Fraud detection and prevention software
See the fraud moving through your own data, and stop the payment before the money leaves.
Trusted by 130+ financial institutions across 30 jurisdictions
Fraud detection and prevention software watches payments and customer behaviour for the patterns of fraud (scams, mule accounts, structuring) and decides whether a payment should go through.
Salv's fraud detection software monitors around 96 million transactions a month across payments, customers and counterparties, and decides on each payment while it's still in flight. When scams change, adjust the rules yourself and tests changes on your own history first.
Over 130 banks, EMIs and payment institutions across Europe run their fraud detection with Salv.
- 1bn+transactions handled a year
- 250M+end-customers covered
- ~96Mtransactions screened and monitored a month
You can usually see the fraud. Stopping it is the harder part.
A customer talked into an investment or an impersonation scam keeps paying until they work out what has happened, and by then several payments have been made.
That money moves on every rail you run, and Salv covers account-to-account, instant and card payments. Instant payments leave the least room, because the decision has to happen in the payment window, not in tomorrow's alert queue. Every loss also lands twice: once as a refund, and again in what the customer tells everyone else about you.
What fraud teams are up against, and how Salv's fraud prevention software can help
-
Change your fraud detection as the scams change
Start from a library of scam scenarios or build your own, then tune thresholds, lists and segments yourself.
Scenarios can also spot warning signs in your customer data, not only in payments — like several customers sharing the same phone number. Every change is tested on up to a year of your own history before it goes live, so you know what it will catch before it raises a real alert.
-
Stop the scam payment before the money leaves
Real-time scenarios return their decision in the API response while the payment is still in flight.
Your system can then hold, release or reject the payment before the money moves. A scenario can check the payment's screening alerts first, so it only rejects once screening is clear. Every alert shows which scenario fired and which transactions triggered it, so your team can see exactly why.
-
When the rest of the case sits at another bank
When the money has already reached another bank, you can take the case to them directly.
If that bank is on Salv Bridge, you send a structured request, conformant with the EPC's FRIDA rulebook. A person there checks the account and replies within rules the network has agreed. The receiving bank then acts under its own obligations.
What to ask any fraud detection software vendor
Each answer below is a mechanism you can ask to see in a demo rather than a claim you have to take on trust.
Why it matters
Scams change faster than a vendor release cycle.
How Salv answers
Thresholds, time windows, lists, segments and suppression intervals are yours to change, and you add your own scenarios to the library.
Why it matters
A flag after settlement is a refund, not a prevention.
How Salv answers
Salv returns its decision in the API response and can carry the action in a webhook; your own system holds, releases or rejects.
Why it matters
An analyst and a supervisor both need the reason.
How Salv answers
The alert names the scenario and lists the transactions that contributed to it.
Why it matters
An untested change lands on the alert queue.
How Salv answers
Yes, on up to a year of your own history, with no real alerts raised.
Why it matters
Outbound and inbound fraud look nothing alike.
How Salv answers
Investment, job and impersonation scams, money mules and structuring, plus patterns in your own customer data such as shared contact details or a cluster of similar new accounts.
Why it matters
Most of the case is outside your own data.
How Salv answers
A structured request to that bank, if it is a member of the network.
Fraud tools that work together
-
Intelligence sharingThe secure network where fraud, AML and sanctions teams settle a case directly with other institutions. Send structured requests, answered in minutes, under rules your network agrees.
-
MonitoringScenarios you can read, test on your own data and change yourself. Watch payments, customers and counterparties, for AML and fraud.
-
Alert and case managementWhen an investigation grows, it becomes a case. Bring the related alerts, customers and payments together and work it from one place as a team.
Fraud detection software FAQ
Software that watches payments and customer behaviour for the patterns of fraud and decides whether a payment should go through. In a bank it sits alongside sanctions screening and case management, and the test worth applying is whether the decision arrives while the payment can still be held.
Detection finds the pattern; prevention acts on the payment before it settles. Detection on its own leaves you writing refunds. A system that is useful to a bank does both.
Three things. Outbound scams, where your own customer is being talked into paying. Inbound mule activity, where the receiving account belongs to a customer of yours. And the part of the case that sits at another bank. Detection built for one of the three leaves the other two open.