Skip to main content
Admins get a view across the whole org — not just their own queue — which is where you spot bottlenecks and manage cost.

​
The Runs view

The Runs view is the org-wide monitor for agent and pipeline runs — status, failures, and HITL, at a glance.
Screenshot 2026 07 27 At 12 18 25 PM
At the top you get the numbers that matter day to day:
  • Live activity — how many runs are active right now, and how many are waiting on HITL review.
  • Awaiting human input — work paused for a reviewer, over the last 7 days and 24 hours.
  • Needs attention — failures that need looking at.
  • Runs (last 7 days) — total and succeeded.
Below that, a volume & outcome chart (succeeded / active / cancelled / failed over time) and top agents by run count, plus a live list of runs awaiting human input so you can see exactly where work is stuck.

​
Reporting

Reporting breaks down by process and by agent, so you can answer questions like “how much R&D work ran this quarter, and how much needed rework.” End users also get a self-service view of their own usage on their dashboard.

​
Using the data

  • If awaiting human input is growing, reviewers are the constraint — add or reassign them.
  • If a specific agent shows a high failure rate, investigate its data source and instructions (see When a run fails).
  • Use token trends to decide where BYOK or open-weight models would pay off (see Billing and credits).

​
Next steps

  1. Your dashboard
  2. Billing and credits
  3. When a run fails