Support
Common Issues
Solutions to frequently encountered problems with Prisme.ai
This page addresses the most commonly encountered issues when using Prisme.ai. For each issue, we provide clear explanations and step-by-step solutions. If you don’t find your specific problem here, please refer to our Troubleshooting Guide or contact our support team.
Authentication and Access Issues
If you’re having trouble logging in, try these solutions in order:
- Password Reset:
- Click “Forgot Password” on the login page
- Follow the instructions sent to your email
- Set a new password and try logging in again
- Check Your Email Address:
- Ensure you’re using the correct email address for your account
- Check for typos or case sensitivity issues
- Browser Issues:
- Clear your browser cache and cookies
- Try using an incognito/private browsing window
- Try a different browser altogether
- Account Status:
- Your account may be locked or inactive
- Contact your organization’s administrator
- SSO Configuration:
- If your organization uses SSO, make sure you’re using the SSO login option
- Contact your IT department to verify SSO settings
If none of these solutions work, please contact support with the specific error message you’re seeing.
Access issues for specific workspaces or projects typically relate to permissions:
- Check Your Role:
- Different roles have different access levels
- Ask your workspace administrator to verify your assigned role
- Workspace Membership:
- You may not be added to the workspace yet
- Request access from the workspace owner
- Project Visibility:
- Projects can be set to private or limited visibility
- Ask the project owner to share it with you
- Organization Boundaries:
- Some workspaces may be restricted to specific organizations
- Verify you’re using the correct organization context
- License Limitations:
- Your license may not include access to certain features
- Contact your organization’s Prisme.ai administrator
For SSO integration issues:
- Configuration Check:
- Verify the SSO settings in both Prisme.ai and your identity provider
- Check that entity IDs, URLs, and certificates match exactly
- Certificate Expiration:
- SAML certificates may have expired
- Generate and upload new certificates
- Attribute Mapping:
- User attributes may not be mapping correctly
- Review attribute mapping configuration
- Browser Issues:
- Ensure cookies and third-party cookies are enabled
- Some browsers block redirects in security modes
- Network Restrictions:
- Check if firewalls or proxies are blocking authentication redirects
- Whitelist necessary domains and endpoints
For detailed SSO troubleshooting, refer to our SSO Implementation Guide.
AI Knowledge and RAG Issues
If your documents aren’t processing as expected in AI Knowledge:
- Check Document Format:
- Ensure your documents are in a supported format (PDF, DOCX, TXT, etc.)
- Convert documents to a different format if needed
- File Size and Complexity:
- Very large files (>50MB) may take longer to process
- Complex documents with lots of formatting, images, or tables may have extraction issues
- Text Extraction:
- Some PDFs contain images of text rather than actual text
- OCR processing may be needed for scanned documents
- Processing Status:
- Check the document processing status in the AI Knowledge interface
- Documents in “Processing” state may need more time
- Character Encoding:
- Documents with unusual character encoding may not process correctly
- Try saving in UTF-8 format
If issues persist, try uploading a simpler version of the document or contact support for assistance.
If your AI Knowledge agent isn’t using information from your documents:
- Check Document Status:
- Verify that documents are fully processed and in “Active” state
- Documents in “Pending” or “Error” state won’t be used
- Query Formulation:
- Be specific in your questions
- Use keywords that appear in your documents
- Try rephrasing your question in different ways
- Chunk Size and Overlap:
- Adjust chunk size settings for your documents
- Smaller chunks may improve retrieval of specific information
- Larger chunks provide more context
- Embedding Settings:
- Different embedding models have varying performance characteristics
- Try changing the embedding model in your project settings
- Retrieval Configuration:
- Adjust the number of chunks retrieved
- Enable query enhancement or self-query features
- Prompt Engineering:
- Review and refine your system prompt
- Include specific instructions about using document content
To address inaccuracies or hallucinations in AI responses:
- Check Retrieved Context:
- Review the chunks retrieved for each query
- Verify if the correct information is being provided to the model
- Adjust Retrieval Settings:
- Increase the number of retrieved chunks
- Improve retrieval relevance with query enhancement
- Prompt Tuning:
- Add explicit instructions to prioritize retrieved information
- Include directives to avoid making up information
- Use phrases like “If you don’t know, say so”
- Model Selection:
- Some models are more prone to hallucination than others
- Try switching to a different LLM
- Document Quality:
- Ensure your documents contain clear, accurate information
- Break down complex documents into more manageable pieces
- Implement Testing:
- Set up automated tests with expected answers
- Regularly evaluate accuracy against known information
AI Builder and Custom App Issues
If your automation isn’t triggering as expected:
- Event Configuration:
- Verify that the event name matches exactly between the trigger and emitter
- Check for typos or case sensitivity issues
- Automation Status:
- Ensure the automation is published and active
- Unpublished changes won’t affect the live automation
- Scope Issues:
- Events may be limited to specific scopes (current socket, workspace, etc.)
- Check the event emission target configuration
- Error Handling:
- Check activity logs for errors
- Add error tracking to your automation
- Permissions:
- Verify that the user has permissions to trigger the automation
- Some automations require specific roles
- Rate Limiting:
- Frequent triggers may be subject to rate limiting
- Implement backoff strategies for high-frequency events
If your custom code functions are not working:
- Syntax Errors:
- Check for JavaScript syntax errors in your code
- Use console.log statements to debug
- Parameter Handling:
- Ensure all required parameters are being passed
- Check parameter types and formats
- Environment Access:
- Custom code has limited access to environment variables
- Verify you’re not trying to access restricted resources
- Execution Timeout:
- Functions have maximum execution times
- Optimize long-running operations
- External Dependencies:
- Custom code can’t import external packages
- Use built-in JavaScript functionality
- Error Handling:
- Implement try/catch blocks in your code
- Return meaningful error messages
Check the activity logs for specific error messages from your custom code execution.
For API integration issues:
- Authentication:
- Check API keys and credentials
- Verify token expiration and refresh mechanisms
- Endpoint Configuration:
- Confirm the API endpoint URL is correct
- Check for typos or formatting issues
- Request Format:
- Verify the request body matches the API requirements
- Check content type headers
- Parameter Validation:
- Ensure all required parameters are provided
- Check parameter formats and data types
- Rate Limiting:
- Many APIs have rate limits
- Implement retry logic with backoff
- Response Handling:
- Add proper error handling for API responses
- Parse response data correctly
Use the fetch preview feature to test API calls directly in the builder interface.
Deployment and Performance Issues
For issues with self-hosted deployments:
- Infrastructure Requirements:
- Verify your environment meets the minimum requirements
- Check CPU, memory, and storage allocations
- Kubernetes Configuration:
- Ensure Kubernetes version is compatible
- Check pod, service, and ingress configurations
- Network Issues:
- Verify network policies allow required communications
- Check DNS resolution and service discovery
- Storage Configuration:
- Ensure persistent volumes are properly configured
- Check storage class and provisioner settings
- Container Registry Access:
- Verify access to container images
- Check image pull secrets if using private registries
- Helm Chart Configuration:
- Review values.yaml for misconfigurations
- Check version compatibility between components
Refer to our detailed Self-Hosting Troubleshooting Guide for specific deployment scenarios.
To address performance issues:
- Browser Performance:
- Clear browser cache and close unnecessary tabs
- Try a different browser or device
- Network Connectivity:
- Run a speed test to check your internet connection
- Check for VPN or proxy issues
- Resource Utilization:
- For self-hosted deployments, check CPU, memory, and disk usage
- Scale resources if necessary
- Concurrent Users:
- High user load can affect performance
- Check usage patterns and consider scaling
- Large Data Volumes:
- Very large knowledge bases may affect performance
- Consider optimizing document chunking
- Model Selection:
- Some AI models are more resource-intensive than others
- Try switching to a lighter model for faster responses
For persistent performance issues, please contact support with specific details about when and where slowdowns occur.
When encountering conflicts while pushing changes:
- Pull Latest Changes:
- Always pull the latest version before pushing
- This ensures you’re working with the most recent state
- Identify Conflicting Items:
- Check the error message for specific items causing conflicts
- Look for items that may have been modified by others
- Resolve Conflicts:
- For each conflicting item, decide which version to keep
- You may need to manually merge changes
- Coordinate with Team Members:
- Communicate with others who might be working on the same workspace
- Establish coordination practices for shared workspaces
- Use Versioning:
- Create a new version before making major changes
- This preserves the previous state as a fallback
- Structured Workspaces:
- Consider splitting large workspaces into smaller ones
- This reduces the likelihood of conflicts
Billing and Account Issues
To change your subscription plan:
- Contact Sales:
- Reach out to our sales team at hello@prisme.ai
- Specify your current plan and desired changes
- Self-Service Option:
- Organization administrators can manage subscription changes
- Navigate to Organization Settings > Billing > Subscription
- Select “Change Plan” and follow the prompts
- Mid-cycle Changes:
- Changes during a billing cycle are prorated
- You’ll receive a credit or additional charge as appropriate
- Contract Adjustments:
- For enterprise contracts, contact your account manager
- Contract amendments may be required
Please note that downgrading may affect feature availability and user limits.
For billing issues:
- Review Invoice Details:
- Check the invoice in your billing dashboard
- Verify the billing period, plan, and any additional charges
- Usage-based Charges:
- Some plans include usage-based billing
- Review your usage metrics in the dashboard
- Payment Method:
- Verify your payment method is current
- Check for expired cards or insufficient funds
- Tax Information:
- Ensure your tax information is complete and correct
- Tax rates vary by location
- Contact Billing Support:
- Email billing@prisme.ai with your account details
- Include the invoice number and specific concerns
For immediate assistance with critical billing issues, please contact your account manager.
To manage users in your organization:
- Administrator Access:
- Only organization administrators can manage users
- Verify you have the necessary permissions
- Adding Users:
- Navigate to Organization Settings > Users > Add User
- Enter the user’s email address and select their role
- An invitation will be sent automatically
- Removing Users:
- Navigate to Organization Settings > Users
- Find the user you want to remove
- Click “Remove” and confirm
- Bulk User Management:
- For adding multiple users, use the bulk import option
- For enterprise customers, SCIM provisioning is available
- License Considerations:
- Adding users may affect your licensing costs
- Some plans have user limits
- Role Assignment:
- Be sure to assign appropriate roles to new users
- Roles determine what actions users can perform
Changes to user accounts take effect immediately.
If You Still Need Help
If you’ve tried the solutions above and are still experiencing issues:
Troubleshooting Guide
Detailed steps for diagnosing and resolving specific issues
Contact Support
Get personalized assistance from our support team
Training Programs
Enhance your skills with our training offerings
FAQ
Find answers to frequently asked questions
When contacting support, please include the following information to help us resolve your issue quickly:
- Detailed description of the problem
- Steps to reproduce the issue
- Error messages (if any)
- Screenshots (when applicable)
- Environment details (browser, operating system)
- Account information (organization name, workspace)
Was this page helpful?