Enterprise

Custom Agents

Create AI agents tailored to your business context and industry

Custom agents let you create AI team members with specialized knowledge for your industry, company processes, and data sources. Available on Enterprise plans.

When to Use Custom Agents

Your industry has domain-specific terminology or metrics not covered by default agents
You need an agent focused on a specific business function (e.g., Supply Chain, Legal, Product)
You want agents that understand your company's internal processes and terminology
You need agents trained on proprietary data sources or internal APIs

Configuration Schema

Define your custom agent with a JSON configuration object:

agent-config.jsonjson

{
  "name": "Supply Chain Analyst",
  "role": "supply_chain",
  "personality": "Detail-oriented logistics expert who optimizes for cost efficiency while maintaining service levels.",
  "expertise": [
    "Supply chain optimization",
    "Inventory management",
    "Vendor relationship management",
    "Logistics cost analysis"
  ],
  "dataSources": ["erp", "warehouse_api", "shipping_tracker"],
  "promptTemplate": "You are {{name}}, a {{role}} at {{company_name}}...",
  "sandbox": false
}

Prompt Templates

Customize how your agent processes queries and generates responses. Templates support variable interpolation for dynamic context injection.

docs.customAgents.promptTemplates.variables

Testing & Validation

Before deploying a custom agent, test it in sandbox mode. Sandbox agents have access to your data but their responses are not visible to other team members.

1Create agent with sandbox: true in configuration
2Send test queries covering your key use cases
3Review responses for accuracy and tone
4Adjust personality and expertise parameters
5Set sandbox: false to deploy to your workspace

Deployment & Versioning

Custom agents support versioning. When you update an agent's configuration, the previous version is preserved. You can roll back at any time from your dashboard.

Integrations