Hello,  

Yesterday Zerodha launched Kite MCP , a protocol that lets users plug their trading accounts into an AI assistant like Claude. The idea is to ask a bot to analyze your portfolio, backtest trades, or build financial dashboards, all in natural language. 

“These AI tools have become so good that you can have a reasonably competent personal financial assistant.”  

- Nithin Kamath, Zerodha 

It sounds liberating to investors and it is quite powerful. But there is something about the news that raises questions we are not discussing enough. The explainability. Are we outsourcing financial judgment to systems we don’t understand, and can’t audit?  

How do we trust a system that can’t be audited or has no explanation of how and why it arrived at certain conclusions.  

While the legalities of AI generated investment strategies are currently in murky waters, it is being celebrated as an innovation of the future. In this edition of Infinite Loop, I will try to cut through the noise and objectively assess three solid questions around AI generated investment strategies.  

Zerodha’s own CTO, Kailash Nadh, frames the tension well: 

"It moves a considerable amount of control from service providers to users, but at the same time, it concentrates decision-making and mediation in the hands of AI blackboxes... What about errors and failures with real-world implications?”  

This moment, where AI doesn’t just advise but starts acting across services, makes it critical to pause and ask: 

• Are robo-advisors legal under current financial law? 

• Is the model commercially viable, or just VC-viable? 

• Should you actually trust an AI with your money? 

• And what kind of robo-advice is actually possible today? 

Let’s get into it. 

Are robo-advisors legal? 

Yes, but with an important asterisk. 

Robo-advisory platforms that provide financial advice using algorithms have moved from novelty to norm in the last decade. With AI now powering everything from investment decisions to tax-saving nudges, it is no longer a question of “can we do this” but “should we”? 

Beneath the sleek interfaces and real time portfolio rebalancing lies a knot of legal ambiguity, liability concerns, and systemic risks that regulators and users alike are only beginning to untangle.  

Although there is no blanket law globally that bans or legalizes robo-advisors. Instead, in a global context, regulators have opted for a tech neutral approach. If you are giving investment advice, whether via human or machine, you fall under existing advisory regulations.  

Where AI is used in financial services, it deems fit to consider that existing regulations on consumer protection, fairness laws, model risk frameworks, anti-manipulation law still apply regardless of whether decisions come from a human or an algorithm.  

In India, SEBI does not explicitly ban robo-advisory either. But it does hold anyone offering Invesment advice accountable under the Investment Advisor Regulation. SEBI’s recent stride on finfluencers offering investment advice without license is a sign that the regulator is tightening oversight on advisory where transparency is in question.  

So yes, they are legal, but they do fall outside the purview of traditional regulations yet.  

Are they viable? 

Technically? Absolutely. Commercially? Not really.  

Platforms like Betterment and Wealthfront have built profitable models, using AI to tailor portfolios to user goals and risk appetites. Globally, AI in the finance market is estimated at $38.36 billion, projected to reach $190.33 billion by 2030 (Yahoo Finance). 

But that viability doesn’t scale across all user segments. As user data shows: 

• Urban professionals love app-based wealth tools 

• Small business owners still prefer Excel or their trusted financial planner 

• Many users abandon robo-advisory tools halfway due to complexity, poor UX, or lack of hand-holding 

This has been the story so far and there’s potential for newer entrants to succeed where giants have mostly failed. It’ll be a game of the technology getting better as well as the companies deploying it doing something out of the ordinary to change their fate.  

Tech alone doesn’t build trust. And in financial matters, trust is everything. 

So, are they worth trusting? 

This is where the shine wears off. 

A robo-advisory may process data faster and more rationally than a human, but rationality ≠ accountability. AI use case in general comes with a historic bias based on the data it has been trained on. The OECD flags key concerns : 

Picture Credit: OECD report  

• Black-box models : Users can’t audit why a recommendation was made 

• Bias amplification : Skewed data = skewed advice 

• Liability gaps : If an AI-led portfolio tanks, who’s accountable? The platform? The CEO? The user? 

If millions follow advice from similarly trained models, what happens when those models fail? We may be looking at a new class of systemic risk , driven not by human greed, but by code convergence. 

The ethical question isn’t “Can I trust an algorithm?” It is: Do I understand what I’m trusting it to do?  

What’s actually possible today? 

Robo-advisors aren’t science fiction. They are already: 

• Offer real-time tax-saving suggestions 

• Calculate retirement gaps 

• Nudge users on savings behavior 

• Track net worth across bank accounts and asset classes 

And increasingly, they integrate with platforms offering paperless KYC , real-time UPI transactions , and embedded credit products . 

But the future isn’t fully robotic. The most promising model today? Hybrid advisory, where AI offers clarity, humans offer context.  

In fact, several regulators, including the US SEC, suggest that this hybrid approach offers the most resilient model for risk oversight, user understanding, and equitable access. 

Is it time for celebration or caution? 

The robo-advisory revolution isn’t illegal, infeasible, or inherently flawed. But it is incomplete. 

The most significant risk isn’t the algorithm itself; it’s our (potentially) blind faith in it. As we increasingly hand over financial control to machines, the burden of oversight, ethics, and user education becomes heavier, not lighter. 

Here's hoping that with industry collaboration, innovation and responsible companies like Zerodha - we can one day find an AI investment advisor that's more like a friend in need, rather than an algorithmic master.