Image credit:iStockphoto/BeeBright

Redefining banking through AI and big data

The rise of the Millennials is creating a revolution in the economy. Banking isn’t exempt from the disruption caused by new technologies. This generation is open to solutions  from outside the traditional financial system. Tech giants and start-ups get the credit of being able to create the necessary change. Artificial intelligence (AI) and data science will most likely be the fuel of the new approach to finance. Already present in retail, HR, and marketing, AI is finding its way in the world of banking.

Defining AI for banking

But how exactly can AI be used for banking? The answer comes down to understanding that AI is an umbrella name for multiple technologies built on big data and neural networks. The most useful in the financial sector will be natural language processing for answering customers’ questions, machine learning for processing back-office operations, replacing humans especially in tedious, repetitive tasks and expert systems with predictive power, able to trade stocks automatically.

AI uses for banking

Customer care

Although it is counter-intuitive, introducing more technology will make banking services more personal and pleasant. Chatbots have evolved enough to create a human-like interaction with clients and are never tired. They can even feel the customer’s moods and react accordingly. Training the robot increases its performance in learning and makes it ready to take on more challenging tasks. One bank already experimenting with such a system is Swedbank, which has employed Nina. The natural language processing system is handling over 30,000 conversations per month, satisfying over 75% of the bank’s clients, who prefer to deal with transactions in the app or online. Nina is already helping cut call center costs.

Recommendation engines

Inspired by Amazon and Netflix, financial institutions aim to create customer-oriented systems that can trace behavior and make appropriate recommendations. These will include notifications regarding spending habits, ideas to improve credit scores or cross-selling the bank’s products. Bank of America is pioneering this by investing in Erica, a chatbot with predictive capabilities, looking to become the personal advisor in your pocket.  Of course, Amazon could not just watch the game. Alexa, the digital assistant, can retrieve balance information, buy stocks, compare insurance ratings and more.

Fraud detection

Security is one of the main concerns of online banking, and automated systems are prone to hacking unless properly secured. Detecting fraud is a combination of artificial intelligence and human expertise. IBM’s Watson system is under training by audit experts. Fighting against multi-channel crime is another track that is upgradeable, counteracting identity thefts and eavesdropping through customer services.

The innovation consists of replacing statistical models with cognitive, predictive models, to fight crime in the early stages or even before it happens, by tracking account activity. Lloyds Banking Group is currently experimenting with Pindrop, a fraud detection mechanism focusing on voice analysis by phone printing. This protects customers from fraud by authenticating calls.

Process improvement

AI is not here to drive up unemployment, but to free human resources from daunting mindless tasks and give them interesting, thoughtful targets. People trapped in repetitive jobs are facing high error levels. AI can help improve accuracy, speed up processes and let people solve problems, while the machine takes care of mundane work. One example is JP Morgan’s COIN project for correct interpretation of legal contracts. The system can analyze over 12,000 contracts per year, saving 360,000 working hours, an impressive achievement from an efficiency perspective.

Challenges of AI in banking

Although the future looks promising for AI in banking, this is a highly-regulated sector, and the challenges and obstacles are numerous. Most companies don’t have a clear understanding of the data they own and how to access it in a useful way. There is still a lack of qualified personnel to handle AI implementation. This affects data analysts, as well as project ownership.

Conclusions

Although there is a lot of exciting innovation underway , banks have a long way to go in adopting state of the art technology. Some noticeable progress is already visible, especially regarding customer interaction via chatbots and process optimization. There is still room for improvement regarding predictive modeling, fraud detection and prevention, as well as automated financial advice.

The progress of big data powered algorithms in banking will be slower compared to other areas such as retail and marketing, due to increased regulations. Clients are expecting the same degree of customization from these services as they are getting from other companies. Instant gratification and 24/7 service are slowly becoming the norm in banking too, pressuring financial institutions to keep up and invest in AI.

Image credit:iStockphoto/BeeBright

Written by Emilia Marius

Emilia Marius is a senior business analyst with 8+ years of experience. She focuses on IT solutions for banking, retail, and ecommerce and has applied her skills to such projects as a sales analysis system for a retail company, a mobile payment solution for an e-shop, an online analytical system for a bank and more.