How AI Could change Finance

The pandemic forced financial institutions to migrate online as lockdown measures were set in. However, they lacked the technical infrastructure and were unprepared - creating errors, fraud and delays. In response to this, interest in AI has accelerated to tackle these very issues. Here are some of the applications of AI in banking to help them navigate this new online territory. 

Exploiting the uncertainty, phishing attacks by scammers have spiked during the pandemic, commonly pretending to represent COVID-19 relief efforts. The jump in online transactions, as brick and mortar sites were forced to shut down, meant that employees struggled to police the digital fraud. However, AI is known for pattern identification and so could analyse irregular behavior such as frequent high-value transactions in a short window and report these findings. 

AI works via machine learning; analysing and refining itself from historical data. The pandemic itself is unprecedented and therefore so are problems that come with it such as fraud. This means that the AI will take time to adapt to the new types of digital fraud that will be born out of the pandemic. Therefore, the most effective applications of AI will involve a human to review and analyse AI recommendations to prevent any false positives.

With historically low levels of consumer confidence, a depressed mortgage market and high default rates, bank profit margins have been hard hit. To compensate, financial institutions need more targeted marketing campaigns; by offering e.g. specific credit card offers, the customer is more likely to buy and use credit on the card. AI could help facilitate this by using tools to analyse historic client records and purchase histories to identify key interests. For example, if the customer has a passion for sports clothes; the bank could offer e.g. JD vouchers in credit card packages. With each ‘deal’ tailored to the specific client.

While the need for the first application has been highlighted by the pandemic, like target marketing, our final utilisation of AI will be needed far after COVID-19 ends. Humans are prone to error, especially during monotonous, consistent and often boring tasks. Such is the case in data entry where employees often input the wrong numbers. AI on the other hand functions on consistent guidelines and therefore can pick up on any data entry mistakes such as personal details on credit applications not matching official documents.

Therefore, the applications of AI in finance are diverse ranging from fraud to marketing; offering both immediate support to cushion the effects of the pandemic and long term solutions.

Thumbnail credit: https://itpeernetwork.intel.com/future-ai-banking/