RiskDesk · Review queue
Manual review
Risk analyst — your shiftM
↩ samples
TXN-0000
€0
Risk score
0
050100
Payment
Customer · device · network
Rules triggered

Interactive HTML5 · simulation · gamified · fintech

You're on the payments risk desk

Flagged transactions are waiting in your review queue. Open each case, read the rules that fired, and clear it — fast. Over-blocking good customers costs the business as much as letting fraud through.

The model risk score is a hint, not the answer — on some cases it points the wrong way, and reading the signals is the job. Every shift is freshly generated, so you can keep training.

ApproveSignals look clean. Release the payment.
DeclineClear fraud pattern. Block before it settles.
EscalateSignals conflict — send to a senior analyst, don't guess.

Keyboard: A approve · D decline · E escalate.

Shift complete.

Cases cleared. Here's how your decisions landed.

A Decision gradeSenior-ready
0
Good payments approved
0
Fraud caught / escalated
0
Fraud let through
0
Good customers over-blocked
Net impact to the business€0

What a senior analyst reads first

  • Velocity beats amount. A tiny charge after five declines in two minutes is card-testing — block it, cheap as it is.
  • Mismatch > single red flag. One foreign IP isn't fraud. IP, BIN country and address all disagreeing is.
  • Ambiguous ≠ decline. When good and bad signals tie, Escalate is correct. Snapping to Decline quietly bleeds real revenue.
See more formats →