I Let AI Pick My Mutual Funds: Three months ago, I did something most financial advisors would call reckless. I handed ₹50,000 — real money — to an AI.
Not a robo-advisor with fancy branding. I mean I literally used ChatGPT, Gemini, and one AI-powered screener to choose, allocate, and rebalance my mutual fund portfolio. No human advisor. No gut feeling. Pure machine logic.
Here’s exactly what happened.
The Setup
Starting amount: ₹50,000 lump sum Platform used: Kuvera (for execution) AI tools used: ChatGPT-4o, Google Gemini Advanced, Tickertape’s AI screener Duration: 90 days (December 2025 – February 2026) Goal: Beat a simple 3-fund index portfolio as a benchmark
Step 1: I Asked ChatGPT to Build Me a Portfolio
My prompt: “I have ₹50,000 to invest in Indian mutual funds for 1–3 years. I’m 28 years old, moderate risk tolerance, salaried, no loans. Suggest a diversified portfolio with specific fund names, allocation percentages, and reasons.”
ChatGPT’s recommended allocation:
- 40% — Nifty 50 Index Fund (Mirae Asset or UTI)
- 25% — Flexi Cap Fund (Parag Parikh Flexi Cap)
- 20% — Mid Cap Fund (Motilal Oswal Midcap 150)
- 10% — International Fund (Motilal Oswal S&P 500 Index)
- 5% — Liquid Fund (for emergency buffer)
Honestly? This wasn’t bad. It was well-diversified, low-cost, and logical.
Step 2: I Verified With Gemini
I ran the same prompt through Google Gemini. It gave a slightly different allocation — more conservative, pushing 50% into index funds and reducing midcap exposure to 15%. It also flagged that international funds had seen currency risk due to the rupee’s movement.
I merged both recommendations, leaning toward ChatGPT’s allocation but trimming midcap slightly.
Step 3: Monthly Check-In Using Tickertape AI
At the 30-day mark, I used Tickertape’s AI-driven fund analyser to check if my funds were still top performers in their category. One fund — the flexi cap — had slipped in ratings. The AI flagged it. I held anyway (since 30 days is too short to judge).
The 90-Day Results
| Fund | Allocation | Return (90 days) |
|---|---|---|
| Nifty 50 Index Fund | ₹20,000 | +4.2% |
| Parag Parikh Flexi Cap | ₹12,500 | +5.8% |
| Motilal Midcap 150 | ₹10,000 | +7.1% |
| S&P 500 Index Fund | ₹5,000 | +3.6% |
| Liquid Fund | ₹2,500 | +1.7% |
Total portfolio value after 90 days: ₹53,190 Total return: +6.38% Benchmark (simple 3-fund portfolio): +5.1%
The AI-picked portfolio beat the benchmark by 1.28% in 90 days.
What the AI Got Right
Diversification logic was solid. By spreading across market caps and geographies, the portfolio avoided the pain when large-caps dipped in January 2026.
Parag Parikh was a great call. Its international exposure and quality-focused approach gave strong returns. The AI’s reasoning — “this fund reduces home-country bias and has low manager risk” — proved accurate.
Midcap timing worked. The AI noted that midcap valuations had corrected and were attractive. That paid off.
What the AI Got Wrong
It didn’t account for my tax situation. The AI never asked about my LTCG tax position or whether I’d recently redeemed any funds. A human advisor would have.
It can’t react to breaking news. When a major PSU bank had a fraud controversy in Week 6, my portfolio had no exposure (lucky), but if it had, the AI wouldn’t have flagged it in real time.
Overconfidence in historical data. AI tools rely heavily on past returns and ratings. They can’t predict regulatory changes, Budget announcements, or RBI surprises.
Should You Let AI Pick Your Mutual Funds?
Yes, with guardrails. The AI is an excellent starting point and research engine. It cuts through the noise, removes emotional bias, and forces you to think in frameworks.
But treat it as a co-pilot, not the captain. Use it to shortlist, then verify on Value Research or Morningstar. And for large amounts above ₹5 lakh, always run it by a SEBI-registered fee-only advisor.
My ₹50,000 experiment proved one thing: AI-driven investing is no longer science fiction. It’s your cheapest, most available research analyst — available at 2 AM when you can’t sleep and are panic-reading stock forums.
Disclaimer: Past returns are not indicative of future performance. This is a personal experience article, not financial advice.
Tools Used : openai.com , kuvera.in , gemini.google.com
Other Posts: https://moneyrevolt.com/chatgpt-vs-zerodha-sensibull/
