Investing Made Easy: How InvestorAi Brings AI to Every Investor

For Gen Z and Millennials, convenience drives every decision—whether it’s ordering food or managing finances. With the same simplicity that Swiggy and Zepto brought to doorstep deliveries, InvestorAi, an AI-powered equity investment platform, aims to make investing seamless and smart.

Founded by Bruce Keith, InvestorAi is among the few fintechs globally that built its AI foundational model entirely from scratch, rather than layering on existing frameworks. Backed by SEBI authorisation and investor Ashish Kacholia, the platform leverages AI, computer vision, and genetic algorithms to translate complex financial data into actionable, outperforming equity baskets for retail investors.

With over 30,000 subscribers, ₹80 crore ARR, and partnerships with India’s top brokers, InvestorAi’s B2B2C model is quietly reshaping how India invests. In a conversation with CiOL, Bruce Keith, CEO & Co-founder of InvestorAi, delves into the company’s AI-first approach, engineering decisions, compliance, and the future of AI investing. 

You chose to build a foundational AI model from scratch rather than using existing ones—what were the decisive trade-offs and lessons from that choice?

Today, most companies build on top of someone else’s large language model, usually American ones. But when we started in 2018, those didn’t exist. So, necessity became the mother of invention.

Fortunately, we had people from IBM Watson backgrounds who knew how to engineer something unique. We wanted to create something akin to Renaissance Technologies—an AI powerhouse, but accessible to retail investors.

Our system is entirely proprietary, right down to the servers used for experimentation. The biggest advantage is complete control over input data, avoiding the hallucinations common in large models.

If I could revisit the process, I’d probably bring in more asset management expertise earlier. But the core system—I’d keep it exactly the same because it gives us full accuracy and control.

InvestorAi uses a mix of AI, computer vision, and genetic algorithms—what unique data sources or modelling approaches give you an edge, and how do you validate their reliability?

The uniqueness lies in how we engineer features. Early on, our system took 4–5 hours to run each day—today, it runs in under a minute.

We took inspiration from what the Chinese did with DeepSeek—reimagining engineering instead of reinventing the wheel. We even use wave technology to detect meaningful signals in noisy data.

Validation happens on two fronts: first, real investor outcomes—if people make money, they stay. Second, internally, our investment operations team monitors all portfolios through risk-adjusted metrics and win rates.

We also maintain two contender models behind every production model to ensure agility. Currently, our MTF basket has an 84% win rate, and our longest-running portfolio has delivered 45% CAGR since 2021.

When you say your equity baskets “consistently outperform market indices”, how do you define and measure outperformance, and how do you ensure transparency for retail users?

From an SEBI perspective, we have to be cautious about phrasing. But yes—our models have outperformed.

Our benchmark is the Nifty 500, as it’s the most realistic comparison for retail investors who might otherwise buy an ETF.

While asset managers can create custom benchmarks, we prefer simplicity. I also welcome SEBI’s upcoming rule requiring CA-certified validation of returns—it’ll bring transparency and allow us to demonstrate audited performance data publicly.

With intraday and MTF baskets aimed at retail investors, how do you manage execution risk, liquidity constraints, and volatility exposure across broker partners?

Initially, our intraday feature offered both long and short baskets. But we realised that on some days, offering both isn’t logical.

So we introduced “Intelligent Entry”, where our AI decides whether to show long, short, both, or none—depending on market volatility.

During unpredictable events, like election weeks or major tariff announcements, we’ve manually suspended recommendations. Missing one opportunity is better than risking a wrong call.

For context, only 8–9% of intraday traders make money according to SEBI data, whereas our recent win rate exceeds 70%—a significant difference.

Operating on a B2B2C model with major brokers, how do you address conflicts of interest and ensure that AI-driven recommendations remain unbiased and compliant with SEBI norms?

The SEBI framework assumes firms use third-party models like OpenAI, where data origin is unclear. Since our model is built entirely in-house, we know every data point, ensuring compliance from the ground up.

Our model’s bias-free nature comes from relying solely on market data, not sentiment data like Twitter, which we found unhelpful.

Bias risks mainly appear in human-in-loop checks—so we monitor those closely. We also re-tune models constantly to prevent stagnation or overfitting.

As AI-driven investing scales, what are the biggest risks to sustaining alpha and investor trust, and how do you plan to mitigate issues like model decay, data drift, or market crowding?

Absolutely. Anyone who thinks a model can run untouched is mistaken.

We separate responsibilities across three teams:
• Investment Operations: Tracks portfolio metrics (performance, Sharpe, and Sortino ratios).
• Investment Office: Oversees regulated activities and compliance.
• Model Group: Monitors model health, decay, and contenders.

We run 20+ production models daily, each with two contenders, totalling 60+ models. Every cycle, we benchmark by running 100,000 random combinations to assess performance reliability.

We’re also advocating for an industry-wide transparency framework similar to PMS Bazaar, where AI investment performance can be publicly compared and validated.

What’s next for InvestorAi—expansion, new asset classes, or deeper personalization for India’s younger investors

We’re expanding deeper into asset management, complementing our broker partnerships, and exploring opportunities for the Indian diaspora via GIFT City frameworks for cross-border investing.

On personalisation—we’re cautious. Financial personalisation can quickly become biased advice. Just because you like tech stocks doesn’t mean you should only get tech recommendations.

We believe in responsible personalisation—helpful, not risky. The goal is to make investing as simple as Swiggy but as transparent as SEBI.