Data Residency
Choose where your AI agent data is stored to meet compliance and governance requirements
Data residency for Airia AI allows global teams to choose the region where their AI agent data is stored at rest, helping organizations meet local compliance requirements and data governance policies.
The data residency feature affects only the data storage location for certain types of data. Airia AI will continue to process data and store certain categories of data in accordance with your service agreement and Airia AI’s Privacy Policy. Data processing may occur outside your selected region to ensure optimal AI agent performance.
If you’re not using data residency, your data will be stored in our default region (North America - US East). For more information, see Airia AI’s Data Management Policy.
Available Data Residency Regions
North America
US East Primary region for North American customers with data centers optimized for low latency across the continent.
Canada Primary region for Canadian customers with data centers optimized for low latency across the continent.
Europe
Netherlands Serves customers across Europe with GDPR-compliant infrastructure.
Middle East
UAE North Dedicated region for Middle Eastern customers with local data sovereignty requirements.
Australia
Australia East Serves Australian and New Zealand customers with local data residency compliance.
How Data Residency Works
Data Stored in Your Selected Region
The following categories of customer data will be stored at rest in your selected region:
- AI agent configurations and training data
- Conversation histories and execution logs
- Uploaded files and documents used by AI agents
- Custom knowledge bases and vector embeddings
- User-generated content and agent responses
- Workflow automation data and triggers
- Analytics data specific to your AI agents
- Application logs and debugging information
Data That May Be Stored Outside Your Selected Region
The following categories of data may be stored in regions outside of your selected data region:
- User profiles and account information
- Billing and subscription data
- Authentication tokens and security credentials
- Usage metrics and platform analytics
- System-generated identifiers and metadata
- Feature usage statistics and performance metrics
- Support ticket data and communications
Data Processing: While your data is stored in your selected region, AI processing and model inference may occur across multiple regions to ensure optimal performance and availability of our AI services.
Data Processing and External Services
Airia AI leverages various external services and third-party providers to deliver comprehensive AI capabilities. While we ensure your data storage complies with your selected region, data processing may involve external services that operate in different regions. Your organization’s administrators have full control over which AI models and external processing capabilities are enabled. When configuring your AI agents, admins can choose which specific models and processing features (such as tool calls, integrations, and AI services) are used, and that may operate outside your selected data residency region. This gives you complete transparency and control to balance functionality needs with data residency requirements.
Types of External Processing
The following types of data processing may occur outside your selected region:
- AI Model Inference: When using third-party AI models (OpenAI, Anthropic, Google, etc.), your prompts and content are processed by these providers in their available regions
- Document Processing: File parsing, OCR, and document analysis may be processed by external services
- Voice Services: Audio transcription and text-to-speech processing through external providers
- Image Processing: Computer vision and image analysis services
- Search and Retrieval: External search engines and knowledge bases for information retrieval
- Integration Services: Data synchronization with external platforms and APIs
- Analytics and Monitoring: Performance monitoring and usage analytics through third-party services
Data Processing Safeguards
To protect your data during external processing, we implement the following measures:
- Data Minimization: Only necessary data is sent to external services for processing
- Encryption in Transit: All data is encrypted during transmission to external services
- Contractual Protections: Data processing agreements with all external providers
- Audit Trails: Complete logging of data processing activities
- Retention Policies: Ensuring external providers don’t retain your data longer than necessary
- Access Controls: Strict controls on which data can be processed externally
Common External Service Providers
The following external services may process your data as part of Airia AI’s operations:
| Service Category | Provider Examples | Data Processed | Processing Regions |
|---|---|---|---|
| AI Language Models | OpenAI, Anthropic, Google | Prompts, conversations, generated content | Primarily US, varies by provider |
| Document Processing | Azure Document Intelligence, AWS Textract | Document content, images, metadata | Multiple regions available |
| Voice Services | OpenAI Whisper, Azure Speech | Audio files, transcripts, voice synthesis | US, EU (varies by service) |
| Search & Retrieval | Various search APIs | Search queries, results metadata | Global |
| Cloud Infrastructure | AWS, Azure, GCP | System logs, performance metrics | Aligned with your region where possible |
| Analytics & Monitoring | Monitoring services | Usage statistics, performance data | Various regions |
The specific external services used may vary based on your configuration, enabled features, and the AI models you choose to use.
Controlling External Processing
You have several options to control how your data is processed externally:
- Model Selection: Choose AI models that align with your data residency requirements
- Feature Configuration: Disable features that require external processing if not needed
- BYOK (Bring Your Own Key): Use your own API keys for supported services to maintain direct control
- On-Premises Options: For highly sensitive data, consider on-premises deployment options
- Data Classification: Mark sensitive data to restrict external processing
When you use AI models or services that don’t support your selected data residency region, processing will occur in the provider’s available regions. We recommend reviewing the data residency policies of each AI provider you plan to use.
Service-Specific Data Residency Details
Core Services
Different AI services have varying levels of data residency support based on the underlying models and infrastructure:
| Capability | Model | NA | EMEA | MENA | APAC | ||||
|---|---|---|---|---|---|---|---|---|---|
| storage | inference | storage | inference | storage | inference | storage | inference | ||
| Document Parsing | External Provider 1 | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| Image Scanning | External Provider 1 | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| Image Scanning | External Provider 2 | ✅ | ✅ | not available | not available | ✅ | ✅ | ||
| Image Scanning | External Provider 3 | ✅ | ✅ | not available | not available | ||||
| Embedding | External Provider 4 | ✅ | ✅ | ||||||
| Embedding | External Provider 5¹ | ✅ | ✅ | ||||||
| Vector Storage | External Provider 6 | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| Voice | External Provider 7 | ✅ | ✅ | ||||||
| Voice | External Provider 8 | ✅ | ✅ |
AI Model Providers
The following table shows the availability and data residency characteristics of different AI model providers:
| Provider | Notable Models | Airia Key Hosted Region | BYOK Available Regions | Data Residency Notes |
|---|---|---|---|---|
| AI21 | Jamba | US | US | Limited to US region |
| Alibaba | Qwen | China | China | China region only |
| Anthropic | Claude | US | US | US region only |
| AWS Bedrock | Titan | US | Most regions globally | Regional availability varies by model |
| Azure Foundry | Various | Multiple | Most regions globally | Depends on deployment type² |
| Cohere | Command | US | US | US region only |
| Gemini | US | US | US region only | |
| Mistral | Mistral | EU | EU | EU region only |
| OpenAI | GPT, o-series | US | US | US region only |
| Perplexity | Sonar | US | US | US region only |
| Replicate | Stable Diffusion, Flux, Llama | US | US | US region only |
| TogetherAI | Llama, Qwen, Deepseek | US | US | US region only |
| XAI | Grok | US | US | US region only |
| Custom (OpenAI Compatible) | Various | Varies | Varies | Depends on your configuration |
When using AI models that don’t support your selected data residency region, inference will occur in the provider’s available regions. Your conversation data and configurations will still be stored in your selected region, but the AI processing will happen where the model is hosted.
Security and Compliance
All data residency regions maintain the same high security standards:
- End-to-end encryption for data in transit and at rest
- Regular security audits and compliance certifications
- Local compliance with regional data protection regulations
- 24/7 monitoring and incident response capabilities
Frequently Asked Questions
Will data residency affect my AI agent performance?
Will data residency affect my AI agent performance?
While data is stored in your selected region, AI processing is optimized globally to ensure the best performance. You may experience slightly different response times depending on your region.
Are there any features that aren't available in certain regions?
Are there any features that aren't available in certain regions?
All core Airia AI features are available across all data residency regions. Some advanced AI models may have regional availability differences as shown in the tables above.
How does data residency work with third-party integrations?
How does data residency work with third-party integrations?
Data from third-party integrations follows the same residency rules as your other data. However, the third-party services themselves may have their own data residency policies.
What happens if I use a model that's not available in my region?
What happens if I use a model that's not available in my region?
If you select an AI model that doesn’t support your data residency region, the inference will occur in the provider’s available regions. Your conversation data and configurations will still be stored in your selected region.