7 mutual fund patterns SEBI RIAs miss — and what AI finds instead

India has over 14,000 mutual fund schemes across 44 AMCs. A SEBI-registered RIA typically tracks 30–80 of them closely and has a mental model for the rest. That model is usually built from the same three sources: AMFI NAV data, a Morningstar/ValueResearch star rating, and the fund's own factsheet. These are good inputs. They are not enough.

We built a multi-agent research pipeline — one AI agent each for fund quality, performance, portfolio composition, and macro/category context — that analyses any Indian mutual fund scheme in under a minute. Here are seven patterns it surfaces consistently that rarely appear in a standard scorecard.

1. Rolling consistency collapse

A fund's 5-year CAGR looks strong at 18%. But its 1-year rolling consistency — the percentage of overlapping 1-year windows where it delivered a positive return — is only 38%. That means in nearly two out of three rolling windows, an investor entering the fund would have seen negative returns within the next year. Most screeners show the trailing CAGR. Almost none show the rolling consistency score.

Every research report on this platform computes both 1-year and 3-year rolling consistency automatically. A fund that is positive in 92% of rolling windows is structurally different from one that delivered its entire return in the last window — even if the trailing CAGR is identical.

2. Category rank divergence

Small-cap funds as a category delivered 28% over 12 months. The fund you are evaluating delivered 19%. A 9% gap against category median could mean the manager is defensively positioned, or that their stock picks have underperformed. The difference matters.

The macro/category agent automatically computes where a fund ranks within its AMFI category — percentile, distance from median, and distance from top/bottom decile. It does this across 1-year, 3-year, and 5-year windows. This peer-relative positioning is computed fresh for every research run, not pulled from a pre-built rating.

3. Mandate drift in holdings

A fund categorised as "Large Cap" by SEBI holds 62% in large caps as of the latest portfolio disclosure. SEBI mandates a minimum of 80%. This is technically visible in the monthly factsheet PDF. It rarely surfaces in a summary report.

The composition agent checks a fund's actual holdings — asset allocation, sector weights, top-10 positions — against the stated SEBI category mandate. If a Large Cap fund has meaningful mid-cap or small-cap exposure, or a Flexi Cap fund is running 90% in large caps (effectively a Large Cap fund at Flexi Cap expense ratios), it notes the gap factually in every report.

4. Hidden concentration in "diversified" funds

A multi-cap fund holds 45 stocks. That sounds diversified. But its top 10 holdings account for 68% of AUM, and the Herfindahl-Hirschman Index (HHI) is 820 — well above the 500 threshold that indicates moderate concentration. The label says diversified; the math says concentrated.

Every research report computes the HHI concentration index and top-10 combined weight from the latest portfolio snapshot. The composition agent interprets these numbers in context — a focused fund with HHI of 1200 is playing its stated strategy; a "diversified" fund with the same HHI is not.

5. Portfolio overlap across client holdings

A client holds three equity funds: a Flexi Cap, a Large & Mid Cap, and a Focused fund. The weighted overlap between the Flexi Cap and Large & Mid Cap is 43%. Eight of the same stocks appear in all three funds, contributing a combined 22% of portfolio weight. The client thinks they are diversified across three funds. In practice, nearly half the money is in the same stocks.

The portfolio overlap tool computes pairwise overlap (weighted Jaccard similarity), duplicate stock exposures across funds, the effective number of unique stocks, and overall portfolio concentration (HHI). For RIAs managing 25+ clients, this turns a 30-minute manual check into a one-click analysis.

6. The SIP timing illusion

A fund shows a 5-year CAGR of 16%. An investor running a ₹10,000/month SIP over the same period earned an XIRR of 11.2%. The gap is not an error — it is the fundamental difference between lumpsum entry at the start and rupee-cost averaging over the period. In a market that rose sharply in the final year, the SIP investor bought most units at higher NAVs.

Every research report computes both the lumpsum CAGR and the SIP XIRR for 3-year and 5-year windows. Showing them side by side makes it immediately clear whether a fund's headline return translates into real SIP investor outcomes. When the gap is large, the NAV trajectory (not just the endpoint) matters.

7. New fund data traps

A fund launched 4 months ago shows a "3-year CAGR" of 55% on some platforms. This is 4 months of return annualised to 3 years — mathematically computed but practically meaningless. An NFO that gained 12% in 90 days is not a 55% CAGR fund. Yet automated screeners often show this number without a warning.

The platform suppresses any metric where the fund does not have sufficient history. A 3-year return requires at least ~3 years of NAV data. A 5-year SIP simulation requires at least 80% of the 5-year period. When a fund is too new, the report shows a clear "Data Notice" explaining exactly what is available, what is not, and why certain metrics display N/A instead of a misleading number.

Informational research only — not investment advice. Flirting with Finance is not a SEBI-registered Investment Adviser or Research Analyst. All data is sourced from AMFI, MFAPI, and public AMC disclosures and may not be complete or current.