R
RetainEQ·Investor Retention Intelligence
Pilot Build · v1Open Workspace →

▎ A new category in Indian wealth tech

Know which investor will redeem
before they do.

RetainEQ is an AI copilot for Indian mutual fund distributors. The wedge is simple: identify which client relationships need attention first, explain why, and help the broker run better retention conversations this week.

Industry · India MF

SOURCE: AMFI · MAR 2026
SIP STOPPAGE RATIO76%
MAR 2026
MF UNITS REDEEMED < 2 YR73%
SEBI FY23
MONTHLY SIP DISCONTINUED50 LAKH
▲ MAR 2026
INDIAN MFDs ACTIVE~80,000
of 1.73 lakh registered
INDUSTRY AUM₹73.73 LAKH CR
MAR 31 2026
ML CHURN PRODUCTS IN INDIA0
12 platforms analyzed

▎ The Problem

Distributors find out a client left after the folio hits zero.

01

AUM walks out silently

On average, 15–20% of an MFD's AUM redeems each year. At 0.83% trail, that is ₹50–80 lakh of recurring income gone — permanently — for every ₹100 Cr managed.

02

No warning system exists

Of 12 Indian MFD platforms analysed, zero use ML to predict redemptions. The category is empty. The first mover defines the playbook.

03

Replacing AUM takes years

A new ₹1 Cr SIP relationship takes 24–36 months to build. Retaining one takes a single well-timed conversation.

▎ How It Works

An anonymized investor book goes in. A weekly attention list comes out.

STEP 01

Ingest

For a live pilot, the broker shares anonymized CAMS or KFintech history only. PAN, name, phone, and email stay inside their environment.

STEP 02

Rank

RetainEQ scores the book, highlights which households need attention first, and turns the book into a weekly queue instead of a spreadsheet dump.

STEP 03

Explain

Each surfaced account comes with plain-language reasons and a suggested next move, so the broker sees why it is on the list.

STEP 04

Act

The broker works from one shared queue: review, call, add notes, override if needed, and keep all follow-up in the same workflow shell.

▎ The Moat

The model isn't the moat. The data from every phone call is.

A weekend hacker can build a churn model. They cannot replicate twelve months of intervention outcomes across hundreds of distributors. Every “Contacted → Stayed/Left” click trains the next version of the platform. Causal ML on top of proprietary outcome data is what makes this category-defining.

L1 · WEDGEChurn dashboardCopyable in a weekend
L2 · WORKFLOWWhatsApp digests, CAS ingestion, scripts6 months to copy
L3 · FEEDBACK LOOPIntervention outcomes trackedRequires 12+ months of usage
L4 · BENCHMARKSCross-distributor scorecardsRequires 500+ MFDs
L5 · MARKETPLACEAMC-side intelligence layerRequires both sides — uncopyable

▎ See It Live

Synthetic demo today. One real pilot away from proof.

The workspace is loaded with a realistic synthetic book and a working broker workflow shell. The goal is not to look broad. The goal is to show one believable pilot: who needs attention, why, and what the broker should do next.

Open The Workspace →