Estimated reading time: 5 minutes

A narrative growing louder in the African market is the appetite for intraregional trade. Historically, trade volumes between, say, Cameroon and France have been larger than with Cameroon’s regional neighbours, which paves a hard path to sustained growth.

As such, the idea that intraregional trade needs to increase is well established. However, the obstacles to this interconnectedness are also well established. From transport infrastructure to payment infrastructure, African systems are not fully insulated from external shocks.

Take roads as an example: roads were built in the Central African Republic (CAR) to facilitate oil exploration in the region, but the landlocked country has long suffered from meagre road infrastructure; the internal 2013 politico-military conflict divided the country and made road construction even more difficult. Impassable roads made northeastern CAR exceedingly isolated, putting the brakes on trade. 

Yet as they say, where there’s a will, there’s a way. Banks are speaking in unison about the benefits which intraregional trade offer both for profitability and for Africa’s stability. 

Can technology support?

Artificial intelligence (AI) is not going to pave a road from Bangui to Kinshasa. To deploy technology appropriately, it’s important to zero in on exactly which tasks AI can be leveraged to help with.

AI can make a stride towards mitigating the compliance case. Compliance is a point of friction so abrasive that it prevents financial institutions (FIs) from servicing markets.

The primary, well-cited factor is that AI can analyse large amounts of relevant data; generative AI (GenAI) can also speed up the process of gathering data. But the arguably more crucial point is that this analysis is particularly useful in building context. This helps financial institutions (FIs) make appropriate judgments on the clients they are considering doing business with. 

A second pain point is working with data that is not standardised or structured and building it into something that supports a bank’s internal process. AI has to accommodate the way governance works with FIs, not the opposite. Banks’ frameworks for managing risk have been developed over a long period of time and are validated with regulators: an AI can’t suddenly start making the decisions for them. What AI can do is feed in relevant information, normalising the data to standardise it across populations.

Third, banks thousands of miles away struggle to ascertain whether a seller in the CAR is selling flowers or narcotics. If AI allows organisations to pull more publicly available information or easily use historical data to contextualise a situation, we’ve established a degree of probability around the feasibility and bona fides of the CAR florist’s trade. 

Data collection, data processing, and data contextualisation: AI should be targeted as such.

With the fragmentation of the market, client experience has come into focus far more. Whilst legacy banks offer similar traditional services, fintech companies are gaining ground for their superior customer experiences that are more customer-focused and user-friendly. For banks to remain competitive and avoid being relegated to mere ‘money movers,’ they must prioritise customer experience through speed and transparency, leveraging AI to deliver these improvements whilst maintaining proper governance frameworks. In summary, AI can get around annoyances.

Assisting in compliance: transformation vs optimisation

AI enables banks to tackle complex regulatory challenges through modularised data workflows that can adapt to different jurisdictional requirements. This allows institutions to expand into previously inaccessible markets by scaling their risk analysis capabilities in real time.

There are two distinct approaches to implementing AI in banking compliance, each offering different benefits depending on an institution’s technical capabilities and risk appetite. 

• Through pre-validation, AI assesses the risk of every corporate, individual, payment, corridor, and invoice in real time, learning from historical data with a latency of less than a second. 
• For banks lacking the infrastructure for real-time integration, the easiest implementation involves optimising existing processes by deploying AI agents to handle routine tasks. Rather than having analysts manually review flagged transactions a month after processing, AI agents can complete nine out of ten standard investigative steps automatically, sitting alongside current processes without requiring extensive system integration.

Data workflows are extremely helpful for achieving compliance, allowing for much more flexibility and accommodating regulatory misalignments. The corridor between Kenya and Central Asia is vastly different from that between Germany and Central Asia or Kenya and Germany. The ability to ask different questions in real time essentially opens up a funnel.

Consider a simple example: in the US, a reporting threshold of $10,000 is different from in the European Union (EU), where the threshold is €10,000. An AI analytical workflow will know not to flag a $10,000 payment in a euro transaction since that’s not the regulatory threshold. 

In practice

A major African banking group approached Elucidate, the risk decision platform for financial crime, with an ambitious goal: to dramatically increase their payment volumes and trade flows into Central Africa’s most challenging markets, including Burkina Faso, Mali, and Nigeria. These were tough, heavily concentrated markets where traditional financial crime risk management simply wasn’t sufficient. The bank’s historical approach of processing everything first and investigating potential issues 30 days later made no sense in such high-risk environments with numerous red flags.

The solution involved deploying AI-enabled risk analytics directly into the payments flow, essentially “shifting left” the entire due diligence process from post-transaction monitoring to real-time assessment. This allowed the bank to define clear risk standards upfront and monitor for compliance instantly, providing immediate transparency to payment originators about what information was required. 

Rather than reducing traffic due to increased scrutiny, the bank doubled its payment volumes within six months, generating more fees whilst creating unprecedented transparency and enabling faster resolution when additional information was needed.

—

There is a precedent in Africa for turning weakness into strength. For instance, the African mobile payments landscape is extremely vibrant: in 2022, mobile money transactions in Africa hit $1 trillion (exceeding the GDP of most African nations). Legacy infrastructure on the continent was not a limiting factor, and they went straight to digital.

Certain challenges are out of the scope for AI to solve. But there now exists an opportunity to strategically deploy AI across the market in a way that allows for lower risk, to improve the integrity of markets, and most importantly, to create opportunities for wealth creation, which will surely come down to trade opportunities for small and medium-sized businesses.