Getting Started
Step-by-step guide to creating and deploying your AI agent system with Control Hub.
This guide walks you through the complete workflow from entering Control Hub to having a deployed, production-ready AI agent system.
Step 1: Agentic System Generation
When you enter Control Hub from a use case, the AI architect automatically generates a multi-agent system tailored to your specific use case. The system creates:
- Specialized agents tailored according to your use case
- Platform agents to manage overall infrastructure, security and observability
- Appropriate model selections for each agent
- Tool configurations based on required capabilities
- Memory settings for conversation management
- System prompts that define agent behavior
Generation typically produces 3-7 agents depending on use case complexity.
Step 2: Review Your Agents
Examine each generated agent to understand its role in the system:
- What problem does this agent solve?
- What tools does it need access to?
- How does it communicate with other agents?
- What model is appropriate for its complexity?
Step 3: Configure Tools
Connect your agents to the systems they need:
- Knowledge Base (RAG) -- Upload documents or connect data sources like Confluence, SharePoint, or S3
- Business Integrations -- Connect to Slack, Gmail, Jira, Salesforce, and other tools
- Web Search -- Enable real-time internet search capabilities
Step 4: Refine with Natural Language
Use the chat interface to modify your architecture:
- "Add a Slack integration to the support agent"
- "Change the coordinator to use long-term memory"
- "Add a sentiment analysis agent to the team"
The AI architect updates your configuration in real-time.
Step 5: Test Your Agents
Switch to the Test Bed to validate agent performance:
- Review the extracted features from your architecture
- Run tests to evaluate agent behavior
- Inspect execution traces to understand agent reasoning
- Identify and fix any failing tests
Iterate between configuration and testing until all features pass.
Step 6: Access Your API
After deployment:
- Copy your API endpoint URL
- Generate an API key (store it securely -- shown only once)
- Start making requests to your deployed agents
Your multi-agent system is now live and ready for integration.
Step 7: Deploy to Production
When satisfied with your agents:
- Click "Approve" to trigger deployment
- Monitor the deployment progress
- Wait for code generation and infrastructure provisioning (3-6 minutes)