FailSafe Radar FAQs
How does Radar differ from other AML and blockchain analytics solutions like Chainalysis, TRM Labs, and Elliptic?
Radar utilises advanced ML to analyse wallet behaviours and transaction-level heuristics that indicate suspicious activity in real-time, while legacy tools are primarily investigation tools that follow this process to determine risk/threat:
First, investigators obtain ground-level truths (evidence is collected through manual research, web scraping, and data partnerships with exchanges and law enforcement) that allow them to attribute addresses to real-world entities (addresses are then labelled accordingly – known scammers, terrorism, gambling sites, etc.);
Then, clustering algorithms are utilised to identify all addresses associated with these attributions, which allows an in-depth view of the wallets that are related to the entity.
There are challenges to this approach – accuracy of data is an issue, as recently exhibited with Mastercard-acquired CipherTrace, a company that had to suspend operations due the inaccuracy of data; the second issue is speed – existing solutions aren't real-time data providers (as the process first requires investigators to submit attributions, after which clustering is applied), meaning fraudsters can easily circumvent detection by spinning up new wallets and on-ramp routes. Furthermore, we're receiving feedback from our customers that Radar serves a use case that the incumbents just aren't providing a solution for: detecting payment fraud and identifying suspicious activity in near real-time that may not be directly attributed to known entities. For example, incumbents aren't able to identify suspicious activity on addresses that we're identifying as part of Sybil/bot clusters and money washing schemes. While using Chainalysis is great for companies who need to demonstrate to their regulators that they are exercising every effort to be compliant, they're not solving for fraud that's ultimately causing harm to customers.
Last updated