> For clean Markdown of any page, append .md to the page URL. > For a complete documentation index, see https://help.sigmacomputing.com/llms.txt. > For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://help.sigmacomputing.com/_mcp/server. # Manage Sigma agents for your organization > Review usage and monitor agents with observability across your organization. > **Note** > > Usage of this feature uses consumption credits. For details about what usage is billed, see [About billable usage events](/docs/about-billable-usage-events). > **Note** > > The use of AI features is subject to the following [disclaimer](/docs/notice-for-enabling-ai-enabled-features-in-sigma). As an admin, gain observability into agent creation and usage across your organization. Monitor how many agents are being created, by whom, and for what use cases. To track usage more closely, refer to the [AI usage](/docs/ai-usage) dashboard. ## Requirements * You must be assigned the Admin [account type](/docs/create-and-manage-account-types). * To see an agent on this page, the workbook containing the agent must have been published (or a version tagged) after August 24, 2026. If you do not see an agent you expect to see, publish or re-tag the workbook. ## Manage agents across your organization 1. Go to **Home** and select the user icon with your initials. 2. Select **Administration** > **Agents**. On the **Agents** page, you can review every agent published across your organization and monitor adoption and usage trends. Review trends for the last 7 days, 30 days, 60 days, or 90 days. ![The Agents page showing metric tiles for active agents, questions, token usage, and active users over the last 30 days, and a table listing 3 agents with their location, owner, and usage details.](https://fdr-prod-docs-files-public.s3.us-east-1.amazonaws.com/sigma.docs.buildwithfern.com/6de8743f579c93961629c3ca6f3c6883521f2a9e1b737d7649de040f839cdee2/assets/docs-images/admin-agent-list.png?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Content-Sha256=UNSIGNED-PAYLOAD&X-Amz-Credential=AKIA6KXJSKKNFOCF7G4B%2F20260929%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Date=20260929T030505Z&X-Amz-Expires=604800&X-Amz-Signature=2f1ff11b8399f583fbccfc61846a0a59b72391facec7b65785af4a2701f7f9fe&X-Amz-SignedHeaders=host&x-amz-checksum-mode=ENABLED&x-id=GetObject) For all the agents that you can access as an admin, review KPIs with details about the agents in your organization: * **Active agents**: The total number of agents used in the selected time window. * **Questions**: The total number of messages sent to all agents. * **Token usage**: The total input and output tokens consumed across all agents. * **Active users**: The number of distinct users who interacted with an agent. An agent table lists every agent in your organization with details about the location, workbook owner, and last updated date for the agent name or description. Sort the table by **Users active**, **Last invoked**, or **Tokens** to find your highest-usage, recently used, or highest-cost agents. > **Info** > > **Users active**, **Last invoked**, and **Tokens** require the AI usage dashboard to be set up. If tracking isn't configured for your organization, these columns don't appear. To set up the AI usage dashboard, see [Configure the AI usage dashboard](/docs/configure-ai-usage-dashboard). ## Review specific agent details and usage On the **Agents** page, select an agent to review more details. You can view a summary of the agent, showing the agent's location, names of data sources, names of configured tools, and instructions. ![The agent details page for Traffic Overview Analyst, showing metric tiles for questions, active users, tokens, and rating, a Usage overview tab with runs by source, and a panel listing the agent's location, data sources, tools, and instructions.](https://fdr-prod-docs-files-public.s3.us-east-1.amazonaws.com/sigma.docs.buildwithfern.com/d95208c757717ba89c1b17a3f43ccf6a8c6c793b0ce52c21553c9200e12baf4a/assets/docs-images/admin-agent-details.png?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Content-Sha256=UNSIGNED-PAYLOAD&X-Amz-Credential=AKIA6KXJSKKNFOCF7G4B%2F20260929%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Date=20260929T030505Z&X-Amz-Expires=604800&X-Amz-Signature=f6349f8af2433bb7b3869c9e3de750b7fab5be0738e2e341d7cd8318bce6174e&X-Amz-SignedHeaders=host&x-amz-checksum-mode=ENABLED&x-id=GetObject) For each agent, you can review the following for the selected time period: * **Questions**: The total number of messages sent to the agent across all versions. * **Active users**: The number of distinct users who interacted with the agent. * **Tokens**: The total input and output tokens consumed by the agent. * **Rating**: The sentiment rating provided by users of the agent, relative to the total number of messages. Users can provide feedback on agent responses by clicking (**Good response**) or (**Bad response**). To monitor specific agent activity, review the details in the tabs: * **Usage overview**: Displays all conversation turns by source, for example a chat element, automated action, the API, or the home page. Review the number of turns in conversations, percentage of total conversation turns, and tokens consumed for each source. * **Execution log**: Lists individual conversations by session, including when each started, who invoked it, its source, execution path, tokens consumed, latency, and outcome. Filter the log by source or error type. Click **View details** on a conversation to review the details, including each turn's question and response. * **Errors**: Displays failed runs by error type, including the number of failed runs, percentage of failures, and when each error type was last seen. Click **View runs** on an error to open the **Execution log** tab filtered to that error. * **Access**: Shows who has access to the agent, inherited from its source workbook. See [Manage access to an agent](#manage-access-to-an-agent). ### Manage access to an agent Access to an agent is controlled by the access grants on the source workbook. If a user is assigned an account type with the **Create and manage agents** permission enabled, they can also create an agent in a workbook. | Workbook access | What it allows | | :-------------- | :-------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **Can view** | Send messages to an agent attached to a chat element. If the user has API credentials or access to the Sigma MCP server, can call the agent from the API or the MCP server. | | **Can explore** | Send messages to an agent attached to a chat element, and change which agent is attached to the chat element in a custom or saved view. | | **Can edit** | Chat with an agent in a chat element, schedule an agent to run automatically, or call an agent programmatically using the API. | As an admin, if you use the [API with impersonation](/docs/impersonate-users#impersonate-users-for-api-calls), you can impersonate a user with **Can view**, **Can explore**, or **Can edit** access to the workbook to call the agent on their behalf. To audit who currently has access to a specific agent, open the agent details and select the **Access** tab. The **Access** tab lists every user and team with access to the agent's source workbook, how each user or team was granted access (directly on the workbook, inherited from a folder or workspace, or as the workbook owner), and how many of the users used the agent in the selected time period. To change who has access to an agent, click **Manage agent access on the workbook** to open the workbook's sharing settings. ## Related resources * [About Sigma agents](/docs/sigma-agents) * [Build Sigma agents](/docs/build-agents) * [AI usage dashboard](/docs/ai-usage)