Chatzuri
Pricing
Guides
Guides
Introduction
  1. 1Getting Started
  2. 2Your Agent
    • Create an agent
    • AI configuration
    • Personality and presentation
    • Customise the chat widget
    • Test in the Playground
    • Voice, video, images, and files
    • How memory works
    • Multi-agent orchestration
    • Performance and cost controls
    • Security and access control
    • Compliance and guardrails
    • Improve response quality
  3. 3Knowledge & Sources
  4. 4Agent actions & Tools
  5. 5Agent Tasks
  6. 6WorkflowsBeta
  7. 7Channels
  8. 8Customers & Conversations
  9. 9Run Your Team
  10. 10Developer Tools
Developer reference →Showcase →
Guides2. Your AgentMulti-agent orchestration
Chapter 2 · Your Agent

Multi-agent orchestration

Pair a supervisor agent with specialist children — route conversations or consult them inline.

5 min read

Orchestration lets one agent (a supervisor) bring in other agents (specialists) when a question falls outside its lane. Instead of asking every agent to do everything, you build a small team — one front desk, a few specialists — and let them cooperate.

Find these settings under Settings → Orchestration on any agent.

Set up a supervisor

To make an agent a specialist of another agent:

  1. Open the specialist's settings.
  2. Set Supervisor agent to the agent that should route to it.
  3. Optionally add Specialization domains (e.g. "billing", "vision", "scheduling") so the supervisor knows what this specialist is best at.

From the supervisor's side, pick an Orchestration pattern. Two patterns are available today:

Handoff

With Handoff, the supervisor decides when to transfer the conversation to a specialist. After the handoff, every following message in that conversation is answered by the specialist — its prompt, model, tools, and memory — until you start a new conversation.

Use handoff when:

  • A question needs deep specialist knowledge for several turns (not just one lookup).
  • You want the customer to clearly feel they've been "transferred to billing" — like a real call centre.

The supervisor announces the transfer to the customer before it happens, so it never feels abrupt.

Sequential

With Sequential, the supervisor stays in charge. When it needs a specialist's input, it consults them privately, receives their reply, and folds it into its own answer. The customer only ever talks to the supervisor.

Use sequential when:

  • You need a one-off lookup ("what's our refund policy?") rather than a full transfer.
  • The supervisor has the customer context and the specialist provides a single fact.

Specialist calls have a depth limit of three, so a misconfigured chain can't recurse forever.

Tip
Children must be in the same team as the supervisor. A supervisor with no children doesn't get the routing tools added — the option is hidden from the model until at least one specialist is attached.

Specialisation domains

Each specialist can declare what it's good at — "vision", "reasoning", "extraction", "translation", and so on. The supervisor sees this list in its prompt and uses it to decide who to call.

Pick domains that describe the specialist's skill, not the product area. "Billing" is fine if the agent's whole purpose is billing; "extraction" is better when the specialist's job is pulling structured data out of messages.

What about concurrent or group-chat patterns?

Two other patterns — concurrent (run specialists in parallel) and group chat (specialists deliberate together) — are not yet available and are hidden from the picker. They need a different execution model and will be added in a future release.

Previous · Your AgentHow memory worksNext · Your AgentPerformance and cost controls
Chatzuri

AI-powered agents are transforming customer interactions by providing instant, intelligent responses around the clock. They help businesses reduce operational costs, improve response times, and scale support without compromising quality. These agents understand natural language, learn from conversations, and integrate with existing systems to offer personalized experiences that enhance customer satisfaction and loyalty.

Chatzuri

AI-powered agents are transforming customer interactions by providing instant, intelligent responses around the clock. They help businesses reduce operational costs, improve response times, and scale support without compromising quality. These agents understand natural language, learn from conversations, and integrate with existing systems to offer personalized experiences that enhance customer satisfaction and loyalty.

Product

  • Pricing
  • Security
  • Affiliates

Resources

  • API
  • Guides
  • Blog
  • Help

Company

  • About us
  • Privacy Policy
  • Terms of Service
  • Cookie Policy
  • DPA

About

  • Teams
  • Singapore, Nairobi

© 2026 Chatzuri. All rights reserved.

Chatzuri uses AI and can make mistakes.

Terms of ServicePrivacy PolicyCookie PolicyChatzuri