Skip to main content

​
Multi-Agent Teams

A single AI agent is powerful. A team of agents working together is unstoppable. Teams in thinnestAI let you combine multiple specialized agents into a coordinated group. Instead of building one monolithic agent that tries to do everything, you create focused agents and let them collaborate.

​
Why Use Teams?

Think of it like a real company. You wouldn’t hire one person to handle sales, support, billing, and engineering. You’d build a team where each person excels at their role. Teams are useful when:
  • Tasks are complex — A customer inquiry might need a support agent, a billing agent, and a technical agent working together.
  • Domains are distinct — Your sales agent shouldn’t need to know your entire knowledge base. Let a research agent handle that.
  • You need routing — Different types of requests should go to different agents automatically.
  • Quality matters — Specialized agents outperform generalist agents on domain-specific tasks.

​
How Teams Work

A team consists of:
  1. Member agents — The individual agents that make up the team. Each has its own model, instructions, tools, and knowledge.
  2. A team mode — How agents coordinate with each other (route, coordinate, collaborate, or tasks).
  3. A team leader (in some modes) — An orchestrator agent that delegates work and synthesizes results.
When a user sends a message to a team, the team mode determines how the message flows between agents.

​
Creating a Team

​
From the Dashboard

  1. Navigate to Agents in the sidebar.
  2. Click Create Agent and select Team as the type.
  3. Add member agents — you can create new ones or select existing agents.
  4. Choose a Team Mode (see Team Modes for details).
  5. Configure the team leader’s instructions (if applicable).
  6. Save and test.

​
From the Flow Editor

The visual flow editor gives you a drag-and-drop interface for building teams:
  1. Open the Flow Editor from your agent’s page.
  2. Drag agent nodes onto the canvas.
  3. Connect them to define the team structure.
  4. Configure the team mode in the Team Config panel.
  5. Deploy when ready.

​
Via the API

curl -X POST https://api.thinnest.ai/agents \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Customer Service Team",
    "type": "team",
    "team_mode": "route",
    "team_members": [
      {"agent_id": "agent_support_123"},
      {"agent_id": "agent_billing_456"},
      {"agent_id": "agent_technical_789"}
    ],
    "instructions": "Route customer inquiries to the most appropriate agent."
  }'

​
Team Configuration

SettingDescription
Team ModeHow agents coordinate — coordinate, route, collaborate, tasks, or broadcast
Member AgentsThe agents in the team
Leader InstructionsInstructions for the orchestrator/router agent
Max IterationsMaximum rounds of agent collaboration (prevents infinite loops)
Response FormatHow the final response is synthesized

​
Example: E-Commerce Support Team

Here’s a practical example — a team that handles e-commerce customer inquiries:
AgentRoleTools
RouterClassifies incoming requests—
Order AgentChecks order status, handles returnsOrder API, Shipping API
Product AgentAnswers product questionsProduct Knowledge Base
Billing AgentHandles payment issues, refundsPayment API
Set the team mode to Route, and incoming messages automatically go to the right agent.

​
Best Practices

  • Keep agents focused — Each agent should have a clear, narrow responsibility.
  • Use descriptive names — The team leader uses agent names and descriptions to make routing decisions.
  • Start simple — Begin with 2-3 agents and add more as needed.
  • Test each agent individually before adding it to a team.
  • Monitor team performance — Check which agents handle the most requests and how often handoffs occur.

​
Next Steps