MyFi: AI Investment Solutions for Financial Literacy in India

In today's tech world, MyFi is dedicated to demystifying investment planning, making it straightforward and convenient for all. Leveraging 256-bit encryption ensures top-notch security for our users' data. As a SEBI-registered Investment Adviser, we offer dependable and personalized guidance, helping you make informed financial decisions to achieve your long-term wealth goals.

Kiran is the Co-founder & CEO of MyFi. He joined TIFIN in February 2022, following the acquisition of 1st Main, as CTO in India to accelerate product development.

Kiran holds a degree in Computer Engineering from Imperial College London. He spent 10 years in Banking Technology at Bank of America Merrill Lynch and Morgan Stanley, eventually leading the Equity Structured Products technology team as Executive Director. In 2017, he left Morgan Stanley to start 1st Main and pursue his entrepreneurial ambitions.

In an exclusive interview with Ciol, Kiran Nambiar, Co-founder & CEO of MyFi, discussed how MyFi leverages AI to provide accurate analysis and personalized recommendations to users. He also highlighted the measures in place to ensure the accuracy and reliability of MyFi’s investment recommendations.

Could you share the journey and inspiration behind the creation of MyFi, particularly focusing on the specific gaps in the Indian financial market that led to its development?

I've personally been interested in the wealth tech space for over a decade now. Coming from an engineering background, I spent about 10 years in banking tech. I first worked in the UK, then in the US. Even while I was in the US, I was keen to do something in the Indian market. The research and groundwork for this started very early on. I remember having conversations around 2012-2013, talking to people about mutual funds. While they understood it was some sort of investment, they didn't know what it was. They didn't understand that there are stocks or debt instruments underneath mutual funds, among other things. There was a massive lack of financial literacy. A survey from around that time indicated that more than 75% of Indian adults lacked basic financial literacy. All these factors combined made me feel there was an opportunity here.

I came back to India in 2017 and started a digital product agency. We design and build websites and mobile apps, and a lot of it is in the fintech space. This is how I started working with TIFIN. We ended up doing so much work for their US products that they acquired us. Over the last two years, in 2022 and 2023, and even parts of 2021, most of my time was spent on one of their US products called Magnifi. Magnifi also used conversational paradigm which has been adopted across TIFIN products like MyFi.

If you look at most TIFIN products today, they are Gen AI-based. We're seeing an opportunity where this conversational experience can change the way people interact with information. In fact,I recently received a message from an equity research team in India saying they wanted to connect all their databases to ChatGPT and ask it questions.

So, it's this combination of the gap in financial literacy, the growth in the Indian market for mutual funds and capital markets in general, and the advancements in AI and India's digital public infrastructure that have brought MyFi to life. Today, we're at around 61 lakh crores in AUM of Mutual Funds, and that continues to grow. Estimates suggest we'll hit 100 lakh crores very soon. SIPs are at an all-time high, around 20,000 crores a month, and only increasing. This combination of factors has brought MyFi to life.

How does MyFi leverage AI to provide accurate analysis and personalized recommendations to users?

So, I think it's important to clarify how we use AI, right? We use generative AI to enable the conversational capabilities in the app. When you say "Hi," "What can you do," or "How's my portfolio doing," we understand what the user is saying and respond with an answer based on generative AI. The actual models that make the recommendations—whether it's a portfolio recommendation, portfolio score, or search—are all driven by the deterministic models that we have.

This approach allows us to keep it conversational while also preventing the risk of hallucination, which is one of the risks we have with large language models (LLMs). We avoid this by controlling the models that respond with recommendations. This is a key differentiator and something we believe is very important to safeguard the recommendations we provide to users.

Could you please explain the role of generative AI in enhancing the user experience?

So, what we do is this: when a question comes in, we use generative AI (GenAI) to understand what they're saying. We then go to our models based on the question. For example, let's say it's a search for finding mid-cap funds. We perform a search based on filters that we apply, like mid-cap returns, etc. We send the response back, and the generative AI is again used to summarize that response for the user, along with the individual widgets where we provide the actual data.

Generative AI is used for input understanding and output summarization, but the actual recommendations are specifically done by our deterministic models.

What measures are in place to ensure the accuracy and reliability of MyFi’s investment recommendations?

The recommendations part is not done using generative AI. What happens is we receive daily feeds from mutual fund houses that provide us with NAVs and other information about mutual funds, as well as any new funds that are coming in. That data is then fed into our models daily.

We re-score each individual's portfolio every day. Any new user who connects is immediately recalculated as well. But all of this is done using deterministic models, not generative AI. So, the way we safeguard our recommendations is by not using generative AI to make the recommendations but to present the information. The calculations are done using our models.

Can you discuss the integration of MyFi with SBI’s regulations and RBI’s Account Aggregator framework?

MyFi, right from the start, aimed to be aligned with the customer rather than with the fund houses in terms of our business model. We are a subscription-based platform and a SEBI-registered RIA. We don't make any money from selling specific mutual funds or receive commission distribution fees or anything like that. We are very focused on being aligned with the investor.

Currently, the mutual fund data is coming directly from MF Central. We are in the process of integrating with the account aggregator as well. There are some gaps in data from the account aggregator when it comes to brokerage accounts and similar data, which we will work through. But the account aggregator is the future. It allows us to consolidate not just investment accounts but also bank accounts, FDs, etc., which truly gives a complete picture of a user's investment portfolio and potential investment capability.

We plan to double down on this over the coming weeks to increase the spectrum of products that we take from the user.

How does this integration ensure data security and privacy for users?

That's a completely valid question. Right from the start, privacy has been at the center of how we design these systems. We ensure that all personally identifiable data, whether it's your date of birth, PAN, or phone number, is encrypted and stored in our database. So even if someone looked at our database, the data they see is scrambled. It's all encrypted using 256-bit encryption with asymmetric keys, which is the level of encryption used for transaction data.

What are some of the key milestones and achievements MyFi has reached since its launch?

So it's really been only three weeks since launch. We've been focusing a lot on organic growth and PR, not performance marketing, because we believe that is the way forward in building a sustainable business. Even with that approach, we've got over 13,000 downloads, and we have 300 crores worth of assets linked on our platform. The growth has been consistent, and the responses so far have been very positive. So, those are some key numbers that we've already hit, and we continue to expand.

What are the upcoming features or enhancements that users can look forward to in MyFi’s product roadmap?

I think one part that a lot of people have requested and that is imminent is introducing stocks on our platform. We wanted to focus on mutual funds first because it was a low-hanging fruit and aligns with our long-term wealth creation view. But stocks are a significant feature that will be coming in the next four weeks or so. We're also looking at expanding into other planning modules and similar features.

Driving the product features based on user demands and ensuring data quality and accuracy is extremely important. That has been a major focus for us as well.
Stocks, execution, planning modules—the list is quite endless. But every time we introduce something, it's about adding more value to the user. That has been our North Star and will continue to be.

How do you plan to expand MyFi’s reach and user base in India, and what specific strategies or partnerships are in the pipeline to achieve this?

We're avoiding performance marketing right now. We are exploring a number of different approaches. We've done a few physical events, including a couple at WeWorks, and we're looking to do more because it gives us an opportunity to interact with users. We want the product growth to be user-led rather than us pushing marketing dollars behind it. That's really our focus. We're also looking at promoting this through companies, employees, and similar events.

So, as much as possible, our focus is on product quality. The users who are on board should be talking about it, driving a very product-led marketing approach.

How does MyFi ensure a judgement-free and unbiased investment experience for users, particularly through its AI-driven approach?

The fact that it is a model answering the user, and these are the models that we have built, allows us to avoid being biased by which fund house might be giving more commissions and things like that. So even our business model, as I was saying, is very aligned with the customer.

Having a subscription based model instead of a distribution based model allows us to be completely aligned with the customer. Our models make recommendations based on data we receive from the market and have been developed by our in-house investments team. This removes and bias of affiliation with specific fund houses.

If the customer does well, we do well. That alignment removes any bias we might have in recommending one fund over another.

What steps are taken to continuously improve MyFi’s AI algorithms to keep them up-to-date with market trends and evolving user needs?

So that is something our investment team focuses on, even with the data we have regarding mutual funds. For example, there was an article that came out saying our AUM in mutual funds has crossed 61 lakh crores. We actually had that data ahead of time. We are constantly looking at AUM data coming from our data providers and monitoring new fund launches. We're not just looking at the funds and categories but also the underlying assets in those funds.

When you look at small-cap funds and tech funds, the actual allocation behind the scenes is something we continuously review, monitor, and make recommendations based on. We also improve our algorithms accordingly. It is a combination of human intervention, data analysis, and model adjustments to account for what we're seeing. This human-plus-tech element allows us to stay ahead of the curve. Having dedicated people looking at this is something that hopefully gives us an edge when it comes to investment data.

Read More Interviews Here: