Revolut uses machine learning to strengthen fraud detection



Customer application

  • Sherlock fraud detection

NoSQL solution

Use case

  • Fraud detection
  • User profile store
  • Digital communication
  • Caching


Key features

revolut white logo

The UK fintech company Revolut developed Sherlock, a machine learning-based fraud prevention system, to counter the growing threat of financial fraud. Sherlock continuously and autonomously monitors card users’ transactions, and if it finds a suspicious transaction it sends the user a push notification for their approval. Revolut selected Couchbase because of its inherent architectural advantages – including speed, agility, and scalability – that address the ever-changing data needs of users and merchants.

“For our customers, the loss of $100 can mean the difference between a pleasant holiday and an experience filled with frustration and resentment. Couchbase has never failed us or our customers.”

Dmitri Lihhatsov
Financial Crime Product Owner, Revolut


  • Fraudsters are evolving to beat traditional predetermined fraud detection rules

  • A mission-critical application required consistent high availability and high throughput for its rapidly growing customer base

  • On average, financial fraud costs institutions between 7-8 cents out of every $100


  • Sherlock’s high speed caching enabled machine learning algorithms to continually learn and update rules – catching 96% of fraudulent transactions
  • Sherlock evaluates transactions for signs of fraud in under 50 milliseconds for Revolut’s 12+ million customers
  • Within the first year in production with Couchbase, a 75% improvement over industry standards saved more than $3M