Understand Operational Replay.

See how exported simulation evidence becomes five readable story moments—and how to distinguish an observation, a recommendation, a later simulated task record, and an outcome without implying command execution or causation.

Back to Operational Replay Download MP4

Recorded presentation

Operational Replay narrated tour

01:20

Simulation results — not live customer dataRead-only presentation — no control or write-backEnglish narration uses an AI-generated voice

How to read the replay

The default view is a recovered simulation-reference transcript; its original source files are not embedded. Recommendations remain advisory. A later simulated task or outcome record is separate evidence and does not prove that the recommendation caused it.

Read the English transcript

Operational Replay turns exported simulation artifacts into a five-step browser story: state, risk, recommendation, simulated task, and outcome. It is not an operating console.
This is the recovered reference transcript; its original files are not embedded. The amber banner labels the entire replay as simulation.
At noon, H1—the only handoff station—goes offline for two recorded hours. Collection continues, but the transfer point into washing is unavailable.
At twelve forty-six, R3 also fails and needs human help. A deterministic briefing flags the assistance backlog; this is not a learned-model claim.
After H1 returns, the baseline recommends unloading R3’s full payload. Execution remains unconfirmed: the recommendation is advisory, not a robot command.
The next screen separates evidence: a shield-allowed decision and R3 later recorded at the handoff. The map shows snapshots—not motion or proof of causation.
The terminal replay shows full service availability, zero stockout minutes, and three assists. These are placeholder-provenance simulation results, not customer performance. Selected artifacts never leave the browser or write back. FacilityState and the simulator remain untouched.