Investment fraud on the rise in 2nd half of 2025 (Belgium)

Belgium is experiencing a sharp acceleration in investment fraud — and the numbers from the FSMA’s latest dashboard leave little room for doubt.

Between July and December 2025 alone, Belgian consumers reported having lost more than €23 million to fraudulent investment schemes .

This is not a marginal increase. It signals structural evolution.

The numbers behind

According to the FSMA:

• 1,622 reports were filed in just six months

• €10.5M+ lost via fraudulent trading platforms and crypto-linked investments

• €9.5M lost through “exclusive stock tips” shared via WhatsApp groups

  
  

A new fraud model emerged in summer 2025:

• Social media ads lure victims into private WhatsApp groups

• Fraudsters impersonate economists or CEOs

• Victims are pushed toward fake crypto apps, manipulated US stocks (“pump & dump”), or data-harvesting schemes

Average loss per victim: €73,000  (yes 73K !!!) Some cases involve losses of several hundred thousand euros. This is not low-value scam traffic.This is wealth destruction.

And the reports keep climbing ... In 2025, the FSMA received 2,911 reports about unlawful activities — an 11% increase vs 2024. Since 2017, reports have grown by nearly 20% per year on average.

More than 65% of warnings published in 2025 targeted fraudulent trading platforms.

Fraud is industrializing.

A structural shift for financial institutions

The pressure is building.

• Customer expectations for reimbursement are rising.

• Trust is no longer based only on security — but on protection.

• Regulators are intensifying scrutiny.

• Fraud teams face heavier, more complex investigations.

And here is the uncomfortable reality:

Most of these payments were authorized by the customer. No stolen credentials. No system breach. No obvious rule violation.

From a traditional fraud engine’s perspective, the transaction looks legitimate.

Why Traditional Models Fall Short?

Most detection systems are built around:

• Single rules

• Static thresholds

• Segment-based deviation analysis

They are designed to catch hard anomalies.

But investment scams — and modern social engineering scams — exploit something else.

They exploit human judgment.

The risk signal is not always in the transaction rules themselves. It sits in the context, the narrative, the behavioral shift from a customer’s normal habits — and increasingly in scenarios deliberately engineered to bypass existing triggers.

These scams are designed around the controls.

Which means part of the risk remains structurally invisible to systems.

And that is the real shift.

The emerging model: prevention loops

This is not just an increase in fraud volume. It is a shift from:

• Unauthorized fraud → Authorized manipulation

• Credential compromise → Psychological compromise

• System exploitation → Human exploitation

And that changes everything, prevention must move:

➡️ from “Is this transaction technically valid?”

➡️ to “Does this situation make sense?”

That requires richer context, earlier intervention (shift left), and, most important, intelligence loops that learn from experts (the team) .

The new model becomes a continuous FraudOps loop — where experts and AI agents work hand in hand:

• AI accelerates data aggregation, alert triage, investigations and compliance reporting

• Human experts apply judgment used to feed knowledge base

• Insights are fed back upstream into prevention

Protection moves closer to the customer — becoming a real-time guardian, not only a back-office control. Earlier in the lifecycle. Before the loss happens. That is the real evolution of fraud prevention.

At Datavillage, we are focused on enabling this evolution — helping financial institutions transform investigations into upstream intelligence and bring contextual fraud prevention closer to the customer decision moment.