AI in Anti-Money Laundering


H3M.IO KROTON modules works hand-in-hand with legacy rule based AML systems and upgrades them with the help of Artificial Intelligence. KROTON product suite includes 5 AI based modules which powerfully combines compliance know-how with data science:


  • KROTON Rule Optimizer optimizes the parameters of existing rules, improving detection performance. Rule Optimizer is powered with Active Learning technology, so not only the rule parameters and thresholds are optimized, missing features in the rules are detected as well.
  • KROTON Suspect Miner utilizes machine learning to find suspicious activity not covered by rules. There are more than 500 features that can be used in detecting money laundering activities. Suspect Miner detects the ones that are too complex to detect with legacy rule based approaches with the help of artificial intelligence.
  • KROTON Graph Miner analyzes graph networks and finds suspicious groups of activities and accounts. Some accounts or customers may not be suspicious when analyzed exclusively, however they can be detected when analyzed as a group. Graph Miner utilizes AI to search through billions of sub-networks in a fast and effective manner.
  • KROTON Link Miner finds suspicious relations in the network. Some accounts cannot be detected by statistical anomalies neither by investigating alone nor by a network analysis. Link Miner uses proprietary algorithms to identify missing links in the suspicious networks.
  • KROTON Rule Miner identifies missing rules in the systems. Some of the findings of Suspect Miner, Graph Miner and Link Miner may give hints about missing rules in the system. Rule Miner utilizes the outputs of these modules and its extensive rule library to help compliance departments detect missing rules in the system.

About Money Laundering

Money Laundering is a global problem: International Monetary Fund (IMF) estimates that about 2-5% of global GDP (up to 2 trillion USD) is involved in money laundry. On the other hand, there are trillions of transactions globally, and it is very hard to detect suspicious operations. In the last decade, banks in US and EU are fined over 340 billion USD related to AML cases.