What is Transaction Risk Scoring?
Transaction risk scoring is the process of assigning a risk score to each individual transaction based on how likely it is to involve money laundering, fraud, or other financial crime.
Transaction risk scoring gives every payment a measurable level of risk. The score combines details about the transaction itself, the customer behind it, and the counterparties involved. Compliance teams then use that score to decide whether a transaction should go through, be held for review, or generate an alert.
Transaction risk scoring works alongside customer risk scoring. Customer risk scoring rates the overall risk of a relationship, while transaction risk scoring rates each payment as it happens. A low-risk customer can still make a high-risk transaction, and transaction risk scoring is designed to catch exactly that.
What factors go into transaction risk scoring?
Transaction risk scoring typically weighs several types of signals, including the following.
- The transaction amount compared with the customer's normal activity
- The payment corridor and the risk level of the countries involved
- Whether the beneficiary, device, or corridor is new for this customer
- The customer's risk tier, PEP status, and adverse media flags
- Recent velocity, such as how many transactions the customer made in the last hour or day
- Links to accounts or entities already flagged as suspicious
Transaction risk scoring often combines these signals into a single composite score. A wire through a high-risk corridor to a first-time beneficiary, at three times the customer's normal amount and soon after a device change, would score far higher than any one of those factors alone.
How is transaction risk scoring used?
Transaction risk scoring drives what happens to each payment. In real-time monitoring, a score above a set level can trigger an automatic hold, a request for extra authentication, or a referral to an analyst before the payment settles. Lower scores let legitimate transactions pass without friction.
Transaction risk scoring also helps teams prioritize alerts. Investigators can work the highest-scoring cases first, and plain-language reason codes show which factors pushed each score up. Linking transaction scores to dynamic risk scoring keeps each decision aligned with the customer's current risk level.
How do rules and models work together in transaction risk scoring?
Most firms combine rules and machine learning models for transaction risk scoring. Rules act as firm gates for risks that always need review, such as a sanctions match. Models then rank remaining transactions by learning from past cases, using signals like amounts relative to the customer's baseline, corridor familiarity, and network closeness to known illicit activity. This hybrid approach keeps scoring explainable while still catching complex patterns.
