a+ Startup studio · Thesis 01 + me

πŸͺΊ Nest Finance Β· Research & Thesis Β· v1

Why India still trusts a human with its money

Apps made investing a few taps. So why do 7 in 10 rupees going into equity funds still come through a person β€” and what happens when AI can do everything except the trust?

A working thesis Β· September 2026


Apps like Groww and Zerodha have made investing easy. Anyone can buy a mutual fund, an ETF or an insurance policy in a few taps. Yet about 7 in 10 rupees going into equity mutual funds still come through a person, not an app.

That's because an app can do the investing, but it can't tell you what to buy. Apps show recommendations, but when it comes to money, Indians trust humans. Usually it's someone smart in their local community whom they call "my financial advisor."

Via a distributor (MFD) Direct / app

Equity MF inflows Β· Franklin Templeton analysis of AMFI data

A note on names: most of these people are officially mutual fund distributors (MFDs). Only SEBI-registered advisers may call themselves "financial advisers" or "IFAs." On this page we say MFD.

What MFDs actually do

MFDs mainly do two things:

  1. Help families decide what to buy, and stay invested. They're the trusted voice who says "don't sell" when markets crash. This is why clients stay with them for decades.
  2. Handle everything around the investment. Sign-ups, ID checks, research, portfolio reports, insurance forms, tracking commissions, answering questions, follow-ups.

The first is the real value. The second eats most of their week.

Job 1 β€” the real value

Trust & conviction

The reason clients stay for decades. Human. AI can't replace it.

Job 2 β€” eats the week

Everything else

Forms, research, reports, servicing, follow-ups. This is what AI can now do.

Nest is an AI-native Prudent

We give India's MFDs an AI-powered back office, so they can remain the trusted face for their clients and let AI do the rest.

Today, big networks like Prudent and NJ provide this back office β€” but with branches across the country, large teams and human relationship managers. They keep a big share of the MFD's income in return.


Prudent: how big this can get

Prudent started in 2000 and launched its MFD network in 2006.

36,880MFDs (Mar 2026)
20.7 lakhinvestors served
143locations
β‚Ή1.4L crclient money (Jun 2026)
β‚Ή75 crlast-quarter profit, +44%
~β‚Ή11,700 crcompany value (Jul 2026)
Dec '21
β‚Ή48,400 cr
Jun '26
β‚Ή1.4 lakh cr

Client money looked after β€” Prudent (current AUM)

Prudent listed in May 2022 at β‚Ή630 a share; it traded around β‚Ή2,860 in July 2026. Last quarter it earned β‚Ή348 crore β€” about β‚Ή190 crore went to its MFDs and β‚Ή75 crore was profit, up 44% in a year. It has also started buying MFD businesses (Karvy's fund business in 2021; Indus Capital β€” β‚Ή2,030 crore of client money β€” for β‚Ή124 crore in 2025) and is already moving on AI with a tool for its MFDs.

NJ: how it grew

In 1994, two fresh graduates in Surat β€” Neeraj Choksi and Jignesh Desai β€” started selling mutual funds. Early on, they recruited insurance agents standing outside LIC offices and trained them to sell funds.

2003
2017
β‚Ή47,500 cr
2026
β‚Ή3.2 lakh cr mobilised

NJ β€” 2003 β‚Ή300 cr Β· 2017 β‚Ή47,500 cr AUM Β· 2026 β‚Ή3.2 lakh cr mobilised (cumulative gross, per NJ β€” not directly comparable to Prudent's current AUM above)

Its commissions earned grew β‚Ή443 crore (FY17) β†’ β‚Ή874 crore (FY21) β†’ β‚Ή1,300 crore (FY22). Today it reports over 56,900 MFDs in its network, according to NJ.


What AI can now do

Everything in "job 2" can now be handled by AI. The trust β€” job 1 β€” stays human.

What the MFD needsPrudent & NJ todayWith Nest
Client sign-up & ID checksForms and back-office staffβœ“ AI, in minutes
Research & fund shortlistsResearch teamsβœ“ AI
Portfolio reviews & reportsPrepared by staffβœ“ AI, in seconds
Insurance needs analysis & formsManual paperworkβœ“ AI, from a recorded chat
Commission tracking & payoutsAccounts teamsβœ“ Automated
Answering client questionsCall centresβœ“ AI assistant, 24/7
Marketing materialDesign teamsβœ“ AI, personalised
Meeting prep & follow-upsRelationship managersβœ“ AI drafts, MFD approves
Trust & conviction (job 1)The MFDThe MFD. This stays human.

Why now

The pay cut. In April 2026, SEBI's new commission rules (replacing TER with BER/SER and separating out GST) cut many smaller distributors' take-home by around 15% β€” hitting the smallest, non-registered MFDs hardest. They're looking for a better deal.

The technology. AI can finally do the back-office work well β€” and even Prudent has started building AI tools, so the race has begun.

The gap. India has about 1.8 lakh individual MFDs, but only about 13,000 run a serious business. The rest need help to grow.

~13,000 of 1.8 lakh individual MFDs run a serious book (AUM > β‚Ή25 cr) Β· Cafemutual

Why Nest wins

MFDs pay twice today. Prudent and NJ keep a share of their commission β€” mainly for research, licences and the platform. On top of that, MFDs pay for their own staff to handle client calls and servicing, which the networks don't really cover. Nest fixes both.

Today
~60–65% blended
Nest
up to 90% (our offer)

Share of commission that reaches the MFD β€” Prudent passes ~60–65% on average; smaller MFDs get less, large ones already 80–90%. With Nest, servicing is also included.

 With Prudent / NJWith Nest
Share of commission that reaches the MFD~60–65% on average β€” smaller MFDs get less, large ones already 80–90%Up to 90% (our offer)
Client servicing: calls, forms, reportsThe MFD hires and pays their own staffIncluded β€” AI does most of it

And when MFDs retire, we can buy their businesses β€” just like Prudent does β€” and run them with AI.

Who we're looking for

This is a starting point, not a decision. If you think it's wrong, or you'd build it differently, that's the conversation I want β€” not a reason to stay quiet.

Co-founder 1 Β· Distribution

Has spent years recruiting and building an MFD network β€” at a fund house or a network like Prudent or NJ. Not an MFD themselves; someone who has signed up, trained and grown distributors at scale.

Co-founder 2 Β· AI

Has shipped production AI products to real users β€” ideally LLM / agent systems that automate document-heavy, regulated workflows (onboarding & KYC, research, report generation) or customer-facing assistants. Zero-to-users, not prototypes.

+ me Think you're one of these two? That's the whole application.
Sources for the numbers above

This is one of the ideas a+ is building. If it’s yours to build, reply β€œ+ me” β†’

See what else a+ is building β†’