The AI Agents Manager lists every agent you can access and is the entry point for creating, opening, sharing, and retiring them. Use it to find an existing agent, check what's deployed where, and take row-level actions like clone, share, archive, or delete.
Overview
The AI Agents manager has two tabs: AI Agents — the agents your team has built — and System AI Agents, which lists the platform's built-in agents. The first tab uses your workspace's name for agents, so it may show a custom label. The System AI Agents tab only appears if your role grants access to system agents; otherwise you just see your AI Agents list.
Click any row to open that agent's editor. Use Search agents to filter the list, use the header to create a new agent, and use the overflow menu to View Archived.
Agent Types
The Type column tags each agent so you can tell at a glance what kind it is:
Type | Description |
|---|---|
Simple | A lightweight agent built in the split-view editor. |
Advanced | A full agent with the complete configuration suite — instructions, tools, knowledge, memory, guardrails, and more. |
Workflow | An agent built as a visual flow. Available only in some environments. |
External | A third-party agent added today via Amazon Bedrock (see External Agents). The Agent-to-Agent (A2A) protocol is used for the reverse direction — exposing this platform's own agents to outside callers. |
Team | A group of agents, managed in the AI Teams area rather than listed here. |
New agents can be Simple or Advanced, depending on your role. Workflow agents are available only in some environments. Legacy RAG agents don't appear in this list.
Note: The word "Agent" itself can be relabeled per workspace, so your instance may show a custom name in its place.
What's on the Agent List
Each row shows the agent's Agent Name, Type, Tags, Shared with, Deployed On (icons for the channels it's live on), Owner, and Last Updated. A Workspace column appears when you're viewing across projects. The list loads 30 rows at a time and sorts by most recently updated.
Two more columns — Created On and Last Updated By — exist but are hidden by default. Turn them on from the table's column-visibility control, or hover the Last Updated date for the same detail as a tooltip.
Each row exposes the following actions:
Action | Description |
|---|---|
Edit | Opens the agent in its editor. |
Clone | Creates a copy of the agent, including its full configuration. |
Share | Grants other users or teams access to the agent. |
Archive | Takes the agent out of active use without deleting it. |
Delete | Permanently removes the agent. |
If you haven't created any agents yet, the list shows an empty state: No AI Agents — "You haven't created any AI Agents yet."
Cloning or Sharing an Agent
Clone starts a copy of an agent. It opens a dialog where you name the copy — it defaults to "{name} (clone)" — and pick the destination workspace, then opens the clone in its editor. The clone carries the agent's full configuration, including its skills.
Share opens the share dialog, where you grant other users or teams access to the agent.
Both actions are available from a row's actions menu and from inside the agent's editor.
Archiving or Restoring an Agent
Archive takes an agent out of active use without deleting it. The confirmation dialog, Archive AI Agent, warns that on archiving:
All knowledge updates will be stopped.
The AI Agent will be undeployed across downstream channels.
The AI Agent can still be restored later.
The dialog also shows how many teams and deployments the agent is part of, so you know the impact before confirming.
Archived agents live under View Archived (the Archived AI Agents list), where each row offers Restore or Delete. Restoring brings the agent back to the AI Agents list.
Deleting an Agent
Delete permanently removes an agent. Because it can't be undone, the Delete AI Agent dialog asks you to type the agent's name to confirm, and spells out the consequences:
All indexed knowledge will be deleted permanently.
Any linked knowledge sets will not be affected.
This action cannot be undone.
The dialog also offers an Export Sessions of this Agent checkbox, on by default, so you can keep a CSV of the agent's conversations before it's gone.
Note: Deleting an agent only removes its own indexed knowledge. Linked knowledge sets survive and stay available to other agents.