Serengeti logo BLACK white bg w slogan
Menu
Serengeti logo GREEN w slogan
Menu
Get a Quote

AI in Finance – Use Cases and Challenges

Serengeti
10.12.2021.

The machine learning market is constantly increasing in size. The most noticeable part of the market is deep learning software, which, according to Statista, will reach almost $1 billion by 2025. An exponential increase in data collection and the need for their fast and accurate processing for forecasting purposes further contributes to this growth.

We can all agree that technology has completely redefined the way we work in various industries. BFSI is one of the industries where AI/ML is causing significant upheavals in 2021 and beyond.

Transforming the Banking Industry

Artificial intelligence and machine learning are altering the way banks operate and perform their functions. They are significantly contributing to a comprehensive and profitable experience for both the bank and the customer. 

Unlike some industries that began with machine learning implementation during the pandemic, financial institutions were early users of artificial intelligence (AI), such as machine learning and other advanced data science methods, before COVID-19.  As a result, the financial industry was one of the largest spenders on AI services in 2018.

Despite the pandemic, interest in AI and ML implementation has been resilient. According to a survey conducted by the Bank of England, around 40% of respondents said that the relevance of machine learning for future operations has increased, with 10% of institutions on a significant rise. The relevance of machine learning has not decreased at any of the banks. The financial sector is heavily utilizing artificial intelligence and machine learning to automate and simplify processes.

why banks use ai
Source: Bank of England

Just in Fintech, AI usage achieved a value of USD 6.67 billion in 2019. It is anticipated that it will reach USD 22.6 billion in only five years. The fintech industry is expected to develop at a CAGR of 23.37 percent until 2025, with several factors accelerating expansion.

Machine Learning in Finance – Use Cases

Machine learning algorithms can improve decision-making, provide custom services to clients, improve fraud and risk management, provide digital assistance, and give better insight into user behaviour. 

reasons banks use ai
Source: Aideo

According to a McKinsey survey, the most commonly used AI technologies are:

- Robotic process automation (36%),

- Virtual assistants (32%),

- Algorithms for fraud detection (25%).

While many financial services companies utilize machine learning for specialized use cases, a growing number of companies are adopting a more holistic approach. Their goal is to implement and integrate AI across the whole lifecycle, from front to back office.

mckinsey ai chart

Let’s look at some popular use cases for machine learning implementation in finance.

Customer service – Today, most clients seldom visit their bank locations. Instead, they use their phones, mobile and internet banking. Nevertheless, good customer experience is still an integral part of bankingLarge banks, including JPMorgan ChaseBank of America, and CitiBank, are investing a lot to enhance customer service while generating income.

We can all agree that in this industry, security is a customer's top priority. But they also value good customer service and unique customer experience.

An example of personalization using the ML algorithm is Eno, the virtual assistant from Capital One. Eno monitors user accounts 24 hours a day, seven days a week, and notifies the user if anything suspicious is discovered. Anyone who has been in a situation where they must return money to an account knows how long and complicated the process can be, so this assistant can be quite helpful in evaluating valid transactions.

Chatbots are another example of successful AI implementation. According to Juniper Research, through chatbot interactions, approximately 826 hours will be saved by banks. The main benefit of chatbots is the possibility to handle a lot of customer requests without human interaction. With chatbots, human call handlers will receive only more complicated inquiries. 

Fraud detection – As already mentioned, security is a top priority for clients and businesses in this industry. Advanced fraud detection and prevention is one of the significant benefits of AI and machine learning. 

Machine learning algorithms learn from natural behavioural patterns. They can adapt quickly to variations in typical behaviour and can recognize patterns of fraudulent transactions. This means that machine learning models can detect suspicious clients before the charge has been made. 

Risk Assessment – By using machine learning algorithms, banks can create accurate spending and revenue estimates by analysing a user's financial history, current transactions, and purchase trends. Automated risk assessment enables banks to automatically give users the optimal loan and credit product conditions based on their risk.

Cost benefits – According to research company Autonomous Next, banks throughout the world should decrease expenses by 22% utilizing artificial intelligence by 2030. Machine learning software is likely to be employed in the banking industry to combat credit card or identity theft. Furthermore, AI lowers labour costs by optimizing capital investment decisions and reducing forecasting risks. Banks, for example, may use predictive analytics to determine whether to accept a loan application based on past data.

Computer Vision in the Banking Sector

Computer vision has a lot of potential in banking. Most of the talk revolves around security and fraud detection.

Content-based image retrieval, code reading, position estimation, character recognition, facial recognition, and shape identification technology all rely on computer vision. All of these technologies have a potential use in this sector, and here are a couple of examples of recent developments.

CaxiaBank has taken it a step further by allowing clients to withdraw money from ATMs using face recognition. The ATM can validate up to 16,000 points on a face picture to completely authenticate the identity of the person making the withdrawal.

Another example is BBVA, where clients may submit a picture of their ID and a selfie to the bank. The system then takes the data and uses computer vision to ensure that the photos are identical. Computer vision replaces manual tasks while increasing accuracy, performance, and efficiency inside the organization.

In April 2021, Russian Sberbank has announced their facial recognition service. This service, when activated, will allow the customer to pay for groceries with just a glance.

Computer vision can help us move closer to the ideal of a paperless and cashless future, in which every transaction is completed digitally via mobile phones and smartwatches. A simple task, such as opening a new account, can be shortened from days to a few hours with computer vision.

Apart from banking, the application of this technology is also possible in insurance, especially if we are talking about insurance for the automotive market. For example, using computer vision with a dashboard camera can facilitate a more detailed assessment of the situation and damage in a driving incident. Thanks to the impartiality and accuracy of the data they would receive from such systems, insurance companies can reduce the possibility of fraud when paying out insurance.

Challenges

Implementing computer vision does not come without challenges. Although today the adoption of new technologies goes faster than before, there is still some resistance in implementing banking beyond the traditional framework.

Some users are still sceptical about the concept of banking, which is not based on physical branches and much paperwork. Even though the level of security can be increased using computer vision, users are often distrustful of anything new when it comes to financial transactions.

A simple example of how computer vision can improve security is the use of biometric data. By scanning the cornea of ​​the user, their identity is established, and the transaction is authorized. As the cornea is extremely difficult to replicate, this authentication approach raises the security of transactions to a higher level. At the same time, the privacy of personal account related data also improves.

Conclusion

Artificial intelligence is helping banks and other financial institutions drive the financial future for their customers through automated call centres, fraud detection, and many other services. Companies that invest in new technology and transform their business with AI stand to gain market share and enhance customer experience and financial performance. 

In the financial industry, reducing the possibility of fraud means creating a huge opportunity to gain customers' trust. Maybe that's why AI adoption is becoming more like a strategic imperative than a business strategy choice. 

If you are interested in AI solutions, feel free to reach out. We will gladly show you all aspects of this technology and help you implement it in your business.

If you’d like to read more about machine learning, download the whitepaper below.

Let's do business

The project was co-financed by the European Union from the European Regional Development Fund. The content of the site is the sole responsibility of Serengeti ltd.
cross