Start here after every run—the bench expands the goal tree and shows how constraints roll up to the top-line mission.
Run the demo to populate the goal tree snapshot.
Need to rerun a single loop? Hit “Nudge” and that agent will rescore its part of the engagement.
Each card summarizes the mission, burn, guardrails, and last output for that agent—perfect for onboarding new viewers.
Chronological log of what each agent just did. Pass numbers match the simulator timeline.
Use the tabs to inspect the artifacts each agent produced during the last loop.
Artifacts load here after running the demo.
Wire the bench into your production estate. Choose a connector, drop credentials, and the fleet will orchestrate live interventions.