---
sourceDocument: Brazil Build workflows
sourceDocumentLink: https://www.servicenow.com/docs/r/build-workflows

 Release :

    - brazil

ft:locale :

    - en-US

ft:publication_title :

    - Brazil Build workflows

ft:clusterId :

    - crworkflow

bundleId :

    - crworkflow

workflow :

    - Creator


---

# Create an intelligent approval

# Create an intelligent approval {#ariaid-title1}

Release version: Brazil  
Updated October 19, 2026  
![](https://www.servicenow.com/docs/portal-asset/ico-clock) 4 minutes to read  
Generate an intelligent approval from an existing policy document.

## Before you begin

Role required: sn_iap.policy_manager or sn_iap.policy_admin

## Procedure

1. Navigate to AllIntelligent ApprovalsIntelligent Approvals Home.  
2. Select Create new.
3. Select the existing policy document that you want to upload.  

   The policy document must be in a PDF format, and the file size must be 10 MB or less.
4. Verify the name of your source policy document, and select Create intelligent approval.  
   The system analyzes your policy document to determine what records the policy applies to, what conditions start the approval process, and the approvals conditions to apply.

5. Review the intelligent approvals scope and trigger conditions.  
   The system displays a Suggested options tab with a text description of the record types that the intelligent approval applies to and the approval start conditions.  
   Important:  
   Review the AI-generated trigger conditions carefully, as outputs may be inaccurate or incomplete.
   Select Edit to override the table name and field conditions of the Suggested options scope and start conditions. Select the table, fields, and field values that you want to use to start this intelligent approval. Alternatively, you can select Define my own to enter a text description of the table and field conditions that you want to use, and AI will convert your text into a new set of suggested approval conditions.

6. Select either the Suggested options or Define my own descriptive text, and then select Next.  
   The system reviews your policy document and generates the intelligent approval. The generation process may take a moment to complete.

   If the system fails to generate intelligent approvals, verify that you have access permissions to both intelligent approvals and the system records that you want approved. For requirements, see [Configure intelligent approvals](https://www.servicenow.com/docs/gwXYb76ghompAO1s5kxJWg "Configure your instance to support intelligent approvals.").
7. Review the Overview and Test results tabs of your intelligent approval policy.  
   A high percentage of AI approved/rejected results indicates that the policy document provides clear, consistent criteria for the system to evaluate. A high percentage of Can't decide results indicates that the policy document needs improvement.  
   Important:  
   AI-generated approvals and rejections may be inaccurate or inappropriate. Review the approval results before publishing the policy.
8. Review the Interpreted policy areas, Source document, Test results, and Suggested improvements tabs.  
   {#create-an-intelligent-approval__table_xsf_lr4_ykc__entry__2}

   | Tab | Description |
   |-|-|
   | Overview | This tab describes how the intelligent approval was generated and its expected automation rate. |
   | Interpreted policy areas | This tab describes the table and fields used to trigger the intelligent approval. It also lists when human review is needed. |
   | Source document | This tab shows a read-only copy of the PDF source document. |
   | Test results | This tab shows the test cases that were generated and run, the decision that the AI made, and the reasoning the AI used to make the decision. |
   | Suggested improvements | This tab shows a list of items that you can add, edit, or remove from your policy source document to potentially improve the automation rate. |
   [ ]

   {#create-an-intelligent-approval__table_xsf_lr4_ykc}
9. From the PDF document viewer toolbar, select Download to save a PDF copy of the suggestions.  
10. Use the suggested improvements to update the approval source policy document.  
    Make edits to the source policy document outside your instance using an appropriate document editor application.  
    Important:  
    AI-generated suggestions may be inaccurate or inappropriate. Review the suggestions before changing the policy.
11. Upload the updated policy document, and review the percentage of AI approved/rejected results.  
    For instructions on uploading a new source document, see [Update source document](https://www.servicenow.com/docs/tB1k5q7zicHTihhvk1BPUg "Update an existing intelligent approval by uploading a new source document. Regenerate the intelligent approval using the uploaded source document.").
12. Select the three-dot menu, and select Potential overlapping approvals.  

    The system reviews your existing intelligent approvals to identify any potential approval conflicts.
    You can resolve overlapping intelligent approvals by one of these methods.
    * You can leave your new intelligent approval in the draft state and update the existing intelligent approval to include the scope and coverage of your new intelligent approval. Choose this method when you want your existing intelligent approval to replace your new intelligent approval.
    * You can deactivate the existing intelligent approval. Choose this method when you want your new intelligent approval to replace your existing intelligent approval.
    {#create-an-intelligent-approval__ul_vqg_dp2_mjc}
13. When the percentage of AI approved/rejected results meets your requirements, select Publish.  

## Result

The intelligent approval policy is published and activated for all incoming requests that match the configured trigger conditions. The system evaluates matching requests as they are created and automatically approves or
rejects requests that clearly meet the policy criteria. Requests that the system can't evaluate remain open for human reviewers to approve.
If intelligent approvals don't start when expected, verify the intelligent approvals start conditions. Use the Edit option to update the table and field conditions.  
If too many intelligent approvals result in a Can't decide decision, review and update your policy document. Load your updated policy and use the suggested improvements to raise the percentage of AI approved/rejected results.  
Note:  
If your policy requires manual review and approval from humans, the percentage of Can't decide approvals will remain high. Intelligent approvals can't automate approval processes that require human review and approval.

*[\>]: and then