Listen to this article
Of late there has been much talk about AI (Artificial Intelligence) and ML (Machine Learning) and many are keen to explore and integrate them into business. But does this make sense or it is merely FOMO (Fear of missing out)?
Also, what are the uses of AI/ML in the mutual fund business?
To get answers to these pertinent questions, Cafemutual hosted Mridul Mishra, Vice President, Fidelity Investments at Cafemutual Confluence 2022 (CC 22). Mridul simplified AI/ML and explained how it supports business operations.
Are AI and ML the same?
AI - In the words of John McCarthy, AI is the science and engineering of making intelligent machines. And, we believe human intelligence to be at the pinnacle of it.
ML - An approach to algorithmically learn patterns in data and use these patterns to predict. To note, it assumes there are patterns in data. If data does not have patterns it doesn’t work.
AI Vs ML - AI and ML are different. In fact, ML is one of the ways of getting AI, which you can also get in other ways.
Where is AI used in the MF industry?
There are plenty of use cases across the life cycle where AI has delivered value.
Taking the most popular use case across categories, Mridul explained:
Personalization - With personalization in marketing, you can not only say if it is better to connect with a person on WhatsApp rather than by mail but you can also tell what time will be right and what will be the best tone to communicate.
Frictionless onboarding - With computer vision, you can process the identification documents and onboard clients within minutes.
Trade execution - AI helps to identify how to trade itself, how to remove the slippage cost and also how to reduce the impact of going into the market
Intelligent call routing - If clients are on the website or mobile app and they end up calling, using AI you will be able to figure out what is the intent of that call. For instance, if a client was on the ‘beneficiary update’ page and calls, most probably the call will be on beneficiaries.
Improved compliance functions - A lot of legal documents have been automated in terms of extracting information.
There are also some common capabilities which have been built on extracting information from documents, doing sentimental analysis of conversations including tweets, etc.
Is AI/ML right for you?
Mridul suggested reflecting on a few questions before deciding to integrate AI/ML into business.
What’s your organizational objective?
FOMO is a bad idea. Considering AI/ML only because your competitors are doing it or because you have read about it over the internet may not go well. Always be clear about why you want it and what the objectives are behind it.
Have you tried and seen benefits with analystics? - Before trying AI/ML try basic analytics and do exploratory statistical inference to see if the data suggests something.
Do you have enough data?
AI/ML is based on data and will not work if you don’t have data.
Do you need explainability and auditability of your decisions?
If you need an understanding as to why AI/ML made certain recommendations, please note that a lot of explainability is not available yet.
Here’s a teaser of Mriduls’s session. You can watch the entire video along with other CC 22 videos at a nominal price of Rs. 499 only. Simply write to us at fouzia@cafemutual.com and get ready to march towards Rs. 100 lakh crore.
