CXO Bytes

AI and Machine Learning in Finance: Enhancing Decision Making and Risk Management

Last few weeks have been eventful for AI as many of the researchers around the world came together and signed an open letter to pause gaint AI experiments claiming that it can pose profound risk to humanity. While some think that these discussions are farfetched, it is clear that AI and machine learning in general present interesting opportunities in the field of computing to manage complex challenges.


Rich data is the fuel for machine learning models. Finance industry in general is data heavy and the number of data sources are only increasing by the day apart from tsunami of ideas and opinions from social media networks. It is important to digest this data to provide timely insights for the decision makers. Industry is also highly regulated, compliance heavy and processes are very well defined and these aspects present a fertile ground for intelligent models to operate.


While many software systems are already in use, managers still struggle as they do not get enough assistance from the applications. The need for AI that works like smart assistants that has access to rich pre-processed data and which can understand the needs of the decision makers and can come up with timely suggestions are the need of the hour. We need systems that can come up with smart proposals and we should be able to cross question the system to check the rationale behind these proposals in the same way we do with the executive assistants. This will help the models to learn and will also help managers to validate the proposals presented by the systems. This will facilitate two way learning where humans can learn from the systems and models can learn based on human feedback.


We live in a highly inter connected world where shocks from one geography travel to another in no time. Unlike earlier times, data from the enterprise is also analysed by several agencies both formal and informal influencers and these are broadcasted in real time. We need robust models to monitor the risks from different directions and take appropriate actions.


When it comes to managing the risk of investors, the present-day systems are designed to assess the risk of the investors during the start of the investment journey and the risks are evaluated in periodic intervals. As the environment is very dynamic, there is a need to analyse the portfolio and the investors situation continuously to take maximum advantage of the opportunities and to take timely decisions. Machine learning models can greatly help here as it can analyse not only from our experience but also from the experience of similar scenarios and can guide the investors at every step. Here again, the models can assess the risk on an ongoing basis based on performance, market dynamics and user preferences. Models as well can prompt users at critical junctures and ensure that costly mistakes are avoided. Systems as well can come up with intelligent proposals to manage and mitigate risks in more innovative ways.


While chat bots are effective in interacting with users in natural language with multi turn conversations, they are used more in the direction of answering questions. We need systems that can ask intelligent and leading questions which will help users to learn. It is possible for users get trapped in their thinking and models can act as an eye opener by asking and answering questions. AI assistants can win the trust of their masters and can help them to manage mundane tasks, avoid common mistakes during execution and point to interesting opportunities and deliver extraordinary outcomes.


While some say, AI can kill of humanity, the reality is, we still have to answer every time in an ATM if our account is savings or current. While the tech companies have always been ahead of the society in painting the vision, we are in a unique moment where the general public has several ideas and we see tech companies lagging in execution and even lobbying to pause the research.


We are witnessing deluge of data and we spend significant time with our devices. Even if AI helps us manage a fraction of this complexity, it will produce remarkable results. It is time to look for incremental improvements before dreaming about sweeping changes that can fix problems of the humanity.



(This article is written by Jayaram Srinivasan, CTO,, and the views expressed in this article are his own)

Leave a Response