This section covers the most common issues you might face with Revelation AI and the steps to fix them quickly, so you can get back to work without delay.
For any non-technical questions, refer to the FAQ page.
⚠️ Sign-in fails or times out
Cause: Token expired or incorrect user role
Fix:
- Re-authenticate using valid Azure credentials making sure you sign in to the correct Azure tenant
- Ensure the user has the Cognitive Services Contributor role on the correct Subscription.
- Complete setup within 60 minutes to avoid token expiration
⚠️ No subscriptions or resource groups appear
Cause: Insufficient permissions or region mismatch
Fix:
- Confirm the user is assigned to the correct subscription and you have selected the correct region.
- Check Azure IAM for appropriate role (Contributor or Cognitive Services Contributor)
- Ensure the resource group exists in the selected region / subscription.
⚠️ “Create AI Resource” button is disabled
Cause: Missing or invalid input
Fix:
- Fill in all required fields: Region, Subscription, Resource Group, OpenAI Resource Name, Deployment Name
- Avoid special characters or uppercase letters in names
⚠️ Model deployment fails
Cause: Region unsupported or quota exceeded / model deprecation or resource name conflict.
Fix:
- Read the error message carefully for the cause.
- Verify the selected region supports your chosen Azure OpenAI model and capacity selection.
- Choose a recommended model or a model that has not been deprecated.
- Check Cognitive Services quotas in the Azure Foundry portal
- Try deploying a smaller model or reduce RPM/TPM settings
- Try deploying to a different region like EAST US 2 or a region that is geographically close to your location.
- Try sign out and back in and re-deploy using the same settings.
⚠️ HTTP 401 Error
Cause: Invalid or outdated API key
Fix:
- If AI Assist chat returns an HTTP 401, the API key has likely been changed or revoked in Azure AI
- Retrieve a fresh key from the Azure AI portal and update it in the Revelation admin area
- Use the “API Key” button in the Revelation AI admin deployments page to enter the new API key.
⚠️ HTTP 404 AI model deployment not found
Cause: Deleted or incorrect deployment reference
Fix:
- An HTTP 404 in AI Assist chat means the model deployment cannot be found—likely deleted in Azure AI
- In Revelation AI admin, deploy a new model and delete the outdated reference
- Ensure the correct deployment is marked as the system default.
⚠️ HTTP 429 or very slow response times
Cause: RPM (Requests per Minute) or TPM (Tokens per Minute) limit reached
Fix:
- Indicates throttling due to system model deployment limits
- Increase RPM/TPM limits in the Revelation AI admin deployment settings
- Adjust based on expected user concurrency or peak usage patterns
⚠️ HTTP 400 (content_filter)
Cause: Prompt triggered Azure OpenAI’s content management policy
Fix:
- The AI was unable to respond due to content filtering
- Review and revise the prompt to avoid triggering sensitive categories (e.g., hate, sexual, violence, self-harm)
- See Azure OpenAI Content Filtering Overview for details
⚠️ AI Assist not appearing for users
Cause: Feature not enabled or access not granted
Fix:
- Go to Admin > Revelation AI > Settings
- Ensure AI Assist is toggled ON
- Assign access to relevant teams or access levels
- Users should log out and back in to get the latest assigned permissions.
⚠️ Unexpected token usage or cost spikes
Cause: High RPM/TPM settings or frequent model calls
Fix:
- Review usage in Azure Cost Management
- Lower RPM/TPM via the capacity slider
- Set up budget alerts and webhook notifications
- Consider re-generating your API key in case it has been compromised and update it in the Deployments tab.