⚠️ Select your experience
The steps depend on your interface version. Select a link to skip to the instructions for your version.
Customers who joined Drata on or after Feb 24, 2026 are automatically on the New Experience.
Instructions for the New Experience ⬇️
AI summaries help you quickly understand why a custom test failed without digging through raw evidence. Instead of manually reviewing findings, Drata AI generates plain-language explanations for each failed resource. You can also export summaries to CSV for deeper analysis or auditor sharing.
Prerequisites
Plan availability: AI summaries are available for select plans.
Permissions: Only Administrators can enable or disable AI features.
Test type: AI summaries are supported only for custom tests.
Drata-provided tests do not support AI summaries.
If AI is not enabled by an administrator, AI summaries will not be available in your workspace.
Enable or disable AI summaries
Administrators can manage AI settings at the company level.
Go to Settings.
Under Configuration, select AI settings.
Enable AI settings and save your changes.
To disable AI summaries:
Turn off Enable AI and save.
Existing AI summaries are hidden, and no new summaries are generated.
Generate an AI summary for a test failure
AI summaries can be generated when a custom test fails.
Go to Monitoring.
Select a failing custom test.
In the test details page, locate the Latest result section.
Select Generate summary.
Drata AI analyzes the test logic and failed resources across all accounts to generate explanations. If many resources failed, summary generation may take additional time.
Once the AI summary is generated, a CSV is downloaded containing failure explanations. The CSV includes:
Provider name
Account alias or ID
Resource name
Resource ID
AI-generated summary
Use this export to:
Share findings with engineering or security teams
Track remediation work
Support audit discussions
Important considerations
AI summaries are assistive, not authoritative.
AI may not extract all details in every scenario.
Always validate AI output before using it for security or compliance decisions.
All standard test data and evidence remain available regardless of AI usage.
View test failure details
Before generating an AI summary, you can review detailed failure information for a test.
Go to Monitoring.
Select a failed custom test to open its details page.
In the History tab, review failure details such as:
Provider name
Account alias or ID
Test status (Failing or Passing)
Select the failed event for more information (such as the raw JSON).
This helps you understand the scope of failures before generating summaries.
Instructions for the Classic Experience ⬇️
Use our AI-powered feature to quickly generate summaries for test failures in custom tests. When a test fails, Drata's AI provides an explanation directly in the test drawer, eliminating the need to switch screens. You can also download a CSV with detailed failure reasons and resource IDs, streamlining issue resolution.
Prerequisites
Only Admins can enable AI summaries. If you are not an admin, you cannot enable or disable AI in Drata. If an admin did not enable AI, you will not have AI enabled in Drata.
Enable or disable AI
Admins can enable AI summarization through the Company settings page.
Go to your Settings page. Under Company Settings, select AI Settings.
Select the AI option and save your choice.
To disable AI, de-select Enable AI option and save your choice. If you disable AI, your AI summaries will be hidden and no additional AI summaries will be generated.
View test failure details
You can view test details such Provider name, Account Alias or ID, and test status (Failing or Passing) within the Last test result section. To view these details:
Navigate to the Monitoring page and select a failed Drata test.
In the test detail drawer, scroll to the Last test result section to access the following information for each account:
Provider name
Account Alias/ID
Status (such as Failing or Passing)
View event link
Select the View event link and you will be redirected to the Audit event for that particular account where you can view the test info and the raw test evidence.
Generate an AI summary of your test failure
Note: AI summaries for test failures are only available for tests built on Custom tests. At the moment, this would only apply to custom tests in your account.
Navigate to the Monitoring page and select the desired failing custom test.
Within the test detail drawer, scroll down to the Last test result section. Select the Generate Summary button below the table.
Drata AI will use the test logic configured and for each resource across all accounts that failed, it will generate an explanation. If the number of resources that failed for the test is large, generating an AI summary can take some time.
Once Drata AI completes the summary, select the download icon to download a CSV with the following information for each failed resource:
Provider Name
Account Alias/ID
Resource Name
Resource ID
AI summary
View AI summary for each account
You can also view the AI summary for each specific account.
Navigate to the Monitoring page and select the desired failing test.
Within the test detail drawer, scroll down to the Last test result section and select View event link.
You will be redirected to the Audit event for that account test failure. Select Raw test evidence section to view the AI generated Summary of test failures. It contains the following info:
Resource name (Buckets for example)
Cause: A short summary for the reason those resources are failing
Failed Resource: Shows the number of resources that failed for this reason.
Summary of the reason those resources failed.
Select the copy icon at the top right of the AI summary to copy all of the text to your clipboard. You're also able to generate a csv for this particular account with the same info as above.
Note: Given the density and volume of information, the AI may fail to extract certain information. We have thoroughly tested this to minimize the chance of this happening as much as possible, but please make sure to always review AI outputs when making security and compliance decisions.