returns plain-text output.
  • Ambiguous or missing commitments retain the existing Resolution Due value.
  • The AI Agent consistently selects the correct commitment according to your defined policy.

After successful testing, publish the validated AI Agent version.

For detailed guidance, see How to Test an AI 2.0 Agent in BoldDesk.

Step 4: Configure AI Trigger Orchestration

Create an AI Trigger Orchestration rule that runs when the target public response is added by a human support agent.

Use a descriptive rule name, such as:

Validate Customer Commitment Timeline

Configure the orchestration rule to:

  • Run for the relevant public agent-response event.
  • Apply only to the required brands, groups, forms, or workflows.
  • Exclude tickets and activities that are not relevant to commitment validation.
  • Invoke the published Commitment Timeline Validation AI Agent.
  • Prevent duplicate execution for the same response whenever possible.

For detailed configuration instructions, see How to Configure AI Trigger Orchestration in BoldDesk.

Step 5: Select the Ticket Context

Include only the ticket context necessary for commitment validation.

Context Field Recommendation Purpose
Ticket ID Include Identifies the ticket being processed.
Last Message Text Include Provides the latest public response from the human support agent.
Resolution Due Date Include Supplies the existing Resolution Due value for comparison and reporting.
Last Modified Time Include only when required Provides a reference point for calculating relative commitments.
Ticket Messages Exclude May include customer responses, historical dates, quoted content, and previous AI-generated notes.
Last Public Note Text Exclude Not required when evaluating a public reply.
Description Text Exclude unless required May introduce unrelated dates and historical context.
Response Due Date Exclude Not required for this workflow.
  • Do not include Ticket Messages in the AI Agent context.
  • Historical ticket conversations may contain earlier commitments, customer-provided dates, or previous AI-generated validation notes that can influence the AI Agent’s decision.
  • Use Last Message Text as the authoritative source for commitment detection.

Step 6: Configure Additional Instructions

Add the following text under Additional Instructions:

Evaluate Last Message Text as the latest public response added by a human
support agent.

If Last Message Text contains a reliable future commitment, return that
commitment in resolutionDue.

Replace the current Resolution Due whether the commitment is earlier or
later than the existing value.

Do not retain the current Resolution Due because it is earlier or later than
the commitment.

Use Resolution Due Date only as the previous value reported in private_note.

If no reliable commitment exists or the commitment is ambiguous, return the
current Resolution Due unchanged.

Use only the exact mapped properties:

- resolutionDue
- private_note

Step 7: Configure Response Mapping

Configure the response mappings as follows:

Ticket Action AI Response Property
Set Resolution Due resolutionDue
Add Private Note private_note

Confirm the following before enabling the orchestration:

  • Property names match the AI Agent output exactly.
  • resolutionDue returns a valid ISO 8601 UTC timestamp.
  • private_note returns plain-text content.
  • The response structure is a flat JSON object without nested properties.

For detailed mapping instructions, see How to Update Ticket Fields Using AI Trigger Orchestration in BoldDesk.

When a human support agent publishes a response containing a valid customer-facing commitment, the AI Agent evaluates the commitment, calculates the corresponding date and time, updates the ticket’s Resolution Due field, and adds a private validation note. If no reliable commitment is detected, the existing Resolution Due value remains unchanged.

Test and Verify the Workflow

After configuring the AI Agent, orchestration rule, and response mappings, validate the workflow to ensure commitments are detected accurately and the ticket’s Resolution Due value behaves as expected.

Test the Workflow

Use a test ticket that does not contain previous commitment-validation notes or historical AI-generated validation results.

Scenario Public Agent Response Expected Result
Earlier commitment We will update you on 2026-11-06T10:32:00Z. Resolution Due is updated to November 6, 2026.
Later commitment We will update you on 2026-11-15T10:32:00Z. Resolution Due is updated to November 15, 2026.
Same commitment The commitment matches the current Resolution Due. No effective change is made to the field.
Ambiguous commitment We will update you soon. The existing Resolution Due value is retained.
No commitment We are currently reviewing the logs. The existing Resolution Due value is retained.

Verify that:

  • The intended AI Agent is invoked.
  • The correct customer-facing commitment is identified.
  • Resolution Due is updated when the detected commitment differs from the existing value.
  • A validation note is added to the ticket.
  • No unrelated ticket fields are modified.
  • The orchestration rule does not execute repeatedly for the same activity.

Review Ticket Messages

Review the ticket conversation and confirm that the generated private note includes:

  • Validation Result
  • Commitment Detected
  • Previous Resolution Due
  • New Resolution Due
  • Reason

Depending on the configuration, BoldDesk may also add a system-generated activity indicating that the Resolution Due value was updated.

Review Ticket History

Open the ticket’s History tab and verify that:

  • The AI Agent execution completed successfully.
  • Resolution Due changed from the previous value to the detected commitment date and time.
  • No subsequent automation, workflow, or SLA process modified the field unexpectedly.

Review AI Agent Activity

Open the ticket’s AI Agents tab and review the execution details, including:

  • AI Agent name
  • Invocation status
  • Associated orchestration rule
  • Returned response payload
  • Generated private note
  • Resulting Resolution Due update

For more information, see How to Review AI Agent Activity in a BoldDesk Ticket.

Resolution Due and SLA Considerations

In most environments, Resolution Due is calculated automatically based on the applicable SLA policy and its configured Time to Resolution target.

Depending on your SLA configuration, the calculated due date may consider:

  • Calendar hours or business hours
  • Business-hour schedules and time zones
  • Weekends and holidays
  • Time spent in Hold statuses
  • The SLA evaluation start time

When this workflow runs, the AI Agent updates the existing Resolution Due value using the commitment detected in the agent’s public response.

Updating Resolution Due can affect:

  • SLA breach calculations
  • Overdue ticket status
  • SLA reminders and escalations
  • Dashboard and reporting metrics
  • Resolution-performance measurements
    Before enabling this workflow, confirm that customer-facing commitments are intended to override the SLA-calculated Resolution Due value.

Be aware that certain SLA-related events may recalculate the ticket’s Resolution Due after the AI Agent updates it. Examples include:

  • Reopening a ticket
  • Moving a ticket from Hold to an active status
  • Updating fields that participate in SLA conditions
  • Applying a different SLA policy

Review your SLA configuration thoroughly to understand how these events may interact with this workflow.

Troubleshooting

Issue Resolution
AI Agent is unavailable in orchestration Verify that the AI Agent has been successfully published.
The orchestration rule does not run Confirm that the ticket meets the rule conditions and that sufficient AI Credits are available.
Resolution Due does not change Verify that resolutionDue is mapped correctly, differs from the current value, and is returned in ISO 8601 UTC format.
The correct commitment is detected, but the existing value is retained Exclude Ticket Messages and previous AI-generated notes from the context. Use Last Message Text as the primary commitment source.
The AI Agent repeats a previous decision Test with a clean ticket and confirm that Ticket Messages are excluded from the orchestration context.
The wrong message is evaluated Remove unnecessary context fields and verify that Last Message Text contains the latest public response from the support agent.
An incorrect date or time is calculated Review your commitment rules, response timestamp, time-zone configuration, business-calendar rules, and time-range handling logic.
Ambiguous commitments unexpectedly change the due date Update the AI Agent instructions to retain the existing value whenever a reliable commitment cannot be calculated.
The private note is not added Confirm that private_note is mapped to Add Private Note in the response mapping configuration.
Resolution Due changes and then reverts Review the ticket’s History tab for SLA recalculations, status changes, workflow automations, or other processes that may have updated the field afterward.
The rule executes repeatedly Refine the orchestration conditions and, where supported, exclude activities that have already been processed by the workflow.

When a support agent communicates a clear customer-facing commitment, the AI Agent detects the commitment, updates the ticket’s Resolution Due value accordingly, and records the decision in a private validation note. If no reliable commitment is present, the existing Resolution Due value remains unchanged.

Frequently Asked Questions

  1. Is an API Action or MCP Tool Required?
    No. This workflow relies entirely on AI Trigger Orchestration response mapping. The AI Agent returns the required values, and the orchestration rule maps them directly to Set Resolution Due and Add Private Note. No API Action or MCP tool is required.

  2. Why Should Ticket Messages Be Excluded from the Context?
    Ticket Messages can contain historical content such as previous AI-generated validation notes, customer responses, quoted replies, and earlier dates. Including this information can influence the AI Agent’s decision and lead to incorrect commitment detection.

For the most reliable results, use Last Message Text as the primary commitment source and exclude Ticket Messages from the orchestration context.

  1. What Context Should Be Included?
    Include only the information required for commitment validation:

    • Ticket ID
    • Last Message Text
    • Resolution Due Date
    • Last Modified Time (only when needed to calculate relative dates and durations)
      Limiting the context helps improve accuracy and reduces the likelihood of unrelated information affecting the result.
  2. Can Resolution Due Move Earlier or Later?
    Yes. Under the policy described in this article, every reliable customer-facing commitment replaces the current Resolution Due value, regardless of whether the commitment date is earlier or later than the existing value.

  3. What Happens When No Reliable Commitment Is Detected?
    If no reliable commitment is identified, or if the commitment is considered ambiguous, the AI Agent retains the existing Resolution Due value and returns a private note indicating a validation result of No Commitment or Ambiguous.

  4. Can SLA Processing Change Resolution Due After the Workflow Runs?
    Yes. SLA-related events can recalculate Resolution Due after the AI Agent updates it. Examples include:

    • Reopening a ticket
    • Moving a ticket from Hold to an active status
    • Updating fields used in SLA conditions
    • Applying a different SLA policy
      Review your SLA configuration to understand how these events may interact with AI-driven Resolution Due updates.

Related Articles

  1. How to Configure AI Trigger Orchestration in BoldDesk
  2. How to Update Ticket Fields Using AI Trigger Orchestration in BoldDesk
  3. How to Test an AI 2.0 Agent in BoldDesk
  4. How to Review AI Agent Activity in a BoldDesk Ticket
  5. How to Create SLA Policies in BoldDesk
  6. How Resolution Due Is Calculated Using Business Hours in BoldDesk
  7. Recalculating SLA Upon Ticket Reopening
Was this article useful?
Like
Dislike
Help us improve this page
Please provide feedback or comments
Comments (0)
Access denied
Access denied
Access denied
Access denied

No articles or sections found
No articles or sections found