How artificial intelligence and machine learning are changing the remittance process

  • Tech
How artificial intelligence and machine learning are changing the remittance process

The main characteristic of the 4th Industrial Revolution is the abundance of data. Businesses and industries are quickly moving towards value addition by collecting, analysing and evaluating this valuable data. 

We wrote a piece on big data, so this is as good a time as any to take a look at it before we proceed.

The financial technology industry is naturally suitable for the application of data science techniques because of its heavy reliance on data, digital systems and databases. However, it is rife with repetitive, mundane and time-consuming tasks. Artificial intelligence (AI) and machine learning (ML) can automate most of these tasks, allowing us to make better use of the valuable insights gleaned from this data.

At Tranglo, we have already begun using AI and ML to improve our fraud detection and financial reporting processes. We use this enhancement to expand our payment network by  creating a more robust ecosystem to integrate different payment rails. 

But there’s so much more. AI and ML have already made a positive impact in many key areas of the remittance industry. Let’s explore. 

Better forecast of global remittances 

A macroeconomic model was built to forecast remittances based on “Monte-Carlo simulation and artificial intelligence”. This model has global application and is useful for analysing the long-term predictions of remittances particularly in scenarios where the sender is from a rich country and the receiver is from a developing country. 

The creators of this model claimed that “by changing the socio-economic characteristics of the countries involved, experts can analyse new socio-economic frameworks to obtain useful conclusions for decision-making processes involving development and sustainability”. 

From a business perspective, the fact that we can predict remittance flows means we can develop tools to strengthen key areas of a process, allocating adequate resources that follow studied trends. This enables us to craft a business model that not only meets the needs of our clients, but also benefits the masses.  

Smart, self-learning remittance system that drastically reduces human error

AI is now advanced enough to provide end-to-end remittance solutions with embedded decision-making and ML models. Such a self-learning system continuously provides insights on data. It is connected to a company’s enterprise resource planning systems and provides highly accurate forecasts of cash flows across all operational and non-operational paths. It does this by analysing and learning about the company’s general ledger transaction data, financial planning and historical bank statements.  

AI can predict dates for invoice payments, validate deductions and forecast defaults. It can even forecast cash flows at both operational and non-operational levels, down to an individual invoice level. With it, leaders can make informed decisions, improving dispute resolution.

A typical end-to-end automated system will electronically capture remittances, process them directly from email attachments without the need for human intervention and reconcile receivables payments directly into corporate account receivables. This significantly reduces human error.

Virtual assistants that answer before you ask

Our heading may be a hyperbole, but AI-powered virtual assistants, or chatbots, are improving user experience in the remittance industry. It replaces a human customer service representative and is available 24/7. 

The chatbots are built on AI models and are trained on extensive data sets (i.e. the possible number of questions that can be asked by a question). Furthermore, the self-learning nature of these chatbots means they automatically learn, record and improve their capabilities when they encounter a new situation. 

In 2018, a remittance provider launched an AI-powered chatbot which makes international money transfer with the help of an SMS. In 2021, the same remittance company launched Pan-Africa remittances chatbot which makes remittances possible with a single SMS without connecting to the Internet.

Cognitive automation of accounts payable

AI is automating low-value and mundane tasks e.g. calculating and managing accounts payable. Typically, companies have to pay USD16-22 to process invoices manually. However, Goldman Sachs, an investment banking company, says that this cost can be reduced to USD6-7 by automating accounts payable. 

AI-powered cognitive automation eliminates the need for human intervention. It not only focuses on automating multiple tasks, but also learning and improving the entire process. 

Safeguarding line-item data

In a typical low-value paper remittance advice process, companies have to pay key-per-stroke. That is why most prefer to store only key information or payment header information of the remittance process, causing significant loss of line-item data. AI and ML can resolve this problem in every step of the process, i.e. email, paper and customer portal. 

AI can capture data directly from check stubs. ML models can be trained using existing or parsed remittance data to learn important keywords based on which remittances can be processed.

This means companies not only save on key-per-stroke costs, but also prevent the loss of  line-item data. This data can then be used to analyse underlying patterns. 

State-of-the-art fraud detection with AI

One of the biggest challenges faced by the financial industry is fraud. According to Javelin Strategy & Research’s “The 2021 Identity Fraud Study”, total fraud losses in 2020 surpassed USD56 billion. A fraudulent transaction that occurs in any bank or remittance provider not only causes financial loss, but also reputational loss. 

AI applications here are manifold. AI-powered models can quickly learn patterns from historical data, i.e. the spending patterns of an ATM cardholder. It can show a warning when that ATM cardholder is spending at an unusual place and therefore warn the bank to take preemptive action. According to the Wolfsberg Group, AI is also being used for sanctions screening, “a control employed within Financial Institutions (FIs) to detect, prevent and manage sanctions risk”.

Similarly, the model can also learn our shopping behaviour, and companies can earn a fortune by providing data and partnering with marketing companies.  

Time-saving financial reporting

On the financial reporting front, AI has a very broad scope for improving the reporting process, freeing accountants from gruesome and repetitive tasks by automating most of the processes. 

A literature review has been conducted on Applications of AI in the financial reporting process. It states that “[Artificial Intelligence can do] Management of accounts receivable and payable, updating customer data, suppliers, invoice processing, automation of authorisation, as well as validation and posting of payments, cash receipts, billing and invoice adjustment with purchase and sales orders”. 

Closing thoughts

Undoubtedly, this is an era of AI and ML. Data science has a lot to offer to the fintech industry not only in terms of modified security and fraud prevention, but also financial reporting, decision-making and strategy formulation.  

Stay updated

Get weekly articles in your inbox for free.

    By submitting this form, you acknowledge that you have reviewed the terms of our Privacy Statement and consent to the use of data in accordance therewith.