How to Build and Scale a Business? Ep.1 — Manoj Kohli & Rohit Kapoor

ELI5 / TLDR

The opening lecture of a six-part business-model course at Masters’ Union. Manoj Kohli (the man who scaled Airtel from zero to 650 million customers) frames a business model as four things: make a product, get it to the customer, give them a good experience, and make “stunning” profit doing it. Swiggy CEO Rohit Kapoor joins by video to argue that the durable money is made by spotting large, unstated consumer problems — bad traffic, urban time-poverty — and that AI is collapsing the cost of building anything to near zero. The blunt career advice: stop talking about AI, stop optimising for a B-school specialisation, and just build for five hours a day until you’re a domain expert.

The Full Story

What a business model actually is

Kohli strips it to its bones. A business model is how a company makes a product or service, how it gets that into the customer’s hand, how the customer experiences it, and — non-negotiable — how the company makes money doing it.

“Company should make profit. If you don’t make profit then it’s worthless.”

His benchmark for himself isn’t profit, it’s stunning profit. Global telecom runs at a 30% EBITDA margin; Airtel hit 63%, double the world average. His explanation is half engineering, half national pride: Indians have good minds and a habit of frugality, so an Indian operator ought to wring more out of less.

The two tests: scalable and profitable

The first quality of a good model, he says, is that it scales. A model can be built for one neighbourhood, one city, one region, one country, or the world — and the goal is always to push it up that ladder. When Airtel went from India to 20 countries, roughly 90% of the core model travelled unchanged; only 10% was customised for local conditions like Nigeria.

This sets up his complaint about Indian industry. The companies are good, he says, but most don’t have scale — only Reliance, Airtel, HDFC really do. And he puts the blame mostly on the industry itself, not the government. Roughly 80% of the problem, in his telling, is that Indian firms are too comfortable making good money at home: low global aspiration, low R&D spend, low hunger. The cure is guts — he cites Airtel’s $10.7 billion acquisition of Zain’s African operations across 17 countries, done without government support, as the kind of bet more companies should be willing to make.

Rohit Kapoor: look for what doesn’t change

Kapoor’s segment is the more interesting half, and he opens by inverting the usual framing. Everyone talks about what’s changing; he wants to talk about what’s eternal. (His aside: “people talk about artificial intelligence — I’m also as curious about natural stupidity.”)

The eternal thing, for a consumer business, is the hunt for the unstated need that can eventually be monetised. The word “eventually” is doing heavy lifting. It’s easy to be profitable at small scale, he says, and India is full of small profitable companies — but there’s no fun in that, and he’s openly critical of the Indian habit of punishing companies that invest for the long term.

His method for finding big problems is almost embarrassingly low-tech: put your phone in your pocket and walk around the city with your eyes open, observing rather than judging. What you’ll notice is that Indian cities are getting harder to live in, and there’s no going back. Traffic gets worse every year.

“In my business a bad traffic situation is actually a tailwind.”

Delivery runs on people’s desire to stay home and go out only when they must. Bad traffic manufactures that desire. Other slow-moving trends he points at: relentless rural-to-urban migration (creating housing and co-living opportunities), pollution, and the “money-rich, time-poor” problem of India’s top two-to-three million households — people as wealthy as a Singaporean or Londoner who still can’t buy a Singapore-grade standard of living the moment they step outside their gated condo. His vivid stat: the average car in Delhi moves at 23 km/h, in Bangalore 17. A 2-crore car crawling at 17 km/h; an electric bike does 25. Spot a large, expensive friction, and you’ve found a business.

Why “TAM” is a bad word

Kapoor’s sharpest point is an attack on Total Addressable Market as a way to evaluate a company.

“What was the time of quick commerce four years back? Zero. Literally zero.”

Five years ago, anyone sizing a 10-minute-delivery market by existing demand would have gotten zero, because the demand didn’t exist yet — it had to be created. He reaches for a Bezos line: across 25 years of Amazon, the one durable lesson was that customers always want things cheaper and faster. Quick commerce works, in his read, because it removes friction. Order from home and it arrives inside the same 10-minute span of attention; anything that takes an hour introduces coordination friction (you might be out, asleep, unavailable) that quick commerce simply deletes.

The AI section: expertise is toast

When he turns to AI, Kapoor’s framing is again “what’s certain vs what’s experimental.” His certain conclusions are stark:

“Expertise is dead. Experience is toast.”

A media manager with 10 years’ experience, a designer, a software developer — “toast,” he says, carving out only true hand-craft like a chef. The flip side is the opportunity: a young person who spends one focused year on AI can be as valuable as someone with 5–10 years of domain experience. He keeps a 25-year-old AI mentor on his calendar for an unmovable daily hour — and that mentor isn’t a CS graduate, just a BBA who spent two and a half years on a voice-AI project.

His warning to students: don’t talk about AI, and don’t confuse typing ChatGPT prompts with knowing AI — “the panwala across the road is also doing that.” The real skill is building context, using coding tools, learning to wield them powerfully. The payoff is that the startup cost of building anything is collapsing — design, procurement, a whole website in three hours with no agency — roughly 20x faster and 100x cheaper. That democratisation, more than the money flowing to the big AI labs, is the real shift.

Inside Swiggy, his AI framework has three legs: (1) the top 400–500 leaders must become fully AI-native regardless of function; (2) where the world has already proven a use case — contact centres, voice calling, co-pilots — adopt at 100% without debate, the main internal job being to feed the tools enough context to stop them hallucinating; (3) run an engine of experiments to find what’s relevant. He describes a new vertical called Crew built “100% AI” by default — humans deployed only where AI can’t cope — and one small city where, for two months, the local team executes AI’s morning instructions without asking questions. “The co-pilot is the human being. The co-pilot is not AI.”

He’s careful not to be a zealot. Both over-leaning (burning money) and under-leaning (losing competitive ground) are real risks; vendor winners change every three months; there’s heavy PR machinery to discount for. But: “If you’re a skeptic on AI, you’re just making life harder for yourself.”

Kohli’s three quotients: IQ, EQ, AQ

Kohli closes by repackaging Kapoor’s advice into a formula. AI multiplies your IQ — take a 100 and it becomes 150 or 200 depending on use. But IQ isn’t enough.

EQ — emotional quotient — comes from leaving the virtual world and meeting real people: shopkeepers, dealers, consumers. As CEO he travelled 20 days a month and spent the first half of every day in the market, listening for both articulated demands and unarticulated expectations. That’s where you find the problem and the solution.

The third, which he claims as his own coinage, is AQ — adversity quotient, a polite word for hunger. Here he gets deliberately provocative: candidates from metros (Delhi, Bombay, Bangalore) have low hunger; small-town candidates have far more, and he’s spent 20–30 years preferentially recruiting them. Hunger and passion, he insists, beat an MBA.

The closing advice: just build

When a student asks what to actually do, the answer from both men converges into one instruction. Put the phone aside, get a decent laptop, pay for AI subscriptions (“tell your parents to fund that”), and build five hours a day — alone, not by watching others build. The hottest profile won’t be a brand or marketing manager; it’ll be the builder. Pick your orientation (brand, sales, whatever) after you can build.

Kohli also fields a speed-vs-quality question with an old Airtel slogan: “speed first, quality always.” In practice that meant shipping a product at 90–95% rather than waiting for 100%, then closing the last 5% within a week or two. “We have to tie our laces while running.”

Key Takeaways

A business model has four parts — make it, deliver it, make the experience good, make profit. Profit isn’t optional; aim for stunning margins (Airtel ran 63% vs a 30% telecom average).
Two tests of a good model: scalability (can it climb from neighbourhood to world?) and profitability. India’s companies are good but under-scaled, and Kohli pins ~80% of the blame on industry timidity, not government.
Find businesses by observing slow, irreversible trends — worsening traffic, urban migration, pollution, the money-rich/time-poor squeeze — rather than by sizing markets that already exist.
TAM is a trap. The big markets (quick commerce, mobile telephony) had a measured demand of zero before someone created them. Bezos’s one durable lesson: customers always want cheaper and faster.
Quick commerce works because it removes coordination friction — delivery inside the same 10-minute span of attention, before you’ve moved or gone to sleep.
AI verdict: “expertise is dead, experience is toast.” One focused year on AI ≈ 5–10 years of legacy domain experience. Hand-craft (e.g. a chef) is the exception.
Don’t confuse prompting with AI literacy — the differentiator is building context and using coding/agentic tools well. Startup costs are collapsing ~20x faster, ~100x cheaper.
Corporate AI playbook: make senior leaders AI-native, adopt proven use cases at 100% without debate, and run an experiment engine. Default new builds to “100% AI, humans only where AI fails.”
Kohli’s formula: IQ (multiplied by AI) + EQ (built by meeting real consumers, not screens) + AQ (adversity/hunger — he recruits from small towns for it).
The career instruction, from both: stop talking, put the phone down, pay for AI tools, and build alone for five hours a day. The builder beats the specialist.
Operating slogan: “speed first, quality always” — ship at 90–95%, close the gap in a week, don’t wait for 100%.

Claude’s Take

This is a course opener, and it behaves like one — light, anecdote-heavy, more vibe than rigour. The transcript is also genuinely mangled in places (auto-captions turned “Zomato” into “Zumato,” “Lenskart” into “Lenskard,” and several sentences into word salad), so some nuance is lost.

Kapoor is the reason to watch. The “observe the eternal, don’t chase the changing” framing and the takedown of TAM are the two ideas worth keeping — both are real, both cut against the standard pitch-deck reflex, and both come from someone running a business where they’re load-bearing rather than theoretical. The traffic-as-tailwind observation is the kind of inversion that’s obvious only after someone says it.

Kohli is more of a mixed bag. The four-part model and the scalability ladder are clean and useful. But the “stunning profitability” and “guts” register is motivational-keynote more than analysis, and the metro-vs-small-town hunger claim is the kind of confident anecdote-as-law that a recruiter likes and a statistician would wince at — it’s a hiring bias dressed as a principle. His IQ/EQ/AQ packaging is tidy but mostly a restatement of Kapoor’s “go outside and look at consumers.”

The “expertise is toast” line is the most quotable and the most overconfident thing in the video. It’s directionally interesting and almost certainly overstated as a blanket claim — but the underlying advice (build, don’t just prompt; one focused year compounds) is sound regardless of whether the dramatic version is true.

A 6: two sharp reusable ideas (eternal-vs-changing, anti-TAM) wrapped in a lot of pep-talk, with a transcript that fights you the whole way.