Can AI manage your stock portfolio? Zerodha users are trying

Zerodha, one of India’s largest stockbrokers, has become the first to integrate MCP with Claude AI, a large language model (LLM) built by Anthropic. With a few simple technical steps, users can now prompt the AI assistant to access and analyse their portfolio, delivering insights in plain English.

Want to know your portfolio’s Sharpe ratio, diversification score, or unrealized losses? Just ask. Claude can also estimate portfolio volatility, assess F&O positions, or even suggest buy/sell actions based on investor-defined goals or benchmarks—all through conversational prompts.

This could be a prompt that simply asks to ‘list my stocks’; current positions; unrealized profit and losses; available margin or buying power; analysis of portfolio’s risk across metrics like beta; estimated volatility of portfolio; diversification score; Sharpe ratio; analysis of stock portfolio; analysis of open F&O positions.

That’s now possible thanks to MCP, an open-source framework released in November 2024 that allows large language models like Claude to access third-party data and tools. By setting up an MCP server, users can give the AI controlled access to their apps and accounts—effectively creating a prompts-driven, portfolio-aware financial assistant.

Early experiments

While Zerodha is the only broker to officially offer MCP integration so far, some users have managed to connect Claude with other platforms like Groww—though it requires more technical skill. And tech-savvy investors have begun sharing their experiments online.

Roughly 300–600 users are trying out the Zerodha-Claude MCP each day, according to the platform, with about 16,000 having accessed it in the past week—still a small fraction of Zerodha’s 7.96 million active user base as of March 2025.

To be sure, Claude’s responses sometimes include disclaimers stating that its recommendations are for educational or general purposes only—but these warnings don’t appear with every output.

One such user, Kirubakaran Rajendran, 38, used Claude to review and adjust his equity portfolio on Zerodha. A Chennai-based full-time algo trader since 2017, Rajendran’s post detailing the interaction quickly went viral on X.

“The rise of AI tool has made automations much more easy, even for non-coders, and cost-effective,” he said.

Rajendran, who has traded on Zerodha since 2010, was clear that the recommendations came from Claude AI, not the broker itself. Zerodha, being a discount platform, doesn’t offer investment advice or research.

He cautions against blind reliance: “With AI, it is important make the right prompts. For example, a prompt like — consider the volatility of the stock and accordingly recommend a stock as I don’t want higher deviation with my portfolio — can lead to recommendations that don’t make the portfolio more volatile. It would identify stocks that it assesses to fit with the portfolio’s overall volatility. Of course, one still needs to do their own bit of research.”

Others are using Claude not just to manage portfolios, but to align them with the philosophies of legendary investors.

Hemanth Kumar Boyapati, a 30-year-old cybersecurity analyst from Vijayawada, prompts Claude to evaluate his holdings in the style of investing legends.

“I have prompted the AI to tell me which parts of my stock portfolio I should hold, sell or add to tune my portfolio in the style of these great investors. I prompt the AI to give me a comprehensive rationale for what it recommends. To fine-tune the recommendations, I upload investment thesis — which are widely-published — of these investors in the AI,” he said.

“Besides this, I also use it for tracking news developments that I may have missed — or the upcoming events related to the stocks in my portfolio. Having said that, I try not to over-churn my portfolio, unless there is a strong case to make any adjustment.”

Can AI be held accountable?

Several licensed research analysts (RAs) or registered investment advisors (RIAs) offer curated stock portfolios through platforms like Smallcase, operating under Securities and Exchange Board of India’s (Sebi) regulations that require explicit registration and accountability for investment advice.

Anyone offering specific investment advice is required to register as an RA or an RIA. But AI tools like Claude, capable of generating stock suggestions and portfolio insights, are challenging that framework.

“As portfolio managers, we are not selling recommendations, we are selling credibility and conviction. Can an investor bet his wealth on AI-driven recommendations? What kind of conviction does the AI has on its own stock calls? Isn’t it susceptible to frequent changes as and when there is change in performances? These are the questions that remain to be answered,” said Abhishek Banerjee, co-founder of LotusDew Wealth, a Hyderabad-based investment advisor.

“There are also questions around accountability—who is responsible if something goes wrong. For anyone to recommend any stocks, you need to fall under the regulatory framework of Sebi by obtaining the RA licence, but how will AI be regulated?” he added.

Others welcome the disruption.

“This is actually something to be welcomed. Now, there is a new avenue of competition. Complacent portfolio managers and advisors will find it tough as AI use becomes more pervasive in the capital markets. But there will always be a place for credible analysts and portfolio managers, who consistently generate outperformance for their clients,” said Dick Hosy Mody, founder of Ethical Advisers, an investment advisory firm.

A new challenge for regulators

India’s investment advisory rules are built around human accountability. But as AI systems increasingly mediate financial decisions, the regulatory perimeter becomes harder to define.

“As artificial Intelligence tools become increasingly accessible to retail investors, especially through seamless integrations like those offered by brokers such as Zerodha, the regulatory perimeter will need to evolve both in depth and clarity,” said Sumit Agrawal, founder of Regstreet Law Advisors.

That evolution, Agrawal said, could begin with a distinct regulatory sandbox or a light-touch code of conduct, especially for brokers and fintechs enabling such access.

“Futuristically, the use of LLMs and AI for personalized investment recommendations, even when initiated by the investor, raises complex questions under India’s Sebi (Investment Advisers) Regulations, 2013. Today, the regulatory framework assumes a ‘human agent’ offering advice. However, when AI becomes the intermediary, and especially if such AI evolves from being a passive tool to an active recommender, the definitions of ‘advice’, ‘execution’, and ‘intermediation’ may require substantial rethinking,” Agrawal said.

“At present, if a platform facilitates AI-driven recommendations that amount to specific buy or sell advice, it could risk straying into unregistered investment advisory territory, depending on the degree of discretion, targeting, and real-time applicability involved,” Agrawal said.

“Questions of auditability, accountability, explainability (of AI output), and investor protection must be front and centre. A hybrid regulatory framework, where human oversight remains accountable for AI outputs or where AI-based interactions are tagged as ‘informational’ and not ‘advisory’, might be a temporary path,” he added.

In a consultation paper dated 20 June, titled Guidelines for Responsible Usage of AI/ML in Indian Securities Markets, Sebi proposed a range of measures, including human oversight, safeguards around client data, investor disclosures, and the use of regulatory sandboxes to test AI applications. However, the paper does not specifically address MCP-style AI integrations.

The regulator also proposed a “regulatory lite” framework for AI/ML use in the securities market—limited to applications that do not directly impact clients or customers of regulated entities through their business operations.

What brokers say

Zerodha has clarified that MCP does not enable Claude to do anything fundamentally new, it simply automates the data extraction process that users could already perform manually.

“The MCP only enables users to pull their portfolio into Claude, which they would otherwise have to manually feed into Claude, which could be done simply using a screenshot of the user’s trading app. The recommendations from Claude can be sought by users with or without the MCP,” said Nikhil Aralimatti, senior business analyst, Zerodha.

Zerodha CTO Kailash Nadh, in a recent blog post, warned of broader systemic concerns.

“Strictly from a user perspective, it feels liberating to be able to access services outside of their walled gardens and bloated UIs riddled with dark patterns. 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 that occur in the midst of AI orchestration across critical services? How would those be debugged and traced? How would accountability be established? Will regulations and governance models emerge for AI-mediated automagical actions across services? Whether the long-term implications of MCP’s viral, cross-cutting spread will be net positive or not, is unclear to me.”

Yogesh Chande, partner at Shardul Amarchand Mangaldas & Co., believes brokers and exchanges may need to upgrade their internal controls.

“The current framework has adequate internal controls and checks at the level of a stock broker internally while providing using AI tools, and externally as well, by way of making sufficient disclosure to the clients intending to use AI tools. As and when the usage of AI tools increases, Sebi and stock exchanges might have to enhance this framework.”

Agrawal added that it may be worth exploring whether existing algorithmic advisory regulations (for robo-advisors) can be extended or refashioned to cover such LLM-powered tools.

“This will be a defining challenge for financial regulators globally—ensuring that innovation does not outpace investor safeguards, especially for retail users who may not fully understand the probabilistic and non-fiduciary nature of LLM outputs.”

Human advisors still have an edge—for now

Despite the technology’s promise, several RIAs believe that AI cannot replace the human touch in financial advice.

“Role of RIA involves several aspects including understanding the goals of the individual — which sometimes may not be explicit — but implicit. It requires multiple rounds of deep conversations across, to create a holistic financial plan — covering risks, goals, succession planning, etc.. So, AI can serve as a research tool, but cannot replace an advisor,” said Kavitha Menon, founder of Probitus Wealth.

“Financial planning requires deeper and higher quality conversations, which can only emerge from human interactions,” said Vishal Dhawan, founder of Plan Ahead Wealth Advisors.