The numbers Brief Episode map Takeaways Charts Quotes Clips Glossary Go next Transcript

Episode 171 · Capital · 58 min

The tag is worth ten per cent

Prequate's founder walked out of KPMG with two other chartered accountants to sell Big Four-grade work to companies that could not afford the Big Four — and found the brand name he had been banking on was worth about ten per cent of the sale. Sixteen years on, his argument to a hall of commerce students is that the credential is the cheapest asset in a finance career, and that the entry-level work which used to buy you a seat is the first thing AI takes.

Founder journey AI & machine learning Education & skilling

Pradyumna

Founder, Prequate · with Vishal Krishna

57:48

Said in this episode

▶ 16:55

400–500

Investment banks serving 6 lakh enterprises

His sizing of the industry: a few hundred organisations covering roughly six lakh enterprises, which is why he calls it a very small niche rather than the money machine students imagine.

▶ 17:40

60–70%

Raises that happen on a personal introduction

Research he cites in conversation: most businesses that successfully raise money do so because someone made a positive introduction, not because the deck won on merit.

▶ 15:52

10%

What the Big Four tag actually sells

His estimate of how much the brand contributed once he was outside it — enough to get a resume read and a better first conversation; the remaining ninety per cent was the founder's own judgement.

▶ 29:44

6×

Energy of an AI query vs a web search

As stated on air. He pairs it with the image of a single web search lighting a bulb in Africa for an hour, which he offers as illustration rather than a measured figure.

▶ 31:45

40k + 20k

The future firm, as he recounts it

A McKinsey chief's description he watched in an interview: forty thousand AI employees and twenty thousand people, against about a hundred thousand employees today. Recounted second-hand.

▶ 4:08

~1%

Indian graduates who land good jobs

Vishal's framing at the top of the session — roughly six million graduates a year, of whom about one per cent get good jobs, with the caveat that the numbers vary by source.

The brief

The argument in sixty seconds

Pradyumna's claim, made on stage at Yeshas Academy to a room of commerce students who have been told all their lives that they are the slow stream, is that the credential is the cheapest asset in a finance career. He barely passed his CA, spent his early years inside KPMG learning how an organisation of a hundred thousand people keeps one low performer from throwing the system into a tizzy, and then left with two other chartered accountants to do Big Four work for medium-sized businesses that could not afford the Big Four. The network they assumed they were carrying out with them turned out to be a fallacy: the tag bought about ten per cent — a better first conversation — and the other ninety per cent was whether a founder believed you understood his business. Investment banking arrived the same accidental way, because clients kept needing money and the firm's one standing policy was to be partners first; the education was discovering that sixty to seventy per cent of businesses that raise capital raise it on a personal introduction rather than on merit. What he hires for now is not the model or the deck — every deck is an AI deck, and he would rather reject the whole pile than hunt for the needle — but neuroplasticity: the chartered accountant on his team who is one of the best semiconductor minds he knows and is talking green hydrogen the next morning. The stakes he leaves the room with are sharper than employability. AI is the first true resource rush since oil, an AI search burns roughly six times a Google search, and the foundation-level work that used to be the ladder into management is the first rung it removes.

Worth your time if you are

Commerce students told they are the slow stream

CAs weighing a Big Four seat against their own firm

Anyone breaking into investment banking without an IIT-IIM badge

Managers watching AI eat the entry-level rung

Episode map

Where the conversation travels

Every block is a chapter, coloured by what it's about. Click any of it to jump straight to that minute on YouTube.

Founder journeyAI & machine learningEducation & skillingHiring & talentImpact & outcomesFundraising

01Cold open: six million graduates, one per cent 0:00 Pawan sets up a live session built to give commerce students outside Bangalore the exposure the CA syllabus never gives them, and Vishal frames the stakes — about six million Indian graduates a year, roughly one per cent of whom land good jobs — before introducing the founder and the student questioner from Zerodha's Tribe programme. 02Why commerce, and why he left KPMG 5:43 He chose commerce because he watched people in big cars call his father sir, then left KPMG because its boundaries — which exist so that one office in Argentina cannot damage the brand worldwide — penalised him for delivering value that was not in the agreement. 03Processes are fantastic, which is the problem 11:02 His counterintuitive answer on process: it exists so the low, average and high performer all operate at the same level, which is compulsory at scale and exactly why you must eventually step outside the framework to find out what is actually out there. 04Risk, payroll and the network fallacy 12:56 Entrepreneurship starts with assuming your own risk is lower than everyone else's, gets survivable once you have solved payroll a couple of times, and punctures the belief that a Big Four network walks out of the door with you. 05Investment banking, by accident 16:36 The firm added capital raising as a layer because clients needed money and its founding policy was to be partners first — and then discovered that sixty to seventy per cent of raises happen on a personal introduction, not on the merit of the deck. 06Shortest path in, and why they love CAs 18:53 The fastest route into investment banking is an IIT then an IIM, where everyone around you is already starting up; for everyone else it is CA plus MBA, or CA plus CFA for private equity, because the CA grind is the only credential he trusts for depth under pressure. 07Neuroplasticity and the two-marshmallow test 20:39 What the job really demands is the ability to learn a business cold in an hour and drop it the next day, plus the patience to work two or three years on a deal — the marshmallow question of giving up one today for two next year. 08Five hires, and the death of the deck 25:00 They hire roughly five people, and look for footballers, dancers and musicians rather than portfolios of identical class projects — because when every submission is an AI deck he would rather reject the whole pile than hunt for the needle. 09The fourth rush: gold, oil, chips, AI 28:40 After gold, oil and the semiconductor rush that birthed Sequoia and Accel, he argues AI is the only genuine transformation since the fifties, compute is now the contested resource, and consumers will one day get an AI bill the way they get an electricity bill. 10Forty thousand AI employees, twenty thousand humans 31:30 He recounts a McKinsey chief describing a future firm of forty thousand AI employees and twenty thousand people against a hundred thousand today, then admits he hit his own enterprise AI rate limit hours before walking on stage — and names the real risk, that the foundation-level tasks young people learn on are the first to go. 11Perception versus perspective 34:05 Vishal's case for supporting real journalism — nobody reads, everybody watches, and the fabric of democracy frays — meets the guest's own indictment of media entering the convenience business, where AI slop is indistinguishable from reporting and only the slow, months-long outfits still do the work. 12FOMO, mental health and a Tuesday morning 38:00 A trained suicide counsellor argues that taboo is what makes a problem grow, that the brain is the most consuming organ you own, and that his generation only had SMS to tell them they were missing out while this one has Instagram doing it continuously. 13The assist nobody remembers 44:10 Vishal's own story of nearly quitting journalism for lack of connections turns into the guest's argument that organisations are families, not leaderboards — for every Zuckerberg there are a hundred thousand engineers, and the competitor in your head does not exist. 14Network is net worth; rockets over beer 47:10 His postgraduate degree in international business mattered less for the syllabus than for the breadth, and he makes the case that the scarce skill of 2026 is relationship building plus the ability to connect two unrelated universes — the reason an investment banker should spend an evening listening to people talk about rockets. 15Accountability, transparency and the harvest 51:34 Accountability is defined as no-follow-up dependability, transparency as the uncomfortable new default where an agent can list every email your team ignored for three days, and Vishal closes on why he left a chauffeured editor's life to stop the machine harvesting his brain.

Takeaways

Ideas to carry out of this hour

The Big Four tag buys the meeting, not the mandate

Three chartered accountants left large firms to do Big Four work for medium-sized businesses that could not afford the Big Four, and assumed the network would come with them. It did not. He puts the tag's contribution at about ten per cent — enough to get your resume read and to have a better conversation — while ninety per cent is a founder deciding whether you actually understand his business. Any company that would have switched from the Big Four to a firm founded yesterday had already found one, because plenty of firms went before them.

▶ 10:14 ▶ 15:38 ▶ 15:52

Process exists to flatten performance, so you must leave it

His answer on whether students should follow the process is deliberately counterintuitive: processes are fantastic, because they make the low, average and high performer operate at the same level, which is compulsory when you employ a hundred thousand people. He credits KPMG for teaching him exactly how large organisations run at scale without one individual throwing the system into a tizzy. But the same framework is what you have to step outside if you want to see the world as it is — and without those foundations, the failure rate on the way out is far higher.

▶ 11:42 ▶ 12:10 ▶ 12:41

Investment banking was a bug fix, and merit is not how money moves

The firm never saw a big opportunity in investment banking; clients kept needing capital, and its founding policy of being partners first meant adding the layer rather than sending them elsewhere. The naive assumption was that effort and email volume would win. The number that reframed the business: sixty to seventy per cent of companies that raise money raise it because somebody made a positive personal introduction. The other education was learning to distinguish a good idea from a good business — the point at which ad hoc help became a vertical.

▶ 16:36 ▶ 17:22 ▶ 17:40

They hire for neuroplasticity, and every deck now looks the same

He loves CAs for the grind — the ability to sit deeply on a problem, with foundational finance assumed — but what he actually screens for is the ability to learn a business cold in an hour, come back out, and do it again the next day on something unrelated. The exhibit is a chartered accountant on his team who is one of the best semiconductor minds he has met and is discussing green hydrogen the following morning. Portfolios no longer help: when everyone submits an AI-generated deck, he would rather reject all of them than hunt for the needle.

▶ 19:55 ▶ 23:20 ▶ 26:43

AI is the first resource rush since oil, and the bill is coming

He places AI in a line that runs gold, oil, then the semiconductor rush that produced Sequoia and Accel, treating the dot-com and e-commerce waves as smaller events by comparison. What is being contested now, in war and in geopolitics alike, is compute. The consumer-facing product is free and the back end is not: an AI query costs roughly six times a web search in energy. His analogy is that everyone gets an electricity bill because nobody can live without electricity, and at some point everyone gets an AI bill for the same reason.

▶ 28:44 ▶ 29:44 ▶ 31:02

The bottom rung is the first thing automation removes

He recounts a McKinsey chief describing an organisation of forty thousand AI employees and twenty thousand people where there are a hundred thousand today, and notes that his own firm's adoption curve took six months. The uncomfortable consequence is generational, not aggregate: the foundation-level tasks young entrants are trained on are precisely what is being replaced, which raises the question of how anyone climbs to management without doing them. He exempts construction, social services, parts of healthcare and some creative fields — and not much else.

▶ 31:45 ▶ 32:01 ▶ 33:33

Media moved into the convenience business; perspective did not

Vishal's charge to the room is that everybody watches, few read, and a generation is building perception rather than perspective while reels do the influencing. The guest, who says he does not read a single newspaper because most of them feel like AI in print, extends it: media entered the same convenience industry as food delivery, optimising for fifteen seconds of interaction over relevance, and half of what scrolls past is AI-generated. The counter-example both land on is the outfit that spends months on one topic and publishes validated work.

▶ 35:05 ▶ 36:20 ▶ 37:33

Breadth is the arbitrage; dependability is the trait

His formulation of the 2026 skill stack is blunt — network is net worth, and relationship building is the one capability a model release cannot make obsolete in two months. Breadth matters more than depth of the same kind: a room of people doing what you do only reinforces or shames you, while an evening spent listening to people argue about the business of rockets lets you weld two separate universes together, which is what the world pays for. Underneath it sits accountability, which he defines as work getting done without a single follow-up.

▶ 49:26 ▶ 50:50 ▶ 53:01

The numbers, drawn

What the episode measures

Every figure below was said on air — timestamps included, caveats kept.

Conversation share

portion of the hour spent on each theme

Founder journey22%

AI & machine learning18%

Education & skilling15%

Hiring & talent13%

Impact & outcomes10%

Fundraising9%

Computed from the chapter map of this episode.

Each wave lands faster than the last

years to transform

As stated in conversation: the internet took fifteen to twenty years, e-commerce seven or eight, and AI is landing in two or three. Upper bounds shown.▶ 30:38

The organisation he expects on the other side

headcount

Recounted second-hand from an interview with a McKinsey chief he watched; the figures are that executive's illustration, not a published forecast.▶ 31:45

Worth keeping

Lines that stay

Whenever I would go outside to deliver things of value, irrespective of whether they were written in an agreement between two organisations, I would usually get penalised — because that's not what you're supposed to be doing.

— Pradyumna ▶ 9:45

Are you ready to give up one marshmallow today if you can get two marshmallows next year? We try and look for people who have that two-marshmallow thinking.

— Pradyumna ▶ 22:08

Everybody gets an electricity bill, because you can't do without electricity. At some point in the future you're going to get an AI bill, because you can't do without AI.

— Pradyumna ▶ 31:02

None of you are reading. None of you are building perspective. You're only building perception.

— Vishal Krishna ▶ 35:05

In 2026, the single skill that will beat any other skill is your ability to build relationships. In two months a model will come out with something that just replaces what you're doing — but it can't change how you build relationships.

— Pradyumna ▶ 49:46

Clips that travel

Short on time? Start here

CAs weighing a Big Four seat against their own firm

Why he walked out of KPMG

Failing the exam, the boundaries that exist so one office cannot damage a global brand, and three CAs deciding it was now or never.

7:44 → 11:02 · 3 min ▶ Watch clip

Founders who think the deck is what raises the money

Investment banking, and the introduction fallacy

How capital raising got bolted on by accident, and the sixty-to-seventy per cent of raises that ride on a personal introduction.

16:36 → 18:53 · 2 min ▶ Watch clip

Anyone breaking into investment banking without an IIT-IIM badge

Neuroplasticity, marshmallows and the AI deck

The clearest eight minutes on what an advisory firm actually screens for — and why a polished AI-built portfolio now hurts you.

20:39 → 28:40 · 8 min ▶ Watch clip

Managers watching AI eat the entry-level rung

The fourth rush and the AI bill

Gold, oil, silicon, AI; six times the energy per query; and the generational problem of automating the work people learn on.

28:40 → 34:05 · 5 min ▶ Watch clip

Students building a career without connections

Network is net worth, rockets over beer

The case for breadth as arbitrage: connect two unrelated fields and you own the one skill he thinks a model cannot copy.

47:10 → 51:34 · 4 min ▶ Watch clip

Glossary

The jargon, unpacked

CA (Chartered Accountant)

India's core accounting qualification, notorious for its failure rates — the credential he barely passed and the one he still hires for, because the grind proves depth under pressure.

Big Four

The four global audit and consulting firms, KPMG among them, whose scale demands the boundaries and processes he eventually left in order to serve mid-sized companies.

Investment banking

Advising companies on raising capital and on transactions; in his sizing, a few hundred Indian firms serving about six lakh enterprises, with pay heavily weighted to closing a deal.

Neuroplasticity

His shorthand for T-shaped learning: the ability to go deep enough on an unfamiliar business in an hour to understand what makes it work, then come back out and repeat it on something else tomorrow.

Marshmallow test

The delayed-gratification experiment he uses as a hiring filter — will you give up one marshmallow today for two next year, which is what a three-year deal cycle demands.

Frontier model

The largest general-purpose AI systems, which he expects to become the next big resource allocators and to arrive with a metered bill attached.

Agentic AI

AI connected directly to your tools — mail, calendar, files — so it can act rather than answer; his example is asking which emails his team left unanswered for three days.

Token

The unit AI systems are billed by; his point is that pricing scales upward with use, so heavier adoption keeps fetching more money from the same user.

Connections

If this resonated, go here next

EP 106 · Enterprise

Full transcript

The whole conversation, searchable

Auto-generated captions, lightly cleaned. Click a timestamp to open that moment on YouTube.