Financial criminals are adept at constructing convincing paper trails. Falsified identity documents, fabricated company registrations, AI-assisted forgeries, and ghost addresses can all fool a KYC process built entirely around records. What they cannot fake is physical reality. Location intelligence is one of the most powerful fraud detection tools available to South African financial institutions, and most are only beginning to use it.
The spatial fingerprint every customer leaves behind
"Everybody has a spatial fingerprint," says Marna Roos, Senior Client Consultant at AfriGIS. "Every time you put down your address at the bank, we can derive a spatial location for it. Every time you swipe your card at a business, a location is attached to that transaction. Through time, it creates a picture that revolves around a core area. That footprint can be triangulated around your home address and your work or business address, and it proves that you are really living where you say you are.”
A fraudster constructing a synthetic identity still has to build years of coherent spatial behaviour, a cost that typically exceeds the return. "It is cheap to create a piece of paper," says Roos. "It is not that cheap to put a physical building onto a piece of land. Fraudsters will always follow the path of least resistance, and spatial reality is expensive to fake."
When an address is geometrically impossible
One of the most immediate fraud signals that spatial data surfaces is the impossible address: one that looks perfectly formatted but cannot physically exist. Roos regularly receives requests from banking clients to verify addresses their own systems cannot resolve.
"I’ll investigate the address and find that it has no physical way of existing," she says. "The format looks perfect: a street name, a suburb, a number. But when you explore that suburb, the street numbers only reach as far as the number fifteen, which means number fifty cannot exist.”
A subtler but equally revealing pattern emerges when a registered business address is cross-referenced against its spatial context. Roos describes a scenario common enough to be instructive: "Everything checks out on paper. The CIPC records are fine, the company is registered, the directors' ID numbers all verify. But when you physically examine the location in our data, it resolves to a farm portion in the Free State with no commercial activity, no points of interest, no built-up coverage. That is where interlinking the data sets starts to build a spatial picture of what is actually happening."
Beyond flat deeds data: the context that changes the risk picture
Many financial institutions already hold deeds data. When a customer buys a property, the transaction is registered and the bank holds a record of ownership. But raw deeds records – the stand description, the registered town, the transfer date, the price – answer only a narrow question: who owns what.
"A deeds record tells you who owns a stand and what they paid. It does not tell you where that stand is, or whether the price makes any sense for its location. We add the street address, the real-world anchor, and suddenly a ten-million-rand valuation sitting in an area where properties trade at five hundred thousand becomes an immediate red flag. The record looked fine, but the location tells a different story," Roos explains.
She points to a broader principle at work: the value of AfriGIS's data lies in the combination of datasets. "It is not just the address or just the deeds record," Roos says. "It is deeds, the national address database, the cadastre, CIPC, points of interest, gated community extents, formal settlement layers, all cross-referenced. Any one of those datasets can be worked around individually. Together, they create a picture of physical reality that is extremely difficult to fabricate, and that is exactly what you need when you are deciding whether to trust what someone has put on paper."
Defensible compliance
South Africa's initial greylisting by the Financial Action Task Force has placed compliance officers under significant and lasting pressure, regardless of being removed from the list last year. FICA and AML obligations are not new, but scrutiny of whether those obligations are being met intensified considerably and have now become the new normal.
Roos summarises the shift that spatial intelligence enables in a single sentence: "Banks will move from a tick box exercise to compliance that is defensible in a court of law”. A process that can demonstrate it cross-referenced a customer's declared address against verified spatial data, checked a business registration against land use classifications, and flagged an anomaly that a conventional KYC review would have missed is a process that holds up to scrutiny.
One of the fastest growing centres of demand for spatial intelligence AfriGIS is currently witnessing is specifically among organisations that need to focus on fraud prevention.
The shift to context-rich spatial intelligence also changes the operational dynamic from reactive to proactive. "With machine learning, once you start building systems that network all this spatial information together, you begin getting automatic notifications that something is amiss," explains Roos. Multiple shell companies registered to the same suburban residential property, a business address resolving to a vacant plot, a customer's declared home in a province where they have never once transacted: each is a signal that a spatially aware system surfaces automatically.
Keeping pace with increasingly sophisticated criminals
Fraud syndicates adapt. Any control mechanism that becomes widely adopted will eventually attract attempts to circumvent it. Roos is candid about this: "Any security mechanism you put in place, eventually someone will find a way around it. That is why it is critical to keep data updated and relevant, and to always be looking for additional data sources that can help surface new anomalies."
Roos says this is exactly why AfriGIS updates its deeds dataset monthly and its CIPC data weekly, and why it is working so hard towards more frequent updates of cadastral and address data. AfriGIS has spent nearly three decades enriching spatial data, but emphasises that consistent updates are what turn such an asset into an invaluable one.
“Fraud, money laundering, illicit shell companies, etc. – they all operate on the principle that mismatches between recorded data and reality can be blurred and then exploited. Throw generative AI into the mix and criminals see even more opportunity to hide their false information under layers of apparent legitimacy,” asserts Roos. “The physical, real world of things and places isn’t so easily ‘faked’. Connecting physical reality to information that is traditionally digital or paper-based and thus proxies for reality is a powerful way to establish truth when falsehood is the tool criminals rely on.”
