Stopping fraud before it happens

We're proactively preventing fraud with realtime payment tracking powered by advanced, serverless technology. 

Fraud prevention - desktop hero

Staying ahead of fraud

Traditional fraud detection relies on historical data, such as past spending behaviour, payment patterns or known fraud signals, to flag suspicious transactions. With this approach, fraud is addressed after it has already happened. But what happens when a fraudster creates a new account? There’s no transaction history to compare it against, making it hard to spot fraudulent activity.

But we’ve changed that.

We record every single payment across every banking channel in real time, using a serverless AWS Aurora database to keep your money safe. Combined with other new technologies, including Memgraph, Flink, Kafka, and SageMaker, our fraud prevention system instantly identifies potential threats.

This real time risk scoring enables us to evaluate payments and proactively block fraudulent transactions before they happen, instead of reacting after the damage has been done. This means that even if we’ve never seen a payment to a particular bank account before, we can instantly determine whether it poses a risk.

When we haven’t seen payments to a particular account recently, or ever, you may be asked to complete a quick security screening. This helps us confirm that the payment is legitimate and protects you from scams.

Blocking payments to known scams

Our app is now directly connected to our fraud centre in real time. What does this mean for you?

When you try to pay a high-risk recipient, you’ll immediately receive a warning. If a beneficiary has already been reported for fraud, our system blocks the transaction before money leaves your account. Even scammers who open brand-new accounts at other banks can be identified and blocked if there are confirmed fraud reports linked to them.

This is a first in South African banking, real time, proactive fraud prevention, proudly built by our own software engineering team.

And our latest fraud prevention measures are proving to be successful.

So successful that an individual complained that they had to open a different bank account to invest in Tiger Agriculture, now exposed as a Ponzi scam, because we had blocked payments to the company.

Why? Because our system proactively flagged it as a scam. And it was right.

This is just one example of how we're protecting your money by staying ahead of fraud.

Lastly, bank impersonation fraud is huge. Scammers know how to imitate bank officials, they even play the banks’ ringtone in the background, transfer you to their managers and send you OTPs by SMS to authenticate. We’ve solved this, our app will warn you if you’re not talking to us. Don’t ignore it.

How much has our AI saved clients from scammers?

Our AI-driven fraud prevention tools saved clients R673 million in the 2026 financial year alone. We stopped more than 394 000 scam payments and blocked 131 000 fraudulent recipients in real time. We also shut down over 80 000 mule accounts used to move stolen money between banks, disrupting the networks scammers rely on to cover their tracks.

How you can protect yourself from scams

  • Always double-check payment details before you send money
  • Never pay for items purchased online from online portals such as Facebook marketplace. Always inspect the goods first or meet the service providers in person
  • Watch out for High Return or get rich quick Investments or Earnings Platforms advertised on social media, WhatsApp or TikTok. These are scams
  • Never share your login details or passwords
  • Watch out for red flags, if something seems too good to be true, it probably is
  • We will never call you to ask to make payments using cash send, vouchers or real time payments to prevent fraud. Our app will warn you if you’re not talking to us
  • Keep your app updated for the latest security features
  • Trust our security warnings if we flag a payment as suspicious

Last updated: 8 May 2026

Was this article helpful?

152

23

You may also like...