EXCLUSIVE: "Fraud Rings Don’t Work in Siloes, So Why Do Banks?" - Baiba Miezere and Hacène Kaci, Eastnets in 'The Fintech Magazine'

Criminals thrive when data is in the dark. Eastnets offers one solution to help shed light on billions of dollars leaking out of the banking system every year

2025 saw the dramatic USA arrest of 13 so-called call centre employees. Together, they scammed more than $5million from 400 unsuspecting grandparents, aged 84 on average. Complete with motivational ‘make every call count’ mantras, target boards and scripts, the teams rang unsuspecting victims posing as their grandchildren in need of bail. The amount of money coerced ranged from $3,200 to $25,000 per call, according to a whiteboard police found on the wall.

Clearly, this is fraud. Impersonation fraud to be exact. In this case, the stolen funds were collected as cash (often with scammers taking their elderly victims to the cashpoint). Runners or money mules then deposited the funds into various accounts, known as layering, before moving them through the banking system and, ultimately, into the hands of an alleged fraud ringleader in the Dominican Republic.

More commonly, money is coerced from victims electronically. Once a bank customer realises they’ve been tricked and reports it to their bank, a fraud team leaps into action. But while staff are busy inputting the details, it’s often already too late. The money is being quietly dispersed across potentially hundreds of crypto and fiat accounts in small amounts (smurfing), making it hard to track and almost impossible to recover.

At this point, though, the bank’s anti-money laundering team don’t even know the crime has happened because, in the siloed data environment in which many banks still operate, departments don’t share information. While criminal money is zipping across the world in seconds, the bank hasn’t even started to join the dots. Banks are aware of this, and there is a growing trend to unify segregated fraud, AML and sanctions screening teams into AFC (anti-financial crime) teams or FRAML (fraud and AML) teams in an attempt to stop fraudulent transactions being authorised.

Unifying staff is one thing; when it comes to unifying the data, though, there are no easy choices. It often requires surgery to the core or a major data migration exercise, both of which are costly and carry their own risks. Eastnets, a payment and compliance solutions provider, offers another alternative – an orchestration layer that screens transactions to reduce what is currently a cripplingly high number of false flags, makes intelligent predictions, and surfaces information from vaults of disparate data to identify potential fraud-busting connections in real time. It could help stop a crime in its tracks, or at the very least, give banks and law enforcement a sporting chance of hunting down those behind it.

“With instant payments moving in seconds, it’s vital to have one risk view that allows banks to make decisions equally fast,” says Baiba Miezere, Group Product Development Director at Eastnets.

The company’s new FinCrime Intelligence Platform connects a range of tools that work across KYC/KYB, AML transaction monitoring, fraud prevention, name and transaction screening, giving everyone eyes on a potential crime.

Data convergence is key, says Hacène Kaci, Pre-Sales Consultant at Eastnets: “We see a lot of data silos in banks where they have onboarding, ongoing due diligence, fraud, screening and transaction monitoring teams working separately.”

But beyond that, the industry needs to ‘get smarter at working collectively on sharing that data’, he adds.

Frictionless protection

Research from Avaloq found that some financial professionals have as many as 15 different applications open at the same time as they try to conduct compliance checks across multiple data stacks in one organisation.

“They spend 90 to 95 per cent of their time switching between systems to collect data and build evidence logs instead of making a decision,” says Miezere. “Our FinCrime Intelligence Platform brings data from different systems – our own or any other vendor’s – into one view.

“Let’s assume that in AML we’ve detected a fraud ring working globally. And now let’s assume that in our bank, we have identified authorised push payments frauds among, typically, our vulnerable older customers. We link that information together and give that intelligence to fraud prevention.”

Crucially, it aims to do this without creating additional unwanted friction for the customer – payments prevented unnecessarily, onerous additional identity checks.

“Our fraud prevention tool combines transaction monitoring with a very efficient fraud layer using behavioural biometrics,” explains Miezere. “It’s very challenging because a bank can’t add friction, neither operationally nor from a customer experience point of view.”

Kaci agrees: “The rule here is not to have a huge amount of false positives, but neither should you miss designated entities. We help financial institutions get that balance right by providing the tools to screen and also to reduce the false positives, by using either a rule-based approach, whitelisting, or AI.”

The legacy conundrum

Worldwide, 80 per cent of financial institutions have pinpointed legacy infrastructure and fragmented data as a major constraint to identifying and preventing fincrime. What’s more, they need to find ways to share that data across regions, without breaching any national or supranational regulations, like local politically exposed persons lists, data privacy laws or the constantly updating sanctions lists, driven by rapidly changing geopolitics, triggering massive money movement, not all of which is legitimate.

“Sanctions change overnight,” says Miezere. “Our World-Check On Demand engine updates sanctions lists every few hours and not only rescreens that data but also provides automated fuzzy logic matching, irrespective of which country that person is sanctioned in.

“It’s very important that banks have that flexibility, speed and accuracy. We help banks to navigate flexibly by combining different data sets behind different lists. You want a local list? Here you go. You want an international list? You can combine them flexibly.”

Professional fraud rings, she highlights, have global footprints bigger than most banks.

One professional fraud ring, for example, targeted more than a million victims across 121 countries with demands for E-ZPass toll payments. It amassed more than $1billion globally, using government impersonation scams and sending 330,000 phishing texts a day. Finding a common denominator can be extremely helpful in unravelling such seemingly overwhelming cases. Unleashed from the false positive mountain by a system that scans using biometric and behavioural markers, bank staff using Eastnets’ FinCrime Intelligence Platform can focus on contextualising the data insights, says Miezere.

“Our artificial intelligence solution helps to identify gaps. Then it tries to identify patterns sitting among different layers,” she adds. “It makes intelligent links between different systems and also across different countries.

“With instant payments moving in seconds, it’s very important that we have that one risk view that allows banks to make decisions equally fast, using auditable, compliant, AI solutions.”

Rise of the machines

Criminals don’t work in siloes and they adapt to and around new technologies with ever-changing tactics. The democratisation of AI, for example, has seen a dramatic rise in the number of synthetic identities created. Research by LexisNexis found that 11 per cent of fraud now involves a fake identity. That figure rises to 48.3 per cent in Latin America. Other criminals have blended societal changes with AI to adapt their scams to modern times. For example, in the UK, five men were jailed in April for creating fake identities on dating websites to hoodwink singles into £2.37million-worth of romance fraud. 

Predictably, they then laundered the money back to themselves. Reported $579billion global losses from AI-driven financial fraud and scams last year are just the tip of a very grubby iceberg. The true cost to the global economy has been estimated at closer to $3trillion, and banks and their customers are not the only victims. In South East Asia alone, it’s believed that 300,000 people have been trafficked into scam compounds where they are imprisoned and exploited. 

Scam rings have even been identified operating from within prisons and pre-trial detention centres. This is especially prevalent in Colombia, Ukraine, South Africa and Russia. Shutting down these criminal networks matters on a global financial and humanitarian scale.

Banks have the skills and capacity to adapt just as fast as criminals – if they can break through the data silos, reduce false positives and screen continuously for constantly updating regulations. Their ultimate weapon, though, is collaboration. As each institution seeks to improve its own data management systems and compliance processes, they can achieve more than the sum of their parts by working together, suggests Kaci.

 As more criminal patterns are detected – through open data, collaboration and with AI tools – blocking fraud will become pre-emptive rather than reactive. The world of siloed data – and siloed thinking – is coming to an end.