Tasks page

Monitor, troubleshoot, and resolve Agent and Workforce tasks in bulk from one place

The Tasks page is where you monitor everything your Agents and Workforces are doing — reviewing outputs, replying to escalations, approving pending actions, and rerunning failed runs. You can act on a single task or apply the same action across every matching task at once.

Tasks from both Agents and Workforces appear here. The same triage and bulk action flows apply to both.

The Tasks page is the project-level monitoring surface. For the per-Agent view inside the Agent builder, see the Run tab.

Tabs

The page is split into tabs, with a subheading below the tab bar describing what the active tab contains:

• All tasks — every conversation across your Agents and Workforces.
• To review — tasks waiting on an approval before they can continue.
• Escalated — tasks the Agent paused on to ask you a question.
• Errors — tasks that hit a tool failure or runtime error.

Side panel

Selecting any task opens a side panel with three sections:

• Output — what the Agent produced and a summary of the actions it took.
• Chat — message the Agent about the task, ask follow-ups, or guide a rerun.
• Details — created timestamp, status, credits used, metadata, schedule, linked Tools, linked Agents.

Failed tool steps show an Edit tool action on hover. Clicking it opens the Tool builder in a new tab so you can fix the failing step and publish. Come back to the Tasks page to rerun the affected tasks.

Bulk actions

When you open a task in the side panel, the page checks for other tasks with the same shape and surfaces a bottom sheet so you can act on the whole group at once. Matching means the same Agent, the same failing tool step (or the same escalation / approval type), and the same payload shape — for example, a Python tool that threw the same KeyError across three conversations would group into one bulk action.

Available bulk actions:

• Reply to escalations with a single shared response.
• Approve or Reject pending approvals.
• Rerun errored tasks from the failing step.

A confirmation modal shows the count before anything is sent (for example, “Reply to 4 agent tasks”, “Approve 3 agent tasks”). Tab counts refresh after the operation completes.

Bulk actions only appear for tasks with prefilled inputs ready to review — tasks that need you to type a custom reply or pick from a dropdown stay one-at-a-time. Up to 50 matching tasks are processed in one bulk run; for larger batches, run the action again, since already-queued tasks are excluded from the next match.