FRS, our fraud detection system, is built on a combination of machine learning and complex rules added by our analysts.
Machine learning is extremely efficient, but sometimes, interpretability can be a problem. That’s why we combine it with a rule-based system.
This hybrid approach combines the best of both worlds to give us the optimal safeguard.
Our machine learning and rule-based systems train on millions of fraudulent transaction data amassed from various sources.
These highly relevant signals come from principals, banks, payment providers, financial partners, fraud honeypots, and many other sources.
Even if an email address is new to Midtrans, chances are, that email has been reported in our partners' network. This lets FRS block suspicious networks from doing transactions.
Our fraud detection system is automatically built-in to our product, so you can be confident that your transactions are secure.
The system analyzes every single transaction, not just a sample of your records. Rest assured, all of your transactions are safe with us.
We know that each business has different needs. That’s why we have designed FRS to be flexible.
Your team can request our risk analysts to assess and include custom rules that may be specific to your business to the system.
You can also put verified customers to allow lists/white list to ensure smooth transactions in the future.
We built FRS in-house, from the ground up with a specific function: to detect frauds in e-commerce transactions.
This is important because this specificity makes FRS superior to third-party systems designed for more generic uses.
FRS is built into Midtrans.
Every transaction is screened.