AI in banking and finance. Stripe, Monzo, and Grab | AI in business #78

AI in banking – introduction

Artificial intelligence is already widely used in many areas of the banking and financial sector. It’s not just chatbots for customer service or well-secured applications. Artificial intelligence is being used in the financial industry for even more serious purposes. Here are the main applications of AI in banking:

• Fraud detection and prevention – advanced algorithms analyze transactions in real time and detect suspicious activity patterns. This effectively protects customers from scams,
• Optimization of financial liquidity forecasting – AI-based predictive models analyze vast amounts of data to precisely predict future cash flows and manage liquidity more accurately.
• Streamlining processes related to creditworthiness assessment – here, too, machine learning algorithms come to the rescue, which, based on the analysis of thousands of credit applications, can accurately assess a customer’s financial credibility,
• Personalization of offers and recommendations for clients – banks utilize advanced recommendation models to tailor financial products to individual customer needs,
• Automation of back-office processes – routine tasks, such as document verification or transaction settlement, can be fully automated with the help of AI.

However, how did companies operating in global markets cope with the implementation of these innovations?

Stripe: transaction credibility through AI in finance

One of the leaders in applying AI to finance is Stripe. It has developed a system called Stripe Radar, which analyzes more than 1,000 features of a transaction in less than 100 milliseconds to assess its reliability. The system has a 99.9% accuracy rate while maintaining a low false alarm rate.

How was this achieved? First, Stripe uses advanced machine learning techniques like deep neural networks. The system is constantly being improved and developed with new capabilities, such as transfer learning.

Second, the company is constantly looking for new signals in transaction data that can help identify anomalies that indicate potential fraud. Stripe’s engineers carefully review each fraud case to understand the criminals’ operating patterns and enrich the system with additional rules.

Stripe Radar is an excellent example of how AI in banking can effectively protect customers from financial scams.

Source: Stripe (https://stripe.com/blog/how-we-built-it-stripe-radar)

Monzo: AI in finance

Monzo, a U.K.-based neobank that operates exclusively in the digital space, has applied machine learning capabilities in a completely different area: optimizing marketing campaigns.

The bank has built models that, based on historical data, can estimate the willingness of a given customer to take advantage of an additional offer, such as opening a savings account, if they receive a specific message from the bank.

Next, to maximize the efficiency of the campaign, the system indicates which customers should receive which promotional message. This allows for precisely targeting the message and achieving significantly better results than in the case of mass, non-personalized communication.

In some cases, the implementation of such optimization has allowed Monzo to increase the effectiveness of campaigns by up to 200%! This demonstrates how AI in banking can help reach customers more efficiently with tailored offers that resonate with them.

Source: Monzo (https://medium.com/data-monzo/optimising-marketing-messages-for-monzo-users-3fe805f24572)

Grab: AI in the classification of sensitive data

Grab is a technological giant from Southeast Asia, offering services such as transportation and delivery. The company has decided to leverage the capabilities of Language Models (LLM) to automate the classification process of sensitive data it stores. This is crucial because the company holds the personal and financial data of its customers.

For this purpose, a set of tags has been prepared describing various categories of data, such as:

• Personal data,
• Contact information,
• Identification numbers.

Next, appropriate queries were designed for the language model to automatically assign these tags based on table and column names in the databases.

As a result, Grab can classify stored information by sensitivity much more quickly and cheaply. This makes it easier to enforce data access and privacy policies. According to the company’s estimates, the solution has saved as many as 360 working days per year that were previously spent on manual data classification.

Source: DALL·E 3, prompt: Marta M. Kania (https://www.linkedin.com/in/martamatyldakania/)

Summary. The future of AI in banking and finance

As the examples of Stripe, Monzo, and Grab show, artificial intelligence is already delivering real business value to banks and financial institutions. It can help prevent fraud more effectively, target customers more precisely, or automate tedious tasks.

In the coming years, the role of AI in banking will continue to grow steadily. We can expect the full automation of many back-office processes, hyper-personalization of financial products, and a closer integration of machine learning models with banking systems.

If you like our content, join our busy bees community on Facebook, Twitter, LinkedIn, Instagram, YouTube, Pinterest, TikTok.