Loading…
Agent Teams
Compose narrow, well-tested agents into teams that hand off, work in parallel, debate or follow a planner — and expose the whole team like a single agent.
Capabilities
A fixed pipeline where each specialist builds on the previous agent's output.
Fan work out to several agents at once and combine the results.
Explicit handoff edges route the conversation to the right specialist, with optional termination keywords.
Round-robin or manager-led speaker selection for review, debate and consensus.
A manager agent plans the task, assigns workers and tracks progress. Plan reviews land in Reviews.
The builder proposes roles, reuses agents you already have and drafts new specialists where needed.
How it works
Choose how members collaborate: pipeline, parallel, handoff, group chat or manager-led.
Reuse published agents or let the builder draft new specialists for missing roles.
Watch every member step, handoff and manager note in one activity timeline.
Publish the team to chat, embed, REST, the OpenAI-compatible API or A2A.
Transparent collaboration
Each step in the activity timeline is tagged with the member agent's name. Manager notes, plans and handoffs are visible, so a team is never a black box.
Manager
Plans the case
01Research agent
Gathers evidence
02Policy checker
Validates rules
03Writer
Drafts the reply
04Under the hood
The specifics your architects and security reviewers will ask for.
Works with
Every capability shares the same runtime, identity and governance.
Use a team when the path depends on judgment — agents deciding who acts next. Use a workflow when the steps are known and must run the same way every time. Workflows can also call teams as a step.
Teams are reachable over A2A, and agents can use external MCP servers and OpenAPI tools, so teams can work with systems outside ZeaIQ.
Book a working session with our team. Bring a real use case — we'll build it with you and show the governance and ROI views alongside.
