Artificial Intelligence in Finance

I recently got admitted to the Simon Business School, University of Rochester for an MS in Business Analytics. The course revolves around using data in a way that can be used to solve complex problems of businesses, deploy models that are capable of doing what humans actually do. Yes, basically using AI in the world of healthcare, finance, marketing, sports and much more.

This made me think that it is the right time to write about what I have learnt so far through various courses in R and python and share it with readers, particularly how data is used to solve the problems of finance.

Before diving into specifics, let us know what Artificial intelligence is, not a dictionary definition, but a definition of my own.

Think about your own self doing a task, like playing a video game, or learning to swim, etc.

Let's go with the video game example. Everyone has played a video game once in their life, may it be Temple Run, Grand Theft Auto, or Among us. At every level, there are new barriers that come up that you have never seen before; increasing the level of difficulties. You fail once, try again, fail again, try the third time and manage to pass that level.

How do you do that?

Your mind collects data on the obstacles that you faced in the game, the ways which resulted in you failing the level and tricks that you could have used to save yourself next time. The more tries you give the more data your mind collects and the more likely you are to successfully complete the level in the next attempt.

In AI programmers do the same, but with machines. A model is like an empty brain at first, which is then fed a huge amount of data on a real-time basis, the model tries to identify the patterns in the data set like we identify the patterns in the obstacles on each level of a game. It then finds a solution to a given problem and predicts a given outcome. Models also fail as we do, and it learns from its mistakes and tries to give an even accurate outcome the next time.

This was a very simple explanation of how AI works everywhere, right from autonomous cars to targeted marketing techniques.

AI is broadly divided into the categories given:

Machine Learning: Identifying patterns in the data and making predictions using different models.

Computer Vision: Detecting, identifying and categorizing images and objects in the images. (What Tesla uses)

Natural Language Processing: Identifying the text or speech and its meaning and completing tasks and solving problems based on that. (Used by Siri, Alexa and OK Google)

Okay, enough about AI, We are Finquity and we want our readers to know how it is used in the finance industry. AI is definitely strengthening its grip in the finance industry giving birth to a new sub-industry known as the Financial Technology industries or fintech in short.

Let's discuss various avenues where it is currently being used along with some real-life examples:

1. Risk Management and Portfolio Optimisation: Machine Learning techniques are being used to assess the risks of the clients along with assessing the risks of the investments the clients wish to make. Various combination of portfolios is then recommended to the clients using historical data on various types of investment. All of this with minimal human intervention. This is actively used by researchers at Narnolia Financial Advisors Limited in Mumbai. From the perspective of a corporate, AI tools are used to assess the risk structure of the projects that they plan to invest in. For eg, predicting the revenues from a launch of a new product. Credit rating agencies like Credit Sussie, etc use models to assess the risks associated with various investments and rate them accordingly.
   

2. Loan Default Protection: This is one of the most common uses of Machine Learning in finance. I myself have worked on data sets solving problems of loan default. Data on lakhs of credit card users and borrowers are collected which includes pointers like, their credit score, frequency of delayed payments, delinquencies in the past, type of occupation and a lot more. The model then tries to predict whether a person applying for a loan or a credit card will be able to repay their dues. Although today the decision-making process is a mixture of both human and artificial intelligence, the day doesn't seem far when machines will decide whom to lend money to. Huge financial corporations like JP Morgan Chase, American Express, etc use these models.

3. Quantitative Trading: It is a non-traditional form of investment that is given birth by the introduction of AI. i. Quant Funds: Unlike mutual funds, the securities in these funds are bought and sold on the basis of the quantitative analysis conducted by various machine learning models. Most of the transactions are conducted automatically. They have turned out to be a very effective and successful means of investment giving promising returns. Some of the examples of quant fund companies are AQR Capital, Renaissance Technologies, etc.

ii. Algorithmic Trading: Rules of trading are set in software that executes trades accordingly and ensures a profit. Humans only have the task to set the rules and nothing else. The securities will be bought and sold on their own by the machines. Many traders and investment firms in India have started using Algo trading actively.

4. Personalised Banking: Large banks have come with their own financial assistants to assist their customers and provide on-demand services. These assistants and models have also been able to provide personalised offers and services to the clients and account holders on phones and computers replacing the need to go to the branch and get the jobs done. EVA assistant by HDFC bank is one of the examples.

The list doesn't have to end here. There are infinite possibilities of how AI can be used in fintech, some of which are yet to be known by us. There are models which can advise you on your personal finances help you spend and invest wisely. The list is endless!

This brings us to a very important question

Will AI take over jobs in the finance industries as well?

The answer is yes and no both. It will definitely replace the jobs that are more manual, like doing the paperwork, buying and selling securities, etc. However, logic intensive jobs like research and analysis of investments, writing complex codes, setting the rules for trading etc will still remain and in fact, grow.

That was my bit on artificial intelligence in the field of finance, feel free to write to us on email and Instagram!

For more such blogs subscribe to us.

#GoSlowBeConsistent