By Oluoma Onyeukwa
The person who steals your money may never speak to you, open the account receiving it or personally move a single transaction. Those tasks can be distributed across an organised network of specialists, automated tools and financial intermediaries.
Fraud is no longer simply a criminal deceiving an unsuspecting customer. It has become a technology-enabled industry with its own infrastructure, distribution channels, specialised services and revenue models.
Yet too much of banking’s response still begins and ends with identifying one suspicious transaction.
I believe this mismatch is becoming one of the financial industry’s most serious weaknesses. Criminal networks increasingly operate like interconnected technology businesses, while financial institutions remain constrained by organisational boundaries, fragmented intelligence and controls that often respond after money has moved.
INTERPOL’s 2026 Global Financial Fraud Threat Assessment describes how organised criminal networks are using artificial intelligence, specialised money-laundering services and fraud-as-a-service platforms to expand their operations.
The implications are significant. Fraudsters no longer need to possess every technical capability themselves. Criminal services can supply impersonation tools, fraudulent identities, phishing infrastructure and access to accounts used to move stolen funds.
The same technologies making legitimate businesses more efficient are reducing the cost of committing fraud.
Consider how a modern scam can unfold.
A fraudulent advertisement reaches a customer through a digital platform. An AI-assisted impersonation convinces the victim to transfer money. The funds enter a receiving account, move through several others and are eventually withdrawn or spent.
Each stage may involve a different organisation. Yet each institution often sees only the fragment passing through its own systems.
Criminals see the network. Banks often see the transaction.
This is the fundamental weakness I believe financial institutions must address.
The scale of the problem is difficult to ignore. UK Finance reported £1.28 billion in payment fraud losses during 2025. Its 2026 report also found that 66% of authorised push payment fraud cases originated online and another 17% through telecommunications.
These figures expose an uncomfortable reality: banks are frequently expected to prevent the final financial consequence of fraud that began and developed outside the banking system.
But the solution cannot simply be more customer warnings, additional authentication prompts or increasingly aggressive account restrictions.
Fraudsters adapt. When one channel becomes harder to exploit, they test another. When one account is blocked, they use a replacement.
The UK’s Financial Conduct Authority provided particularly revealing evidence in September 2026. Its research found that fraud proceeds were commonly withdrawn or spent after passing through two to five money-mule accounts. More than 238,000 suspected mule accounts were closed by financial firms in 2025.
Those closures matter. But closing accounts is not the same as dismantling the infrastructure that repeatedly supplies them.
If criminals can replace compromised accounts faster than banks can identify the networks behind them, the industry is measuring disruption without necessarily achieving it.
Technology must change that equation.
Banks should move beyond evaluating transactions predominantly in isolation and develop stronger capabilities for detecting relationships between accounts, devices, beneficiaries and suspicious payment patterns.
Graph analytics can help reveal networks that conventional transaction monitoring may overlook. Machine learning can identify unusual connections and emerging fraud patterns across multiple signals. Real-time intelligence sharing can help receiving institutions recognise suspicious funds before they disappear through further transfers.
But these capabilities become substantially more useful when institutions cooperate.
A fraudulent payment should not become another bank’s problem simply because the money has crossed an institutional boundary.
I would argue for a shared fraud-intelligence framework that allows authorised financial institutions to exchange high-confidence risk indicators, coordinate interventions and trace suspicious transaction networks without unnecessarily exposing customers’ personal information.
Technology companies and telecommunications providers must also participate. Platforms where scams originate cannot remain peripheral to prevention while banks and victims absorb the financial consequences.
There is, of course, a serious danger in expanding automated surveillance. False positives can freeze legitimate accounts, damage reputations and exclude innocent customers.
That is why shared intelligence must include strict data safeguards, independent oversight, explainable decisions and effective routes for customers to challenge mistakes.
Fighting organised fraud should not mean treating every customer as a potential criminal.
Success must also be measured differently. Rather than celebrating only the number of blocked transactions or closed accounts, banks should examine how quickly fraudulent networks are identified, how effectively criminal funds are intercepted and whether disrupted operations reappear through new accounts.
Fraud has become an industry because criminals understand the value of operating as a network. Until financial institutions become equally effective at defending the network, they will continue winning individual battles while struggling to dismantle the business behind the crime.
About the Author
Oluoma Onyeukwa is a banking and financial operations professional with experience spanning fraud and card operations, financial data and cloud technologies. Her professional interests include financial crime prevention, data analytics, digital banking security and the responsible use of emerging technologies in financial services.
