What Are AI Providers?
AI Providers are secure connections to external AI services that power Elementum’s intelligent features. Think of them as bridges that allow Elementum to access advanced AI models for tasks like text generation, data analysis, and intelligent automation.AI Providers are the foundation for all AI features in Elementum. You’ll need to configure at least one provider before you can use AI Services, AI Search, or Agents.
How AI Providers Work
AI Providers establish authenticated connections to external AI services, allowing Elementum to securely access their models and capabilities. Once configured, these providers enable you to:- Create AI Services for specific tasks (LLMs, embeddings)
- Power intelligent automations with AI-driven actions
- Enable AI Search across your data
- Deploy conversational Agents for customer support
Supported AI Providers
Elementum supports three major AI providers, each with unique strengths:OpenAI
Best for: General-purpose AI tasks, conversational agents, and text generationModels: o3, o4-mini, GPT-4 Turbo (LLM services only)Setup: API key authentication
Snowflake Cortex
Best for: Data-native AI with direct access to your warehouseModels: Claude Sonnet 4, Claude 3.7 Sonnet, Claude Opus 4, Arctic L V2.0Setup: Uses your existing CloudLink connection
Google Gemini
Best for: Multimodal AI capabilities and Google ecosystem integrationModels: Gemini 2.5 Pro, Gemini 2.5 (LLM services only)Setup: Google Cloud service account
Provider Selection Guide
Choose OpenAI if:- You want the most mature and widely-used AI models
- You need reliable conversational AI capabilities
- You’re building customer-facing agents
- You prefer straightforward API key setup
- You already have a CloudLink connection configured
- You want AI that operates directly on your data warehouse
- You need data privacy and residency controls
- You’re building data-intensive AI workflows
- You need multimodal AI capabilities (text, images, etc.)
- You’re already using Google Cloud services
- You want cutting-edge AI model performance
- You need flexible model options
Authentication & Security
All AI Providers use secure authentication methods:- API Key Authentication: OpenAI uses encrypted API keys
- Service Account Authentication: Google Gemini uses service account credentials
- CloudLink Integration: Snowflake Cortex leverages your existing secure connection
Security Best Practice: Never share your API keys or service account credentials. Store them securely and rotate them regularly according to your security policies.
Setup Process Overview
Setting up AI Providers follows a consistent pattern:1
Prerequisites
Ensure you have the necessary accounts and permissions for your chosen provider
2
Obtain Credentials
Get API keys, service account files, or configure CloudLink access
3
Configure Provider
Add the provider in Elementum’s Organization Settings
4
Test Connection
Verify the connection works correctly
5
Create AI Services
Set up specific AI Services using your configured provider
Getting Started
Ready to set up your first AI Provider? Choose your preferred provider and follow the detailed setup guide:Set Up OpenAI
Configure OpenAI API access for general-purpose AI capabilities
Set Up Snowflake Cortex
Enable AI directly on your data warehouse
Set Up Google Gemini
Connect to Google’s advanced AI models
Multiple Providers
You can configure multiple AI Providers simultaneously to:- Use different models for different tasks
- Implement failover strategies for reliability
- Optimize costs by routing to the most efficient provider
- Access specialized capabilities from each provider
Pro Tip: Start with one provider to get familiar with AI features, then add additional providers as your needs grow.
Next Steps
After configuring your AI Providers:- Create AI Services - Set up specific LLMs and embedding models
- Configure AI Search - Enable intelligent search across your data
- Build Agents - Create conversational AI assistants
- Use AI Actions - Add AI to your automation workflows
Common Issues
Connection Failures
Connection Failures
Symptoms: Unable to connect to AI providerCommon Causes:
- Invalid API keys or credentials
- Network policy restrictions
- Insufficient permissions
- Verify credentials are correct and active
- Check network policies and firewall rules
- Ensure proper permissions are granted
Model Availability
Model Availability
Symptoms: Desired models don’t appear in service creationCommon Causes:
- Provider not properly configured
- Account limits or restrictions
- Regional availability issues
- Verify provider configuration is complete
- Check account status and limits
- Confirm model availability in your region
AI Providers are the foundation of Elementum’s AI capabilities. Choose the right provider for your needs and start building intelligent workflows today.