It was in the 1950s when Alan Turing posed the question, “Can machines think?” and since then artificial intelligence has been there. Amazon Alexa and Siri have become household names. However, it was the finance industry that caught up with it quickly. According to Citigroup, the financial services industry is the biggest spender on AI outside technology. With the evolution of big data, cloud computing, innovative hardware, and faster special-purpose systems, this growth has been faster in the last few years. Today, AI applications are widespread and are actively used in Algo trading, fraud detection, chat-bots, market and regulatory compliance, market impact analysis, and stress testing.

Detecting Financial Frauds

An increase in e-commerce has not only increased convenience but also gave rise to online frauds. According to a KPMG article, some banks spend around 500 million dollars every year to fight Anti Money Laundering (AML) and in KYC practices. US banks spend around $70 billion on compliance every year. In India too, 53% of the total registered companies have been victims of online financial fraud. This is a serious situation as customers will get wary of using credit cards and other devices. To protect the firm from such malpractices and maintain customer loyalty, it was important for organizations to strengthen their operations. AI came as a game changer here as ML (Machine Learning) algorithms analysed billions of data points simultaneously to keep fraudulent behaviour at bay. This is just impossible for humans.

Financial crimes don’t come with a pre-known set of rules and regulations. One financial fraud is unique from the other. With ML came the power of self-learning and calibration so that financial risks are met successfully. Using this same self-calibration power by AI, banks today load their historical data into monitoring systems to classify events as fraud-based or non-fraud-based. This involves algorithm training, back training, and validation sequences.

Organizations nowadays are also going for AI-based fraud prevention alternatives for a stronger experience. A good example here is DI ( Decision Intelligence) tool by Mastercard. By using this, rather than classifying a transaction on the basis of historical data (as it gets limited to a set of conditions), a baseline is created to compare new customers. Now, against this baseline, every new transaction is scored in a real-time environment. This helps Mastercard in two ways. Firstly, the fraud rates get reduced and secondly the ‘false positives’ that arise out of strict algorithms are in check. False positives refer to transactions that are declined if they don’t conform to the algorithms but are actually clean. As per the Javelin Strategy Report 2015, transactions wrongly declined for fraud account for $118 billion in retail losses.

AI and ML have also made such an impact on the stock market that all leading traders vouch for them. “ Artificial intelligence is to trading what fire was to cavemen”, said one industry player. All thanks to AI-powered sentiment analysis. With this, the system learns new trends in the financial sector on its own and enables traders in making their intuitive decisions a lot smarter and more accurate.

Banking Chatbots and Roboadvisory Service

Chatbots and roboadvisory services were unheard of 10-15 years back. It was after the 2008 depression when the power of customer engagement was realised and these technologies became popular. A ‘chatbot’ is a computer program designed to simulate conversations with human users over the internet using AI technology. Roboadvisory services, on the other hand, don’t involve robots at all. These are algorithm-based programs (user customised) that help users in making financial and investment decisions.

Chatbots are synonymous with customer engagement for banks and financial institutions today. Erica, a chatbot by Bank of America, helps to send notifications to customers, getting credit reports, advising users on money-saving plans, paying bills and carrying other simple transactions. Plum too is a widely used finance bot that helps users save money. Here, after linking one’s account number with the mobile app, the AI engine of Plum analyses the spending habits of the user to predict how much he can save. Similarly, JP Morgan Chase uses the bot COIN, which helps the firm analyse legal documents more closely.

Roboadvisory services have brought transparency and quality to the wealth advisory and management business at lower costs. An interesting case here is of the debt collectors in China. The P2P debt in China stood at 200 billion $ in May 2018. There was no way that the debt collectors would be able to recover the amount. To improve the situation, a roboadvisory dialogue service named Ziyitong was developed. This AI-based debt collection system used borrower’s and their friend's information to create language phrases that help in getting the money back. Ziyitong will then call such debtors and repeat those phrases. According to the firm involved in its development, the recovery rate with Ziyitong is 41%. Not only this, roboadvisory is actively revolutionising the way stock trading, investment making, etc. was done.

Algorithmic Trading (AT)

Algorithmic trading or automatic trading refers to feeding trading rules into a program and then using it for trade. This involves minimum human intervention. AT when combined with AI and ML, creates a structure that predicts the results much faster (also known as HFT) and with more accuracy. Such trading has the ability to evaluate the best prices while taking multiple market sentiments in mind. According to Aldridge and Krawciw(2017), the market share of AT today is 40% and is used extensively in trading houses, hedge funds, corporate, bank trading options, etc. One well-known AI trading tool is Katana, which shows human traders different ways to gather bond prices more rapidly. As per the press reports, Katana has led to faster decisions 90% of the time when compared to traditional methods.

Without doubt, AI is the bandwagon that can make the finance industry a lot smarter, safer and more efficient. With billions of investments being made by firms in this field, it won’t be long before we witness a completely AI-enabled finance sector.

By: Ashish Malhotra
Date: 2nd August, 2021