In a world where people can order biryani on Swiggy or groceries on Zepto with a single tap, it was perhaps only a matter of time before someone tried to do the same for wealth management. InvestorAi, an AI-powered equity investment platform founded in 2018 by Akshaya Bhargava, Bruce Keith, and Sarthak Behura, wants to strip away the jargon, middlemen, and complexity of stock picking, replacing it with one-click investing.
Backed by the Securities Exchange Board of India’s (SEBI) approval and investors like Ashish Kacholia, the company has positioned itself as a rare player that built its own AI foundation model from scratch, moving away from existing large language models (LLMs). The question now is whether InvestorAi can truly democratise investing without diluting the responsibility and rigour that financial decisions demand.
The Democratisation Ambition
InvestorAi was built with a clear focus: making wealth management accessible to a much larger pool of investors.
Bruce Keith, cofounder and CEO of InvestorAi, vividly recalls the early motivation. “What we wanted to do was use AI to help democratise wealth and asset management. We thought: if a typical wealth advisor can look after 50 clients, AI could make them efficient enough to handle 100.”
That early model of AI tools for wealth advisors was gaining traction across the US, UK, Europe, and India. But the pandemic (Covid 19) disrupted the business. As wealth advisors froze decisions, the team realised they needed to flip their approach. Instead of making the supply side efficient, they would go directly to investors.
In April 2021, InvestorAi launched its first equity basket on Smallcase. The results were striking for the startup. “That basket has compounded annually at 45% since launch, compared to the market at 16-17%. It just knocked it out of the park,” Keith says.
A Foundational AI Model, Not a Wrapper
Unlike many fintech firms that lean heavily on existing LLMs, InvestorAi has taken the road less travelled: designing an in-house foundational AI. “We don’t use any LLM tech in our principal model,” says Keith, explaining that they didn’t need ‘language’ but data for identifying stocks.
“On a daily basis, we run about 4 billion data points for each model portfolio, across 20-30 portfolios. That means close to 90 models running every day,” he adds.
The engineering progress has been dramatic. In 2021, the first AI baskets took several hours to run overnight, he said. Today, the same models refresh in under a minute. InvestorAi’s tech stack blends computer vision with proprietary techniques, allowing the platform to adapt to shifting market conditions almost in real time.
But with grand scale comes the risk of overfitting, as AI chases patterns that don’t hold up. Keith says they take a different approach. “You should never train for returns. If you do, you might get something, but it will be horribly volatile for a retail investor.”
Instead, InvestorAi trains to minimise risk, drawdowns, or maximise Sharpe ratios. “We also run contenders for every model, so we can switch if performance skews,” Keith adds.
For retail investors, AI’s opacity is often a dealbreaker. How do you trust an algorithm you can’t understand?
Keith recognises that retail investors often rely on past performance, mistakenly believing it indicates future success. To counter this, he provides key metrics like Sharpe and Sortino ratios. He emphasises the importance of tracking win rates as no one is correct all the time.
Crucially, InvestorAi also focuses on exit signals. “Anyone can tell you what to buy. Hardly anyone tells you when to sell. That’s the real trick: helping people lock in profit targets, rebalance portfolios, and compound over time.”
Convenience with Guardrails
InvestorAi markets itself with the promise of “investing as easy as Swiggy.” But unlike food delivery, stock markets are volatile, and investor behaviour under stress can be unpredictable.
Keith notes: “It is one click, but you see everything before you click. The machine doesn’t panic like humans do. In extreme volatility, we’ve even manually switched off intraday products rather than risk over-trading.”
To protect younger, less experienced investors, the platform builds in safeguards. “On our intraday product, every recommendation comes with an automatic stop-loss. You can change it, but the default guardrails are there. I think we have a moral obligation to put protections in place, even if the regulator doesn’t mandate it.”
When asked what 2030 might look like, Keith predicts volatility, seeing the market tumble in the next five years. However, he says, maintaining trust with the customers and the community will make them long-term users of the platform. “Trust comes from constant dialogue, showing why the machine made a call, even when performance dips,” he adds.
The Business Engine
InvestorAi’s revenue model today is B2B2C, and the annual revenue reached ₹13.4 crore as of March 2023.
Keith acknowledges growth limits, having surpassed last year’s figures three and a half times and aiming for five times by year-end. To address future caps, they’re developing institutional products, such as a US long-short fund in collaboration with a Singapore hedge fund, to better democratise investing.
The company recently received its RIA license, allowing it to charge variable fees rather than fixed subscriptions. Direct-to-consumer products are also on the horizon.
With regulators also tightening scrutiny on AI in finance, compliance is no longer optional. Here, Keith strikes a confident note: “When I read SEBI’s AI/ML consultation paper, I was happy. They’re right to worry about companies building on someone else’s foundational models, not knowing what data goes in or out. We were compliant before the regs even came in.”
InvestorAi is walking a tightrope between accessibility and accountability. By fusing cutting-edge AI with convenience, it hopes to make investing feel as intuitive as ordering pizza. But unlike pizza, portfolios come with risks.
The post This Fintech Startup is Making Investing as Easy as Ordering Pizza appeared first on Analytics India Magazine.
