Catch the pattern before it becomes a habit.

Anomaly detection on voids, refunds and till discrepancies, built as an extension of the AI Agent's audit trail.

Loss Prevention setup guide

Estimated setup time: 10 min

Before you start

Setup steps

1
Enable Loss Prevention
Turn it on under AI Agent → Loss Prevention — it reads from the same audit trail already logging every action.
2
Set baseline sensitivity
Choose how far outside normal patterns a void, refund or discount rate needs to be before it's flagged, under Loss Prevention → Sensitivity.
3
Choose who gets alerted
Set which managers receive flags, and whether it's immediate or a daily digest, under Loss Prevention → Alerts.
4
Review your first flagged pattern
NUA explains what triggered a flag (e.g. one staff member's void rate is 3x the team average) — investigate or dismiss it.
5
Tune over time
Adjust sensitivity after the first few weeks — too sensitive buries real signals in noise, too loose misses them.

Tips

This flags patterns, not individuals — a flagged pattern is a prompt to look, not an accusation.
Start with a daily digest rather than immediate alerts, so you're reviewing patterns with context, not reacting to every single event.
Cross-reference flags with the roster under Staff & Rostering — some patterns are fully explained by who was on shift.

Troubleshooting

Too many flags to act on.
Lower sensitivity under Loss Prevention → Sensitivity — it's tuned too tight for your venue's normal variation.
A real issue wasn't flagged.
Check the sensitivity threshold isn't set too loose, and confirm the relevant action type (void, refund, discount) is included under Loss Prevention → Scope.

Related modules

Turn the agent from a monitor into an operator.

Rosters built from forecast demand, not gut feel.

Still stuck on loss prevention setup?

Email us directly, a real person on the team replies.