01 · Ask
Start with the outcome
Use requests such as “Fill tomorrow’s open shifts” or “Find two guards for Parramatta from 6pm to midnight.”
Status → Context → ActionRequest understood
Fill tomorrow’s open shifts. Avoid unnecessary overtime.
Roster Copilot
Tell Iron Watch what needs covering.
Ask for the roster outcome in operational language. Iron Watch interprets the request, works through valid options and presents a roster for authorised review.
See Iron Watch in ActionOperational view
28shifts tomorrow
27covered
0credential conflicts
96%roster confidence
Human-controlled decision support
02 · Verify
Deterministic rules first
AI is not the source of truth for licence dates, credentials, availability, conflicts, shift facts or tenant boundaries.
- Only eligible guards move to ranking
- Future-date credential checks
- AI failure does not stop core rostering
Status → Context → Action50 guards checked
Eligibility rules found 17 valid options for the open shift.
03 · Explain
Recommendations with reasons
The scheduler sees why a candidate is ranked: availability, credentials, site familiarity, workload and other authorised factors.
Status → Context → ActionMichael is the strongest match
Available, inducted, current credentials and no shift conflict.
04 · Control
Approve the roster
Roster Confidence is based on measurable coverage, credential, availability and unresolved-assignment conditions — never an invented AI score.
Status → Context → ActionRoster proposed
27 of 28 shifts covered. One requires attention.
Ready before it becomes urgent
Put the operation in one controlled environment.
See how Iron Watch connects guards, shifts, credentials, sites and operational decisions.