Bypass, refiling, grey routes and settlement leakage compound monthly on unmonitored corridors. We run detection as a discipline: per-corridor baselines on live CDR streams, machines proposing, engineers confirming inside a 15 minute window, recalibrated every quarter.
OTT bypass and SIM box converting international calls into local ones.
CLI refiling, spoofing and masking arbitraging origin-based rates.
A2P bypass, grey routes and pumping draining messaging revenue.
Settlement, VOD revenue share and mobile money reconciliation.
Plus quality-side monitoring that catches what fraud breaks: Voice Transit, Routing Quality and Mobile Money Monitoring.
Machine flagging alone produces a meaningful share of false positives, so every detector runs with a named engineer behind it and a measured error rate in front of it. That is the difference between detection as a product and detection as a discipline, and it is the reason our AI principles make human accountability principle number three. If you want the full framework, the wholesale fraud guide maps all four families and the five-layer prevention stack.
