There are multiple acceptable versions of every address in SA: This “almost accurate" data is threatening compliance and banking KYC

There is a 50 million-point complexity problem in South African location data upon which compliance and banking Know Your Customer (KYC) depends: there are five times more acceptable descriptive "names" for places than there are actual "places".

While South African banks invest in artificial intelligence and advanced risk modelling as a means of strengthening their compliance, it doesn’t cater for the fundamental data quality issue undermining their KYC processes: incomplete address information.

The gap between “acceptable”, "sufficient" and "complete" address data represents a significant compliance risk and operational burden for financial institutions. It’s one that criminals love exploiting, and impacts the financial institutions’ efficiency and profitability.

The hidden complexity of South African addresses

The challenge is not a lack of data, but how that data is structured and understood, according to Johannes du Plessis, Sales Manager and Geospatial Expert at AfriGIS. "Many institutions can find a customer," he explains. "But that doesn't mean they have the full picture. If you don't have the complete story, you introduce uncertainty into every decision that follows."

The challenge is particularly acute in South Africa's diverse address landscape, which encompasses everything from traditional suburban addresses to farm portions, informal settlements, and rapidly evolving urban developments.

To put the scale of this fragmentation into perspective, South Africa has roughly 8 million to 10 million unique addresses – yet there are over 50 million valid aliases for those locations.

This means each property has several legitimate identifiers, from the street-facing delivery address to the municipal erf number used in bond registrations. Without a way to link these variations, financial organisations essentially have a fragmented view of their customers.

When fragmentation becomes risk

A single location in South Africa can legitimately exist under multiple formats: street addresses, stand numbers, farm portions, sectional title references, or estate naming conventions. These variations are not errors but different representations of the same place, used across legal, municipal, and everyday contexts.

"What we see is that one property can have multiple valid versions of an address," says Du Plessis. "If those aren't connected, you're effectively working with fragments of the same truth."

This fragmentation creates direct compliance implications. FICA and KYC processes rely on the ability to verify identity, assess risk, and detect anomalies. When multiple versions of an address aren't linked, critical links or signals can be missed.

"Different documents may use different versions of the same address," Du Plessis explains. "If those aren't recognised as the same location, you lose the ability to connect the dots."

The operational impact

The implications extend beyond compliance into daily operations. Address data is often captured and stored differently across departments, from onboarding to credit, insurance, and collections. Without standardisation, the same customer may exist in multiple formats within the same institution, leading to duplication, rework, and inefficiencies.

This creates particular friction in digital onboarding environments and disproportionately affects customers in less formal address environments, such as farm portions or informal settlements. If systems cannot accommodate how a location is actually described, onboarding becomes unnecessarily complex and can even become exclusionary.

The solution lies in approaching address data differently. Rather than treating an address as a single input field, the shift is toward structuring it as a set of standardised, validated components linked to a precise geospatial location.

This includes capturing multiple address aliases for the same property and connecting them through a single spatial reference point. "The key is that spatial link," explains du Plessis. "Once you tie all those versions to one location, you can start to build a much richer understanding.”

From there, additional information layers can be associated with that location: property data, business registrations, demographic indicators, and risk-relevant datasets such as crime or environmental factors.

Unlocking new opportunities

Complete address verification enables banks to onboard previously "unverifiable" customer segments while maintaining rigorous anti-money laundering standards. When enriched with demographic, crime, and weather data layers, verified addresses become powerful tools for risk assessment and business intelligence.

"If you're doing insurance on a property, you need to understand flood risk, fire risk, and crime patterns in that specific location," Du Plessis explains. "For banking, complete address data helps detect anomalies between declared income and property values, or identify potential beneficial ownership issues in sanctions screening."

This enhanced data also supports strategic decisions about matching branches and services across different locations.

Balancing precision with inclusion

Crucially, this approach doesn't require forcing all customers into a single address format. In South Africa's diverse environment, inclusivity depends on accommodating different ways of describing a location while ensuring those descriptions can be verified and standardised behind the scenes.

"You absolutely need to support different ways of capturing an address," emphasises Du Plessis. "If you don't, you exclude people. But if you can link those variations, you can still be inclusive and maintain compliance."

This type of flexibility allows financial institutions to extend services into areas that have traditionally been difficult to onboard, without increasing risk exposure.

As financial institutions invest in advanced analytics and AI-driven decision-making, there's growing recognition that these capabilities depend on the quality of underlying data.

"There's substantial focus on modelling," Du Plessis notes. "But often the issue isn't the model; it's the data “going into” it."

Improving address completeness supports functions across the organisation, from compliance and risk management to customer experience and strategic planning. When the foundation is solid, everything built on top of it becomes more reliable.

A practical next step

For many organisations, the most effective starting point is treating address data as a shared, standardised asset. This includes centralising address datasets, enabling structured capture across systems, and ensuring that different representations of a location are linked and validated.

"If you fix just one thing, fix this," says Du Plessis. "It strengthens compliance, improves customer experience, and reduces risk across the business in a lasting, foundational way."

As regulatory expectations shift and change, and competition intensifies, the ability to move from "almost accurate" to complete, verifiable data is becoming essential for doing business, both locally and globally.

In an environment where precision is paramount, addressing the address problem may be the most practical step financial institutions can take to strengthen their operations.