Support Learning and Insight

It’s more important than ever to understand the challenges facing financial executives. Support the Financial Education & Research Foundation today.


How US Businesses Set to Benefit from the Use of AI in Banking

by David Dua

As banks begin to offer businesses the visibility they require, there is increased adoption of technologies that make use of artificial intelligence and machine learning.

For SMEs, gaining visibility and control of spending is essential to the financial health of the company and pivotal to maintaining adequate cash flow. However, this is often a challenge with research from Fraedom finding that 40% of SMEs in the US have a clear picture of spend at the end of each month but little insight on a day to day basis. This lack of visibility has resulted in 22% of SMEs having to regularly spend significant time and money investigating who spent what. As banks begin to offer businesses the visibility they require, we are seeing increased adoption of technologies that make use of artificial intelligence (AI) and machine learning (ML). Subsequently, businesses are set to benefit from a wider range of capabilities, tools and controls with AI likely to have a major impact on the following aspect of a business’ finances: 

Reducing incidences of fraud
Currently, businesses lose an average of 7% of their annual expenditure to fraud, however, as banks make better use of AI for fraud detection this figure will drop. The effectiveness of AI for fraud detection is demonstrated by its use at Visa, which has reduced global fraud rates to less than 0.1%. Overall, businesses will benefit from improved security features as AI is used to detect anomalies in their accounts and fraudulent activities much quicker than previously possible. This works by the model having an understanding of what is ‘normal’ for each account or card and recognising patterns based on past transactions and behaviours. For example, if 99% of the transactions for one account happen Monday to Friday, a transaction that occurs at the weekend will be seen as abnormal and flagged as such. Of course, anomalous transactions aren’t always fraud. Often, they’re just out of the ordinary, requiring some more investigation and flagging them to the business would allow for this. These new technologies will ensure businesses are able to deal with discrepancies in their accounts immediately, rather than finding out about them months down the line when it’s harder to get a clear picture of events at the time when the transaction took place.  

Greater controls
AI will allow banks to more accurately forecast how much credit businesses require and limits on spending will be set automatically, enabling businesses to gain a better understanding of their spending. This can then be implemented within the organisation as it will enable businesses to redistribute credit limits based on what different employees regularly spend. This means that credit will be allocated in an optimal way, ensuring the amount of credit employees are given reflects their spend history. This ensures that those employees who often make large transactions are given the credit to do so, while those who use their company accounts for lower-cost transactions don’t receive as much, therefore ensuring that credit is being used to the greatest effect. 

Simplifying expense management 
In addition to providing enterprises with a greater degree of control and understanding of their finances, banks are also beginning to use AI to offer businesses extra tools and services. A prime example of this is expense management systems which use AI to simplify the expense process and reduce the amount of time employees and finance departments spend on such tasks. As with fraud detection, the system would establish patterns based on the employees historic spending behaviour. For example, it may pick up that once a week the sum of $10 is spent in a coffee shop which the user then applies a particular expense code to. Once this behaviour has been demonstrated enough times, it becomes a pattern. So, the user will no longer have to code the transaction themselves, the system would automatically identify the type of expense it is and code it correctly. 

As the system establishes more patterns and understands what the user or business is doing, smart coding could start to be applied to a greater number of transactions. This would significantly reduce the amount of time spent manually sorting through and coding expenses as the employee then only has to check that the correct codes have been applied. 

As banks make greater use of AI and ML, they will ultimately build up a fuller and more accurate picture of their business customers. This will be of inestimable value to organisations and SMEs in particular – providing them with a greater level of control over their accounts, improved visibility and a better understanding of their finances. As this is realised, businesses will begin to reap the rewards of their employees spending less time manually interrogating accounts and instead being able to focus on more value-adding tasks. 

David Dua is a Data Science Stream Lead & Principal Data Scientist at Fraedom.