AI stock assistant for Indian retail investors: A practical guide to smarter trading with Swastika Sarthi

Key Takeaways

• Sarthi is Swastika Investmart's AI stock research assistant. It compresses the reading, screening and comparison work that sits behind an investment idea, but it does not place orders and it does not hand you a verdict.
• The order of operations matters. You open a demat and trading account with a SEBI-registered broker and finish your KYC first, then bring an AI assistant into the research that happens inside that account.
• The most valuable thing an AI research tool produces is not a recommendation but a structured bull case and bear case that you can interrogate before you commit money.
• Anything the assistant surfaces should be verifiable against primary sources, which for Indian equities means company disclosures on nseindia.com and bseindia.com and regulatory material on sebi.gov.in.
• A sustainable routine is quieter than most beginners expect: a few minutes on most days, a longer weekly review, and a genuine rethink of each holding once a quarter as results arrive.
• Position sizing, diversification and exit discipline decide your results far more than idea quality does, and no software can enforce those habits on your behalf.

Why Retail Research in India Is Harder Than It Looks

The problem facing an Indian retail investor in 2026 is not a shortage of information. It is the opposite. More than five thousand companies are listed between the NSE and the BSE, every one of them files quarterly results, shareholding patterns, related party disclosures and a steady stream of material announcements, and all of it is public. On top of that sits a layer of commentary: brokerage notes, television panels, newsletters, and an enormous volume of unattributed advice on social media and messaging groups. The raw material for a good decision is freely available, which is precisely why the bottleneck has moved elsewhere.

That bottleneck is attention. A working professional who invests on the side has perhaps thirty focused minutes a day. Reading one annual report properly consumes several hours. So most people quietly stop reading and start reacting instead, buying on a headline or a forwarded tip because the alternative feels impossible. The result is a portfolio assembled from fragments, where the investor cannot say in plain language why they own any particular stock. When the position falls twenty percent, there is no thesis to test against, so the decision to hold or sell becomes purely emotional.

Sarthi exists to attack that specific gap. It is an AI research assistant built into the Swastika ecosystem, and its job is to do the first pass of reading and organising so that your limited attention lands on judgement rather than on data gathering. Ask it about a company and it will pull together the financial trend, the valuation context, the recent disclosures and the visible risks, then lay out the case for and against the stock in language you can actually argue with. What it will not do is decide for you, and that boundary is the most important thing to understand before you start.

How to Start Investing in the Indian Stock Market With AI Tools

Before any research tool becomes useful, you need somewhere for the decision to land. In India that means two linked accounts: a demat account, which holds your shares in electronic form with a depository, and a trading account, which routes your buy and sell orders to the exchanges. Both are opened through a broker registered with SEBI, and at Swastika the two are opened together in a single application. The process is almost entirely digital now, and the practical requirements are short:

• PAN card, which is mandatory and is the identifier your holdings are mapped to.
• Aadhaar for identity and address verification, usually completed through an OTP-based eKYC flow.
• A bank account in your own name, with a cancelled cheque or recent statement, since payouts can only be made to a bank account matching your name.
• A signature specimen and a short in-person verification step, generally handled by a guided video call.

Once the account is live, resist the urge to trade on day one. The more useful first step is to fund the account modestly and define, in writing, what you are trying to do. Someone building a retirement corpus over twenty years and someone trying to generate income from swing trades need entirely different research, and an AI assistant cannot infer which one you are. Write down your horizon, the maximum you are willing to lose on any single position, and roughly how many stocks you intend to hold. This takes ten minutes and it changes every conversation you will later have with the tool.

Sarthi enters at the third step, once the account exists and the mandate is clear. Used well, it works in three modes. The first is screening, where you describe the kind of company you want in ordinary language, something like a midcap industrial with falling debt and improving margins, and the assistant narrows thousands of listed names to a handful worth a closer look. The second is risk assessment, where you point it at a specific stock and ask what could go wrong, prompting it to surface leverage, promoter pledging, customer concentration, governance flags and liquidity problems that a price chart hides completely. The third is framing, where you ask it to build the bull case and the bear case side by side so you can see what each side of the trade is actually assuming.

The habit worth forming early is treating the assistant as a well-read colleague rather than an oracle. Push back on what it tells you. If it says margins improved, ask over what period and whether the improvement came from pricing, from input costs, or from a one-off item. If it flags rising debt, ask whether the borrowing funded capacity that is already generating revenue. This back and forth is where the real learning happens, and it is also how you catch the occasional error, because language models can be confidently wrong and the only defence is a reader who checks.

Manual Research vs. Sarthi-Assisted Research

The comparison below is about where your time goes, not about replacing your own thinking. Notice that the final row does not change, because the decision stays with you in both columns.

Navigating Indian Market Regulations: SEBI, NSE, and BSE Guidelines for Retail Traders

Using an AI tool responsibly in India starts with understanding who regulates what. SEBI, the Securities and Exchange Board of India, is the statutory regulator for the securities market. It registers and supervises brokers, investment advisers and research analysts, and it sets the conduct rules those intermediaries operate under. The NSE and the BSE are the exchanges where trades are matched, and they also publish the disclosures that listed companies are required to file. Between sebi.gov.in, nseindia.com and bseindia.com you can independently verify almost anything material about a listed Indian company, and you should treat those sites as the final word whenever a secondary source and a filing disagree.

The distinction that matters most for tools like Sarthi is the one between research and advice. Personalised investment advice, meaning a recommendation tailored to your particular financial situation, is a regulated activity in India that only a SEBI-registered investment adviser may provide. Research and analysis is a different thing. A research assistant that organises public information, summarises financials and lays out arguments is giving you material to think with, not a personal recommendation to act on. Reading Sarthi's output as though it were bespoke advice is a misuse of the tool, and it also removes the accountability structure the regulator built for actual advice. If your situation genuinely calls for a personalised plan, the right step is to engage a registered adviser and check their registration number directly on the SEBI website.

A few other rules are worth knowing because they shape what you can and cannot do inside your account. Indian equity trades settle on a T+1 basis, so funds and shares move faster than many older guides suggest and you need to plan liquidity accordingly. Exchanges apply circuit limits to individual stocks and to the broader indices, which means a position can become untradeable at exactly the moment you most want out. Insider trading rules apply to everyone, so acting on information that is not yet public is an offence regardless of how you came by it. And the pool of illiquid smallcaps that screeners love is subject to surveillance measures that can restrict trading with little notice. None of this is exotic, but all of it is easier to absorb before it affects you than during a bad week.

Building a Sustainable Investing Routine With AI and Data-Driven Insights

Most retail investors fail not because they pick bad companies but because their process is unsustainable. They begin with intense enthusiasm, monitor prices constantly for a few weeks, burn out, and then stop paying attention entirely until something frightening happens. An AI assistant makes it possible to run a much lighter routine, because the time-consuming part of staying informed is exactly the part it does well. The goal is a rhythm you can keep for years, not a sprint.

On a daily basis, ten minutes is enough. Ask what actually changed for the companies you own, look for genuine developments such as results, credit rating actions, large orders, management changes or regulatory notices, and consciously ignore price movement that carries no news with it. The point of this check is to notice when a fact underlying your thesis has shifted, not to admire or worry about the day's percentage.

Once a week, spend closer to an hour on the portfolio as a whole. Look at your allocation across sectors and market capitalisations, since drift happens silently when winners grow. Revisit any position that has moved sharply and ask whether the reason is something you understood when you bought it. This is also the natural slot for looking at one new idea properly, which is a far better cadence than reacting to whatever is trending. One well-understood addition per month beats twelve impulsive ones.

Quarterly, when results arrive, do the real work. For each holding, write two or three sentences explaining why you still own it, then use Sarthi to challenge that statement by asking for the current bear case. If the reason you originally bought has quietly stopped being true, that is the signal to exit, and it is a much more reliable trigger than a price target. Keeping these notes creates something no tool can give you, which is an honest record of your own reasoning over time. Reading last year's notes is uncomfortable and it is the fastest way to improve.

A Worked Example: Using Sarthi to Shortlist a Stock

The walkthrough that follows is entirely hypothetical. The company does not exist, every number below is invented purely to illustrate the process, and none of it should be read as a view on any real stock. Think of it as a hypothetical example of what a session looks like when the tool is used properly.

Suppose our imaginary investor wants exposure to Indian industrial capex without buying the largest and most expensive names. She starts by describing that to Sarthi in ordinary language: midcap manufacturers supplying capital goods, revenue growing at least fifteen percent a year over the last three years, debt reducing rather than rising, and a market capitalisation between five thousand and twenty thousand crore rupees. The assistant returns a shortlist of a handful of names, and she picks one fictional company, call it Bharat Precision Systems, to examine in detail.

The first pass looks encouraging. In this invented scenario, revenue has grown from roughly 900 crore to about 1,500 crore over three years, operating margin has improved from twelve percent to a little over sixteen percent, and net debt to equity has fallen from 0.9 to 0.35. The stock trades at around 32 times trailing earnings against a sector median near 24. Left there, the story writes itself as a quality compounder that the market has begun to notice, and this is exactly the point where an enthusiastic investor buys.

Instead she asks for the risks, and the picture becomes more textured. The assistant flags that a single client accounts for something like thirty-eight percent of revenue, that the margin gain coincided with a period of falling steel prices rather than any change in pricing power, that receivable days have stretched from 62 to 91, and that a large capacity expansion is due to commission next year with the associated depreciation already contracted. None of these are fatal. All of them change what she is actually betting on.

She then asks for both cases explicitly. The bull case is that the capacity expansion arrives into a genuine capex upcycle, the client base broadens as new lines qualify with additional customers, and earnings grow fast enough that the premium multiple looks reasonable in hindsight. The bear case is that steel prices reverse and take the margin gain with them, the anchor client renegotiates or dual-sources, and the new capacity sits underutilised while depreciation and interest hit the profit and loss statement immediately. Reading them side by side, she can see the two variables that decide the outcome, which are input cost direction and client concentration, and that clarity is the real output of the session.

Her decision reflects that uncertainty rather than resolving it. She verifies the concentration disclosure and the receivables figure in the company's own filing on the exchange website, because those two numbers now carry the thesis and she is not willing to act on a summary alone. She then takes a starter position of two percent of her portfolio instead of the five percent she might have taken on the first impression, and writes down one line: "revisit after the next two quarters, specifically to check whether the client concentration falls and whether margins hold if steel prices rise." If the concentration improves she will add. If margins collapse with input costs, she has already decided what that means. The assistant did not make this decision, but it made a considered decision possible in under an hour.

Frequently Asked Questions

Is Sarthi Giving Me Investment Advice?

No. Sarthi is a research assistant that organises and analyses publicly available information about listed companies. It does not know your income, liabilities, tax position, time horizon or risk tolerance, and personalised investment advice in India is a regulated activity restricted to SEBI-registered investment advisers. Use its output as material for your own judgement, and consult a registered adviser if you want a plan built around your specific circumstances.

Can an AI Assistant Predict Which Stocks Will Go Up?

No tool can do that reliably, and any tool claiming otherwise deserves suspicion rather than money. What Sarthi can do is explain what a company's numbers and disclosures currently show, and set out the conditions under which the investment case would work or fail. That is genuinely useful, because it lets you monitor those specific conditions instead of monitoring the price and guessing.

How Much Money Do I Need to Start?

Far less than most people assume, since you can buy a single share and there is no minimum portfolio size. The more useful framing is that your first few months should be sized so that mistakes teach you something without hurting, which for many beginners means a few thousand rupees a month invested regularly. Increase the amount once your process is stable rather than once your confidence is high, because the two are not the same thing.

Should I Verify What the AI Tells Me?

Yes, and this is not optional for anything material. Any number you are actually going to act on should be confirmed in the company's own filings on nseindia.com or bseindia.com, because language models can misread or misattribute figures with complete confidence. Treat the assistant as a fast first reader whose work you spot-check, which is roughly how you would treat a capable junior analyst.

Does Using an AI Tool Mean I Can Skip Learning the Basics?

The opposite is true. The tool is most valuable to someone who understands what a cash flow statement is, why promoter pledging matters and how dilution affects per-share value, because that person can ask sharper questions and recognise a weak answer. Someone with no foundation cannot tell a good analysis from a plausible-sounding one. Sarthi shortens the reading, but it does not remove the need to understand what you are reading.

How Often Should I Check My Portfolio?

For a long-term investor, a brief daily glance for genuine news, a weekly portfolio review and a proper quarterly reassessment around results is enough. Checking prices many times a day reliably increases anxiety and trading costs without improving returns. The discipline worth building is reacting to changes in the business rather than to changes in the quote.

Conclusion

The honest promise of AI in retail investing is narrower than the marketing around it, and considerably more useful. It does not tell you which stock will double. It removes the hours of reading that stand between you and an informed opinion, applies the same risk checklist to every company regardless of how tired you are, and makes it easy to hear the argument against your own idea before you fund it. For someone who invests around a full-time job, that shift in where the effort goes is the difference between a portfolio built on reasoning and one built on tips.

The parts that stay with you are the parts that matter most. You still decide what to own, how much to risk on any single name, and what would make you sell. You still verify the numbers that carry your thesis against the filings on the exchange websites, and you still understand that only a SEBI-registered adviser can give you advice tailored to your own situation. An assistant that respects those boundaries is far more valuable than one that pretends to dissolve them, because the boundaries are what keep you solvent through a bad year.

If you are ready to put this into practice, the sequence is simple. Open your account, write down your mandate before you place a single order, and let research assistance do the reading while you do the thinking. Open your trading and demat account here