Knoxy capability · Watches
Fraud and anomalies caught before money moves
Fraud and anomaly detection is the Knoxy capability that watches scans, claims, redemptions, payouts and orders across IMAST 360, spots activity that doesn't fit normal patterns, and alerts the right person before rewards or payouts go out.
- Scans, claims, redemptions, payouts and orders
- Patterns across products, not one screen
- Alerts to the person who can act
The problem
Why channel programs need anomaly detection
Any program that pays rewards or incentives attracts some misuse. The common forms are well known: the same QR code or invoice claimed twice, edited invoice amounts, many accounts run by one person, and redemptions that happen faster than any real partner would make them.
Rules catch the obvious cases, but the costly ones often only show up when you look across data — a partner whose claims rise while their orders don't, or a cluster of new accounts redeeming from the same device. That is the kind of pattern people miss in daily reports and an AI can watch for continuously.
Knoxy, the AI Chief of Staff in IMAST 360, adds that layer. It works alongside LoyaltyBoard's AI fraud detection and Reward Max's anomaly checks, and brings what it finds to the person responsible.
What Knoxy watches
Where anomalies show up
Knoxy looks across the products your teams already use.
Scans
Duplicate and unusual QR scans by device or location.
Claims
Duplicate or out-of-pattern claims.
Redemptions & payouts
Unusual redemption speed and payout patterns.
Orders
Order patterns that don't fit a partner's history.
How it works
From signal to decision
What happens when Knoxy spots something.
- 1Watch
Continuous checks
Knoxy watches activity across loyalty, rewards and orders.
- 2Flag
Anomaly found
Activity that doesn't fit is flagged with the reason.
- 3Alert
Right person told
The loyalty, finance or sales owner gets the alert.
- 4Decide
Your team decides
Your team reviews and acts before the payout.
Best practice
Making fraud controls work without slowing partners
Most partners are honest, so controls should stay out of their way. Automatic checks on every transaction do the routine work; people only look at what is flagged.
Start with identity. Digital KYC with PAN and Aadhaar eKYC before redemption removes a large share of risk up front, and makes every later alert easier to act on.
Feed decisions back. When a flagged case turns out to be genuine — a large retailer that really does scan many codes — note it, so the same partner is not held up again. Over time, alerts become fewer and more useful.
- Automatic checks on every transaction
- KYC before the first redemption
- Clear owners for loyalty, finance and sales alerts
- Review outcomes to sharpen future alerts
Making it work
Turning anomaly alerts into fewer losses
Detection only saves money if someone acts on it. Decide before launch who reviews each type of alert — program managers for scan and claim anomalies, finance for payout anomalies, sales managers for order patterns — and how quickly they should respond.
Review outcomes as well as alerts. When a flagged case turns out to be genuine, note why. A large retailer that legitimately scans many codes, or a seasonal spike in a region, shouldn't keep raising the same alert; adjusting rules for that segment keeps attention on the cases that matter.
Combine detection with prevention. Digital KYC before redemption, unique product codes, OCR checks on invoices and bank account validation for payouts remove much of the risk before anything needs to be flagged.
Finally, share what you learn. Patterns found in one program or region often appear elsewhere, and updating schemes and rules across the channel closes the gap for everyone.
- Assign an owner for each alert type
- Record why flagged cases were genuine or not
- Pair detection with KYC, unique codes and OCR
- Apply lessons across programs and regions
FAQs
Fraud and anomaly detection FAQs
What is anomaly detection in a loyalty program?
It is the automatic spotting of activity that doesn't fit normal patterns — such as duplicate scans, out-of-pattern claims or unusually fast redemptions — so it can be checked before rewards are paid.
What does Knoxy check?
Knoxy detects fraud and anomalies across loyalty scans and claims, redemptions and payouts, and orders in IMAST 360.
How is this different from LoyaltyBoard's AI fraud detection?
LoyaltyBoard's AI fraud detection checks loyalty activity such as redemption velocity, duplicate claims, device fingerprints and location. Knoxy works across IMAST 360, so it can also bring in rewards and order data and alert the right person.
Who gets the alerts?
Knoxy alerts the person responsible — for example the loyalty program owner, finance or a sales manager.
Does KYC still matter?
Yes. Digital KYC with PAN and Aadhaar eKYC ties each account to a verified person and is one of the most effective controls.
Will honest partners be slowed down?
Automatic checks run in the background; only flagged activity needs review, so most partners are not affected.
Is there an audit trail?
IMAST 360 keeps immutable audit logs.
Who should review fraud alerts?
Assign owners by type: program managers for scans and claims, finance for payouts, and sales managers for unusual order patterns.