Observatory
Digital fraud: the scenarios have changed, and so must our reflexes
Fraudulent messages have become credible and multi-channel. Understanding how they work lets you verify in the right place and respond according to what has been handed over.
Jeremy Kraft · 7 min read
In this article
The message no longer contains mistakes. The logo is in the right place. The person on the other end knows your name and your bank and can even imitate the voice of someone close to you. Digital fraud has improved in quality, but its underlying mechanics remain surprisingly stable: create a credible situation, trigger an emotion, then reduce the time available to check.
Understanding these mechanics is more useful than memorising a collection of fake messages. The settings change every week. The action requested, however, always comes back: hand over a secret, install something, approve a transaction or send money.
Fraud is often a story in several acts
We still picture the fraudster as a stranger sending a clumsy link. Today's campaigns instead build a journey. A text message announces a problem. A call follows a few minutes later. Someone introducing themselves as the "security department" repeats the information from the first message. Finally, they ask you to confirm with itsme or to move your money to a supposedly protected account.
Each channel reinforces the previous one. The text message gives a reason for the call; the call gives a human face to the fake website; knowledge of a few personal details creates the impression that the caller has access to your file.
This scenario explains why careful people can be caught out. They are not reacting to a single element, but to a coherent story crafted for them.
What artificial intelligence really changes
Generative tools make it possible to produce correct messages quickly in French, Dutch, German or English. They make personalisation from public information easier and make imitating a voice more accessible.
They do not, however, give fraudsters magical powers. A synthetic voice still has to convince you to act. A fake adviser still has to get you to disclose a code. A fraudulent advert still has to lead you to a platform controlled by the scammer.
Defence must therefore shift from appearance to procedure. A message can be perfect and still be fraudulent. A voice can be familiar and prove nothing. What matters is being able to confirm the request through a channel the other party does not control.
The most common scenarios
### The fake bank adviser
They tell you about a suspicious payment and use your worry to get you to approve a transaction. They may ask for a code, an itsme confirmation or the installation of remote assistance software. A bank will never ask you to transfer your funds to a "secure account" chosen over the phone.
### The fake invoice
The document seems to come from a real supplier, but the account number has changed. Sometimes the supplier's or customer's mailbox really has been compromised: the message then slips into an existing conversation. Any change of bank details must be confirmed with a known contact, independently of email.
### CEO fraud
An employee receives an urgent, confidential request supposedly sent by a manager. Hierarchical pressure here replaces fear about the bank. The fraudster is counting on the person preferring to comply rather than disturb their superior.
### The fake investment
An advert, sometimes featuring a well-known figure, promises an exclusive method or an innovative platform. An "adviser" then calls regularly. The dashboard shows gains, but these figures do not correspond to any real asset.
### Fraudulent recovery
After a first loss, a company promises to recover the money for a fee. Victims are targeted because they have already paid and want to put the situation right quickly. The second fraudster often belongs to the same network or uses lists that have been resold.
Urgency, trust and shame
These three emotions structure most frauds.
Urgency prevents comparison. Trust comes from a logo, a familiar name, a voice or a personal detail. Shame appears after the first action and delays reporting. The fraudster sometimes uses all three in this order: "act now", "you can trust me", then "don't tell anyone, you've made a mistake".
Naming this sequence helps to break it. When you feel pressure to act without a witness, the right decision is precisely to introduce a delay and a second pair of eyes.
The independent channel rule
For any request involving money, a code or access:
1. break off the conversation;
2. use neither the link nor the number provided;
3. open the official app yourself or use a number you already know;
4. explain the request to someone else before confirming.
This method works even against a very convincing imitation. The fraudster controls their message, their website and their number. They do not control the banking app you installed long ago, the number printed on your card or the colleague sitting next to you.
In a family, agreeing on a question or a code word for unusual financial requests can also help. It is not absolute proof, but it is an effective obstacle when a call uses an imitated voice.
If you clicked without handing anything over
Close the page. Do not download anything or fill in any field. If a file was opened or an app installed, disconnect the device from the network and ask for technical help. A single click does not always mean a compromise, but you need to look at what actually happened.
Forward the suspicious message to `suspect@safeonweb.be`. Reports feed Belgium's blocking mechanisms and can protect other people.
If you disclosed a password
Change it from a clean device, starting with your email. If that password is used elsewhere, change it on the other accounts too. Revoke active sessions and check the recovery methods, connected apps and automatic forwarding rules in your mailbox.
Then enable a second factor. An authenticator app is useful; a security key or passkey offers better resistance to fake websites.
If money or a bank approval is involved
Contact your bank immediately through its official channel. Timing can determine whether a transaction can still be blocked. For a card, Card Stop can be reached on `078 170 170`.
Keep the exchanges, numbers, addresses, URLs, payment references and times. File a complaint with the local police in the event of a loss or extortion. Do not pay any "recovery fees" offered afterwards.
Designing an organisation where people can say stop
Awareness fails when it asks employees to be infallible. A better organisation makes dangerous requests hard to carry out: dual approval of payments, verbal confirmation of account changes, named accounts, phishing-resistant MFA and a clear reporting procedure.
People must also be explicitly given the right to halt an urgent request, even if it seems to come from management. The employee who calls to check is not slowing the company down: they are protecting its cash.
The right indicator is not just the number of people who click during an exercise. You need to measure the time to report, the quality of the information passed on and the team's ability to contain the situation.
Official sources
Editorial information
- Written by
- Jeremy Kraft
- Last reviewed
- Method
- Public sources, editorial review and proportionate guidance.
Discussion
Comments
No comment yet. Feel free to open the discussion with a question or your own experience.