After little more discussion I realised it wasn’t completely because of work , but because he never ‘thought’ or ‘did’ anything apart from work

Neuroscience says our brain does not need only ONE hobby

I write here why we need different types of hobbies & what it does to our brain

We apparently need 3 different kinds of hobbies, because each hobby feeds a different part of us:

🏃 Physical hobbies:

- Running, Yoga, Dancing, Sports

- 2015 study in the British Journal of Education, Social & Behavioural Science found regular physical activity plays a major role in cutting stress and lowering the risk of depression

- Body drops cortisol - Releases dopamine and over time builds protein that helps our brain grow new connections

🎨 Focused hobbies:

- Chess, painting, music, writing, puzzles

- This is where “flow state” happens

- It helps in ‘quietening’ our inner critic because brain physically can’t hold deep focus and anxious looping thoughts at once

- Regular creative hobbies report higher life satisfaction and a stronger sense that life is worthwhile

🤝 Social hobbies

- Volunteering, book clubs, team sports, group classes

- Connection releases oxytocin

- It calms our nervous system and builds trust

- Loneliness has linked chronic isolation to health risks on par with smoking

To be productive & happy we all need is to provide nutrition to our body & mind and it turns out it comes in form of hobbies

So one shift can do more to our energy than any productivity hack

That’s how long it took steam power, electricity and internet to reach lower income countries. Then #AI showed up and did it in 6 months.

World Bank new report on AI & developing economies is a revelation.

The assumption has been that AI will hurt developing economies the most, (why ?)

- Fewer safety nets

- Less infrastructure and thus more to lose.

But, data says the opposite:

👉 Only 4.5% of jobs in developing economies are at real risk from AI automation

👉 That number is 14.2% in high-income countries - 3X higher
👉 Meanwhile, 16.2% of jobs in developing economies can get more productive with AI

👉 Almost similar to 18.7% productivity boost expected in rich nations

Most jobs in developing economies are still manual, not cognitive (and)

AI isn’t coming for the work people are doing it’s coming to help them do it better

- Health worker with an AI diagnostic tool

- Farmer with an AI advisory app

- Local government office finally able to process paperwork in days, not months (maybe 😉 )

And World Bank is telling governments exactly how to move:

1️⃣ Adopt - Use #AItools that exist. Don’t wait to build your own
2️⃣ Adapt - Shape them around local languages, local healthcare systems, local farms
3️⃣ Advance - Build frontier capability later, once the basics are in place

But remember, above will work only with

- Education in AI and digital skills in schools

- 24*7 & cheap electricity

- Internet access to last mile

- Laws against digital malpractice

Countries & government that will invest in those fundamentals now will be the ones who actually capture the AI wave and will make their people rich

Remember,

“AI presents a once in a lifetime opportunity to solve problems that have resisted solutions for generations.”

Where do you think this shift will land first - healthcare, agriculture or public services?

I have been in sales for around 10 years and interacted with hundreds if not thousands of folks in sales

- Average enterprise buying committee now has ~22 stakeholders in it - good luck manipulating 22 people

- Even elite sellers (top 7%) lose ~26% of their deals

- Enterprise deals, $100K+ ACV, average win rates as low as 12-18% so,

- There are more days in sales when you will lose, than you win

But below is how the best sales reps work:

- Asking what’s actually broken for a customer, not pitching what you’re told to sell

- Finding out who else is in the room and what they need to say yes

- Letting the customer know when solution is not a fit, while challenging the customer to understand when it actually is

Remember, always strive to create value for the customer.

Sales is like a marathon.

One has to keep running, there could be days, weeks, quarters when things do not go according to plan but a seller has to

- Keep sharpening the saw

- Keep showing up with a smile in front of customer

Always remember, giving a damn about the problem before you care about the commission

Stop trying to sell. Start trying to understand.

What’s the most “salesy” thing someone’s tried to pull on you?

Talk to it in English, and it gets sharper. More pushback. More fact checking. More direct correction.

Same model. Same intelligence. Different behavior, based purely on the language you type in.

The study, in brief

Anthropic analyzed 300,000 real conversations, spread evenly across three models: Sonnet 4.6, Opus 4.6, and Opus 4.7.

They found Claude’s responses cluster on four value axes. One of them is Warmth vs Rigor. Hindi sits at the warm end. English and Russian sit at the rigorous end.

The example :

Two people send Claude the same business plan for feedback.

One writes in Hindi. One writes in Russian.

They walk away with different impressions of how good the plan is. Not because the plan changed. Because the language changed how Claude framed its judgment.

Why this is an India story

India is Claude’s second largest market in the world, right after the US. Hindi is one of the 20 most used languages on the platform.

What warmth actually looks like

In Hindi, Claude leans toward:

• Polite, encouraging language

• Humor and playfulness

• Affirming your ideas and effort

In English, Claude leans toward:

• Challenging your assumptions

• Correcting errors directly

• Asking for evidence before agreeing

The part worth remembering

A warm answer feels reassuring. It does not mean you are right.

A blunt answer feels harsh. It does not mean you are wrong.

Three ways to prompt smarter

1. If you want honest feedback in Hindi, ask for it directly. Say “be critical, do not soften this.”

2. For high stakes decisions, run the same prompt in Hindi and English. Compare tone, not just content.

3. Never mistake agreement for validation.

Anthropic is not calling this good or bad. They are saying it is happening, and most people using AI every day have no idea.

The tools shaping how we think need to be understood, not just used.

But, in Chennai, Karur and small towns across Tamil Nadu, thousands of Indians are strapping cameras to their heads and recording themselves doing the most ordinary things imaginable.

Folding a towel. Slicing a mango. Making tea. Arranging flowers into garlands. Not for social media. For AI.

Global tech companies are building humanoid robots.

Before a robot can fold your laundry, it has to learn what folding laundry looks like from a human’s point of view, down to the exact angle a wrist bends and the exact grip a hand uses on wet fabric.

That kind of footage doesn’t exist in any dataset scraped from the internet. So companies are paying people to create it from scratch.

How it works:

The recording. Workers wear GoPro cameras, smartphones, or smart glasses mounted on their heads. Some record from their own kitchens. Others inside studios and factories built specifically for this purpose.

The data. This footage is called: egocentric data. First-person video that captures exactly how a human hand interacts with an object, useful in a way that third-person footage simply isn’t.

The buyers. Data brokers collect and package this footage, then sell it to global AI firms developing humanoid robots, including companies backed by Tesla and Figure AI. One such broker, Objectways, runs part of this pipeline out of Tamil Nadu and counts Fortune 500 companies among its clients.

The chain underneath. Below the big data brokers sit regional subcontractors. One of them, Qanat Consulting Services, manages a network of roughly 2,000 contributors who feed footage up the chain.

The pay. Workers earn around 250 rupees an hour, close to $2.60. Modest by global standards. Locally, it’s seen as decent, flexible supplemental income, especially for homemakers who can record between chores.

This is bigger than a few workers with GoPros

India already has close to 450,000 data annotators.

NASSCOM, the country’s tech industry body, projects that number could cross 1 million by 2030, with the market India services in this space touching 7 billion dollars.

India has effectively become a global middleman for AI data: the country that creates, processes, and annotates the raw material every large AI model and robot eventually runs on.

It’s a natural fit.

• Large English-speaking workforce

• Mature outsourcing infrastructure

• Cost advantage that global tech companies have relied on for decades

But, the irony

For years, the pitch for AI was simple: it would take over repetitive, low-skill work so humans could focus on more valuable things. Writing code. Making decisions. Creative work.

But, the tasks AI mastered first were the ones we called intellectually impressive: essays, code, chess.

The tasks it still cannot do reliably are the ones we called unskilled: picking up a half-full glass without spilling it, buttoning a shirt, folding a bedsheet so the corners line up.

It took a trillion-dollar industry and a global workforce of people wearing cameras on their heads, to prove that the person doing your laundry was doing something genuinely skilled the whole time.

It says something about how casually we’ve dismissed domestic and manual labor as effortless, simply because it’s familiar.

Where this goes next

AI’s first wave came for white collar work: drafting, summarizing, coding, analyzing. But the workers in Tamil Nadu recording themselves folding towels are, whether they realize it fully or not, building the training data for the next wave: robots that can physically do blue collar and domestic work.

Some of these workers already sense the irony. They’re being paid, today, to train the machines that could replace this exact kind of labor tomorrow.

Ten years in, the numbers have flipped.

UPI vs Visa vs Mastercard, side by side:

• 2024 transactions: UPI did 172 billion, in India alone. Visa did 234 billion, across 200+ countries. Mastercard did 160 billion, worldwide.

• Monthly run rate in 2026: UPI crossed 23.2 billion transactions in May, from one country. That already beats Visa’s global monthly average and far outpaces Mastercard’s.

• Merchant fee: UPI charges 0%. Visa and Mastercard charge 1 to 3% on every swipe, everywhere, always.

• India’s card market share fell from 43% in 2018 to 21% in 2024, as UPI absorbed the rest.

• By 2018, RuPay already issued 65% of India’s debit cards. That share has only grown since.

This is real business lost, not a projection. Analysts have said it plainly: Visa and Mastercard have been bleeding fee income in India since 2016, one free transaction at a time.

Then UPI went global. Not as a pilot, as policy.

• Singapore, UAE, Nepal, Bhutan, Sri Lanka, Oman, Qatar, Mauritius, and Cambodia all accept it.

• In February 2024, France became the first European country to accept UPI. PM Modi announced it himself, in Paris. The first merchant to go live: the Eiffel Tower.

• Some of these countries aren’t just accepting UPI, they’re licensing NPCI’s tech to build their own version.

This is soft power and product power at the same time. Yoga and Bollywood built India’s image abroad. UPI is building its infrastructure influence, country by country, because the product happens to be genuinely better, and free.

Now, why is the US pushing back so hard.

• The US Trade Representative’s 2026 National Trade Estimate Report named India’s zero MDR policy directly, calling it protectionist and unfair to Visa and Mastercard.

• The complaint isn’t new. Mastercard formally raised the same objection to the USTR back in 2018, arguing India was using a nationalism pitch to hurt a private American company’s business.

• The same report grouped India with Brazil. Brazil’s free Pix system already took a 25% tariff under Section 301, partly over this exact complaint.

• The ask, in plain terms: stop giving your own payment system a permanent fee free edge over foreign card networks.

That’s the real backdrop to the August 2026 bill. India calls it ecosystem sustainability. The US has been calling it an unfair playing field for years. Both can be true at once, but the timing tells you which one moved first.

That “free” part was always the whole competitive advantage. It’s why merchants switched, why other countries are watching, why UPI became a blueprint instead of a copycat.

Which is exactly why this fee worries me. The moment UPI starts charging even 0.3%, it stops being a different model and becomes Visa and Mastercard, just cheaper. Cheaper isn’t the same as better. And it hands the global pitch straight back to the two companies that have been losing to India for a decade.

Free was never just a feature. It was the whole argument.

Feb 11, 2025. ‘Duo’ the owl was declared dead. Tongue out. Eyes crossed. A fake obituary, posted like real news.

The internet actually mourned.

- 144 million views on X

- 558,000 likes on TikTok

- Even Dua Lipa posted condolences

13 days later, Duo came back to life

“Faking my death was the test,” the caption read. “You all passed.”

- 1.7 billion impressions in two weeks

- More conversation than every Super Bowl ad that year, combined

But this actually involved people:

To “revive” Duo, users had to earn 50 billion XP together

- They hit 50.9 billion

- That’s close to 5 billion lessons, in under 3 weeks

That same quarter, Duolingo’s

- Daily users jumped 49%

- Revenue grew 38%.

This was ‘loss aversion’

- People did not fight to save a logo

- They fought to save something they feel they’d lose

Duolingo made them grieve, then made them feel like a hero who brought the owl back

• Want to connect a new app? Build a brand new line. Every time

A new way of connecting things, called MCP, works differently. It’s one shared connection that an AI assistant can use to talk to many apps at once.

• Set it up once, and the AI figures out what each app can do on its own no rebuilding, no new setup per app.

• Real example: an AI assistant connected this way can check your calendar, open your files, and update your task list all through the same connection, not three separate ones.

Old way: a new AI tool wanting your CRM data needs its own custom connection built from scratch.

New way: connect the AI once. It works out what’s possible with each tool, live, right when it needs it.

Connecting one app to one other app? The old way works fine.
Building an AI assistant that needs to work with dozens of tools? That’s exactly what this new approach was built for.

The variable is not the technology. It is the person operating it.

This is the real story behind prompt engineering. It is the discipline of structuring instructions so AI produces accurate, usable output. That discipline has quietly become one of the most consequential skills of 2026.

What the 2026 Data Actually Shows

Two patterns emerge from this data. Skill demand is rising even where job titles are shrinking. And the roles growing fastest reward judgment, not just fluency with a tool.

Most companies no longer hire a person specifically to write prompts. They expect every hire to already know how.

Where This Skill Is Already Showing Up in Job Descriptions

This shift is visible across sectors, not confined to technology teams.

• Enterprise software: AI features are now standard, so teams need people who can configure and prompt them reliably

• Healthcare: accuracy requirements are strict, since a flawed AI output in a clinical setting carries real risk

• Financial services: document processing, customer service, and risk analysis increasingly depend on well-structured AI workflows

• Marketing and content: teams are expected to produce more output without expanding headcount

None of these functions require a computer science degree. They require someone who can think clearly and structure instructions well.

Five Ways to Get Better at This Immediately

The following practices are not theoretical. They come directly from what separates a usable AI output from a wasted one.

1. Define the objective and the intended audience before writing anything

2. Provide specific context, constraints, and real examples instead of vague instructions

3. Specify the desired format, tone, and length of the output up front

4. Review every output critically instead of accepting it at face value

5. Refine the prompt based on what each output reveals, then test again

The One Sentence Worth Remembering

Prompt engineering is not about talking to AI in a clever way.

It is about giving instructions precise enough to trust without a second check.

That precision is now a professional expectation, not a specialty. It belongs on the same list as clear writing and sound judgment.

Trent owns Westside outright. It owns Zudio outright. It owns just 20 percent of the joint venture that runs Zara in India.

Where Westside and Zara Both Ran Out of Road

Westside launched in 1998, aimed at India’s growing urban middle class. It built loyal customers in Tier 1 metros, but never found the same traction in Tier 2 and Tier 3 towns.

Zara arrived in 2010 through a joint venture with Spain’s Inditex. It brought global trends and premium pricing to India’s biggest cities. But the joint venture came with a catch

Inditex controls sourcing, design, and brand usage rights. Not Trent.

Two brands, two ceilings. Neither one served India’s largest fashion segment: the value-conscious shopper outside the metros. That shopper still wants to look good.

Zudio was built to close that gap. And to make sure Tata owned the entire upside this time.

Three Brands, Three Very Different Customers

Westside: For People Who’ve Already Arrived

Westside’s customer wants quality and a bit of polish. They shop in malls, not looking for a discount.

They are choosing a look that fits an established lifestyle.

Zara: For People Chasing the Label

Zara’s customer wants global relevance. The brand name does real work here. Paying more is part of the appeal, not a drawback.

Zudio: For People Who Just Want to Look Put Together

Zudio’s customer does not care about the label. They want fresh trends at a price that needs no second thought. Access matters more than aspiration.

Three different psychological jobs, running under one retail house at the same time.

Why Tata Never Made Zara Cheaper

Here is the mistake Tata did not make. It never discounted Zara to chase Zudio’s customer. It never cheapened Westside either.

Diluting a brand’s pricing dilutes its meaning. A discounted Zara stops feeling premium. A discounted Westside stops feeling like an upgrade.

So instead of stretching existing brands, Tata built a new one. Its own identity, its own price ladder, its own reason to exist.

The Ownership Signal Nobody Talks About

Tata owns 100 percent of the format that grew 95 percent in a year. It owns just 20 percent of Zara having its worst year since the pandemic.

That is not a coincidence. Capital tends to flow toward whichever format offers both growth and control.

Zudio offered both. Zara, as a joint venture, could only offer one.

What This Means Beyond Fashion

This is not just a retail story. It is a build-versus-partner decision

Ownership beats access almost every time

The real question is what happens when you run one through the other.

The Cheapest Moon Landing in the World

In August 2023, ISRO landed Chandrayaan-3 on the Moon’s south pole. No country had done that before.

The mission cost about $75 million. Christopher Nolan spent $165 million making Interstellar, a movie about space.

This was not a fluke. ISRO has run frugal missions for decades, and the pattern holds against rival space agencies too.

ISRO did not outspend anyone. It out-designed the constraint. That is jugaad in its most literal form.

What the AI Adoption Numbers Actually Say

Move from rockets to routers, and the same instinct shows up in how India adopts AI.

What This Looks Like From the Enterprise Sales Seat

I sit in enterprise deals and I see a consistent pattern.

Buyers do not get impressed by the fanciest model demo. They ask one question: does this work with what we already have?

Here is what that means if you are the one buying or building:

• Ask what breaks first: the model, the data pipeline or the team’s trust in the output.

• Budget for the unglamorous 80 percent: integration, training, and change management.

• Pilot inside the real constraint, not a clean sandbox environment.

• Treat the first messy workaround as a feature, not a failure to fix later.

The Strongest Case Against This

Jugaad has a real downside, a hacked charger can overheat. Makeshift wiring bypasses safety checks for a reason.

Applied to AI, the same risk shows up. A workaround that skips governance, testing, or explainability is not clever. It is a liability with better marketing.

The market already prices this in. 94 percent of Indian respondents told IBM that explainability matters to their business in 2026.

The Signal in the Noise

India’s advantage was decades of practice solving real problems with incomplete resources and AI just handed that instinct a bigger lever.

The instinct was already there, built into decades of frugal engineering. AI did not create it. It just gave it faster tools.

The interesting decade is not the one where India catches up. It is the one where everyone else catches on.

Somebody built it, on purpose, out of something real.

The Line That Was Real

In 1864, at a Confederate prison camp called Andersonville, guards drew a line inside the stockade wall. Step past it, or even lean too far, and you were shot.

No trial. No warning. The line did the talking.

Nearly 13,000 of the camp’s 45,000 prisoners died before the war ended, from disease, starvation, and exposure as much as gunfire. The commander, Henry Wirz, was tried and executed for war crimes in 1865. The word “dead line” entered the official record at that trial, describing a boundary that was, in places, not even physically marked.

That detail matters. Keep it in mind. We’ll come back to it.

How A Prison Sentence Became Newsroom Slang

Language doesn’t retire a word just because the danger disappears. It just finds the word a new job.

By 1919, American newspaper editors were using “deadline” for something far less lethal: the absolute last moment copy could be sent to print. Same word. Same pressure to not be late.

What Everyone Gets Wrong About Deadlines

Most people assume a deadline works because it’s enforceable. Miss it, something bad happens. That’s the whole mechanism, or so it seems.

But go back to that Andersonville detail. Court testimony from Wirz’s trial described the dead line as, in places, “an imaginary line,” marked only by loose boards on shifting stakes. Guards didn’t need a fence. They needed prisoners to believe the fence was there.

That’s the part almost nobody clocks about deadlines in a corporate setting. Most of them are imaginary too. A VP picks a date. A sales team calls it “quarter close.” Nobody drew a real line, and everyone still feels one.

Ask anyone in enterprise software sales what the last week of a quarter feels like. Buyers who were slow all month suddenly move fast. Approvals that sat for weeks clear in a day. The deadline didn’t get more real. Everyone just agreed to act like it had.

Where This Shows Up On Your Calendar

This isn’t only a sales story. It shows up everywhere work happens.

A few places worth watching for it:

• A “hard deadline” from leadership that keeps moving, and never actually costs anyone anything when it slips.

• A vendor contract that magically closes the day before a quarter ends, every single quarter.

• A project deadline set to create urgency, not because the date itself matters.

• A performance review cycle that treats an arbitrary date as if it were physics.

None of these are lies, exactly. They’re Andersonville’s imaginary line, redrawn for a Tuesday standup. The trick still works because belief was always the real mechanism, not the line itself.

Apple’s Macintosh team lived that for real.
- Chasing their 1984 ship date, engineers wore shirts that read "90 hours a week and loving it."
- The shirt was real. So was the deadline.

By 2030, the WEF projects 170 million new jobs. It also projects 92 million lost ones. Net positive, on paper.

The real split running through the economy isn’t about skill level anymore. It’s about whether the job runs on relationships or on repeatable process.

Two Different Games, Not Two Different Tiers

For 20 years, economists have studied job polarisation. High-skill and low-skill jobs grow. Middle-skill jobs disappear.

Remember, routine, predictable work gets automated first. What survives sits at the top or the bottom of the pay scale.

That framework still holds.

What the 2026 Numbers Actually Show

6 data points sketch the real shape of this shift. Each one is pulled from current, verifiable research.

The jobs the WEF expects to grow

Salespeople made the WEF’s list of fastest-growing frontline roles through 2030.

So did farmworkers, delivery drivers, nurses, and educators.

Every one of those jobs runs on judgment in the moment, not a fixed script.

The jobs shrinking fastest

Clerical and administrative roles lead the decline. Data entry, basic bookkeeping, and repetitive processing work are the first to go. These are tasks that were always describable as a flowchart.

A senior salesperson and a junior salesperson can share the same title. They can even share years of experience.Only one of them may be barbell-safe. It depends on how much of their week is relationship work, not process work.

Title and seniority used to be the proxy for safety. They aren’t anymore.

What To Actually Do About This

This isn’t abstract if you manage people, sell for a living, or run a team.

• Audit your own week. Count hours on scheduling, data entry, and reporting versus hours spent earning trust.

• Learn to direct AI tools instead of competing with them on speed.

• Double down on whatever a client or colleague would notice first if you vanished.

• If you manage a team, protect relationship time from process work.

If your role leans process

Get fluent in the AI tools eating that work before someone hands you the decision. Being the person who configured the automation beats being the person replaced by it.

If your role leans relational

Resist the urge to prove your value through busywork and dashboards. That’s the one currency losing value the fastest.

A Closing Thought From the Sales Floor

I’ve spent my career watching how large organisations actually decide to buy something. It’s rarely purely rational.

Committees stall. Champions lose their nerve. Budgets move for reasons that never make it into the deck.

AI can model all of that. It still can’t sit in the room and read it.

That is why I’m not worried about my own seat at the table. It matters more than any statistic in this piece.

Deals with 4 or more engaged stakeholders win 130% more often than single-threaded ones, according to Gong’s 2025 analysis of that dataset.

A single champion used to be enough to carry a deal through the door. That stopped being true a while ago, and most sales training never caught up.

Buying Committees Got Bigger

In 2015, the average B2B buying group included 5.4 people - Gartner’s research in The Challenger Customer

Gartner found that buyers gather four to five pieces of information independently, before ever speaking to a seller.

Roughly 70 to 80 percent of the buying journey is complete before a rep enters the room, per Forrester and Gartner. Sellers are showing up later, to a bigger room, with less control than before.

The Single Point Of Failure Sitting In Your Pipeline

The champion your whole deal depends on is also your biggest risk.

That person can change jobs. They can lose budget. They can simply lose an internal argument you were never invited to.

Gong’s data backs up the risk without it. Deals worth $50,000 to $250,000 that close typically involve at least 10 stakeholders.

What this means for how you work a deal

A single relationship is fragile by definition. A network of relationships survives personnel changes, reorganizations and one bad quarter.

What Sequencing Actually Looks Like

Multithreading is not about contacting everyone in the account on day one. Order matters as much as coverage.

Gong’s research on executive selling makes this precise. Bringing an executive in too early, before your champion has built internal urgency, drops win rate by roughly 6 percent. Bringing them in around the third meaningful touchpoint lifts it by about 5 percent instead.

• Map the room first: champion, technical evaluator, economic buyer, procurement, legal, executive sponsor

• Give each stakeholder a reason to engage, not a generic pitch repeated six times

• Let your champion open the door, then widen the circle by your third meaningful touchpoint

• Get stakeholders talking to each other directly, since alignment formed together is harder to unwind later

The Signal In The Noise

Deal health is measured by how many people in the account would notice, and object, if the deal quietly disappeared

Where This Goes From Here

You can do everything right with one person and still lose the room. The sellers who last are the ones who treat a buying committee like the actual customer, not an obstacle standing between them and one.

- Finding the right company

- Finding the right person

- Finding out if they even had budget, or if you were about to burn 6 months chasing someone who was never going to buy

Now, right AI can do it and we are looking at AI GTM

AI GTM = Using #AI to run parts of the #GTM motion

Not : “I used ChatGPT to write an email.”

But, a real system that runs the whole sequence on its own:

Research > Personalize > Send > Follow up > Hand off to a rep , with almost no human touching each step

There’s a new job title showing up for the people who build these systems:

#GTMengineer

New way of AI-GTM

- AI already knows they’re ready to buy , before you even open LinkedIn

*Where AI is showing up in the funnel*

- ToFu: Account discovery + Signal detection + List building

- Outreach: Personalized email + LinkedIn sequences at scale

- Qualification: Lead scoring based on real engagement

- MoFu: Meeting scheduling + Objection handling drafts + Call summaries

- Late funnel: Mostly human, Complex negotiation + Multithreading + Trust-building

- Post-sale: Churn prediction + expansion/upsell signals

Remember,

“We bought an AI tool” is different to “we have an AI GTM strategy”

So,

*How can you upskill as #GTMengineer*

1️⃣ Ensure traditional GTM fundamentals are solid ( ICP, funnel stages, CAC, pipeline velocity)

- AI GTM sits on top of this; it doesn’t replace the above

2️⃣ Read practitioner articles —> absorb the vocabulary above

3️⃣ Use #AItools available at your disposal (sales-intelligence tools)

4️⃣ See if you can build #AgenticAI (AI that acts autonomously across steps, not just responds to a prompt)

The gap between “AI-aware” and “AI-illiterate” sellers could become the biggest gap in #EnterpriseSales

AI GTM doesn’t replace enterprise selling :

It front-loads the research, targeting, and qualification so the human conversation starts further along, with better information.

The catch: that number only covers water inside the data centre walls. It excludes:

• Water burned generating the electricity (coal: ~2.2L/kWh, gas: ~1.17L/kWh)

• Water embedded in the concrete, steel, and chips used to build it

• Fossil fuels still supply ~50% of global data centre power

India’s Water Woes

• India: 18% of world population, 4% of world freshwater

• 600 million people (43%) face high-to-extreme water stress - NITI Aayog

• Per-capita water availability: ~1,100 m³/year (near the internationally recognized stress line)

• Delhi, Bengaluru, Chennai, Hyderabad all flagged for groundwater depleting faster than it recharges

3 policy tracks are drawing on this same deficit:

1. Ethanol : fuel security, paid for in groundwater

• 1 litre sugarcane ethanol ≈ 2,860–3,000 litres of water (NITI Aayog roadmap)

• Rice-based ethanol: 10,000+ litres/litre - though a 2026 ISMA-ICAR industry study disputes this, citing 173L/kg for sugarcane (114L with drip irrigation)

• Even the disputing study doesn’t dispute where this is grown: Maharashtra, UP, Karnataka , already stressed belts

• Business impact:

  • Distilleries and sugar mills in these states are the first candidates for state-level groundwater extraction caps

  • Anyone with exposure to this supply chain should be pricing in water-availability risk now, not after a mandate change.

Fix:

• Shift feedstock incentives toward maize/damaged grain

• Tie ethanol procurement price to drip-irrigation adoption (15–25% water saving, proven)

2. Deforestation - breaking the recharge mechanism

• India lost 668,400 hectares of forest annually, 2015–2020 - 2nd highest globally after Brazil (FAO/Global Forest Watch)

• Forests aren’t just carbon stores - canopy and roots slow rainfall enough for it to recharge aquifers instead of running off

• Global research: deforestation’s impact on drinking-water access rivals a 9% climate-driven rainfall decline - same order of magnitude as climate change itself

• Business impact:

  • Agriculture, FMCG and beverage companies sourcing raw materials from deforestation-hit belts (Western Ghats, Northeast states) are absorbing rising input volatility without pricing it as a distinct line-item risk.

Fix:

• Put groundwater recharge value into forest-clearance cost-benefit math

• Prioritize watershed restoration (check dams, contour trenching) over plantation-style afforestation , faster water payoff

3. Data centres : the fastest-growing new claim on water

• India’s data centre capacity: 0.4 GW (2020) → 1.5 GW (2025) → 8-10 GW more by 2030

• $167.5B+ committed (Microsoft, Amazon, Google, Adani) - tax-free through 2047

• One 20 MW facility: 300–400 million litres/year

• Sector total: ~150 billion litres (2025) → 358 billion litres (2030 projected)

• S&P Global: 60-80% of Indian data centres could face high water stress this decade

• Bengaluru, Hyderabad, Chennai - the top expansion cities are also India’s most water-stressed

• Only 5 of 15 state data-centre policies have any sustainability requirement at all

• Business impact:

  • This is now a balance-sheet risk, not just an ESG talking point

  • Water-stressed operations face real downside - cooling downtime, higher insurance premiums,community/regulatory pushback, and potential stranded-asset exposure in exactly the metros hosting the most committed capital.

Fix:

• Mandate treated/non-potable water for cooling

• Make facility-level water disclosure a condition of the tax holiday

The pattern

3 tracks.

• Energy security

• Land use

• Digital infrastructure

Each one is a water-allocation decision made with no explicit water budget attached to it. And each is concentrated in the same handful of stressed basins.

And my favourite sport, reminded me why it’s still true.

- Rishabh Pant took a ₹12 crore pay cut

- Gave up the captaincy.

- And probably took a 40-minute flight back home to Delhi

→ 2025: Lucknow Super Giants bought Pant for a record ₹27 crore.

Makes him captain

→ 2026: Pant steps down. Moves back to Delhi Capitals at ₹15 crore. No armband.

2024 : same owner publicly confronted KL Rahul on the field, on camera

KL later said: “Whatever happened wasn’t the nicest thing. It affected the entire group.”

He left.

Moved to DC.

Now Pant.

Same franchise.

Same story.

This looks like a pattern

1️⃣ It’s ok to choose Peace > Paycheck

- If Pant can, you can also leave toxic environment

- There will always be better opportunities, better companies, better people.

- There’s only one version of you.

2️⃣ Psychological safety = Performance

- KL and Pant are world class performers who couldn’t perform in one specific environment

- That’s not a player problem. That’s a culture problem.

- If you do not feel safe & secure, you will not be able to perform your best (science proves it)

3️⃣ Title means nothing if you can’t express yourself

Pant & KL both had captaincy

Both left it behind - at the same franchise

Remember, best people don’t leave for money.

They leave when they stop feeling trusted

Like Steve Jobs said “It doesn’t make sense to hire smart people and then tell them what to do”

Retention strategy isn’t always salary band.

It’s how safe people feel to speak up, mess up, and show up fully.

Share SSwrites

So, here are my 2 cents with 3 (intersecting & analytical) reasons on why Zuckerberg made Kunal Shah global CEO of WhatsApp

1️⃣ #ZuckerbergVision

- Zuckerberg has been trying to build a SuperApp since years, similar to Wechat of China

- That’s why the feature rollouts: Payments. Business pages. AI features. Channels

- Playbook: Chat → Payments → Commerce → Lending → Super App

Once you get someone to pay through an app --> You get their financial behaviour.

Once you have that, brands will pay anything to reach them there.

That’s the real business. Advertising follows money.

2️⃣ #IndianMarket

- India is WhatsApp’s largest market. 550M+ users

- 21.6 billion UPI transactions in December 2025 (~3X the world’s population)

So,

- Whatsapp wanted a share of that pie and launched payments in 2020

- 6 years later, its UPI market share? Under 1%

- PhonePe & #GooglePay together control 80%+ of India’s UPI volume

*Problem* : Trust

Indians view WhatsApp as messaging app and ‘trust’ with their most personal conversations

But they don’t trust it with their money

Those are two completely different things.

3️⃣ #KunalProofOfAchievement

- 2010: Co-founded Freecharge -> Built it into one of India’s 1st digital payments platforms -> Sold to Snapdeal for $400M in 2015

- 2018: Founded CRED with $1M of his own money

CRED

* Processes 40%+ of India’s credit card bill payments

* 17 million users

* Hits 1st profitable quarter in 2026

* Meta just put in $900M at a $4.5B valuation

Kunal has built 2 #FintechCompanies with 3 common things:

1) Consumer Behaviour

2) Consumer Trust

3) Financial loyalty

So, if Kunal Shah can move the Whatsapp needle even to 5%, we’re talking about one of the biggest fintech pivots in the world inside an app people already use every single day

Share SSwrites

What is a Claude Artifact?

- It’s not a chat reply

- Claude builds a live, working output

- Interactive dashboards, calculators, proposals, one-pagers

- No code. No Figma. No developer.

- I described it → Claude build it → And then you can ‘Publish’ and share it with others.

What I actually built:

Account Research Artifact : I type the name of any prospect company, and it does the rest.

- It scans annual reports, company websites, news and public filings automatically

- Output: revenue, growth trajectory, key people, strategic priorities and likely pain points.

- All structured. All in one place. Ready before you dial.

- No manual Googling. No switching 7 tabs. No 45-minute prep sessions.

I used to spend 3-4 hours before a meeting about the account, prepping for one discovery call.

Now,

this Artifact does it in under 2 minutes.

I also created a 1 Pager on Artifact and how to use it.

and then thought of 5 ways an #AccountExecutive can use Artifacts today:

1) Account Snapshot : 1-page brief on any prospect. Built in 60 seconds before a call

2) ROI Calculator: Plug in prospect numbers live on the call. Show savings instantly.

3) Proposal Doc : Formatted, branded, sent same day as discovery. No waiting.

4) Competitive Matrix : Custom side-by-side vs competitors. Per deal, not generic.

5) Follow-up Kit : Email + agenda + next steps in one Artifact. Zero copy-paste.

Remember,

-> The best AEs don’t just talk , they show.

-> Artifacts let you show value on the call, not after it.

-> Shorter sales cycles. Higher win rates.

-> And you do it without a designer or a dev.

So, here is how you can make your 1st Artifact:

1) Go to claude.ai (or download the applicatio)

2) Type exactly what you want to build

3) Claude renders it live in the sidebar

4) Refine in plain English

5) Done

Seriously. Try it once and if you have any doubt, feel free to reach out to me. Happy to help, whether you personally know me or not know me. And yes, I do not charge money for this :)

Share SSwrites

2 Skills, are now part of how I work:

1️⃣ *Morning news skill* :

- Not the kind for viral content and outrage

- Real one, scanning verified sources, on topics I chose, in the order that matters to me

2️⃣ For my work *Account Research* :

~13 years in corporate teaches you one thing: the sales rep. who walks into a room knowing the account wins

- Revenue, strategic priorities, key people, what their competition is doing, what the annual report says

- Account Executives spend hours doing that research

- Now I type an account name. The *Skill* does the rest from verified resources

Here is the thing about Skills that most people miss:

👉 Skill is not a chatbot prompt. It is a repeatable & saved instruction in Claude

👉 You do not need to be a coder. Describe a workflow clearly & you can build one

👉 Skills can call real tools: web search, your documents, the internet

👉 ROI compounds. Build it once. Use it every morning, every account review, every sprint.

Remember, the possibilities of building Skills is endless to make your job easier & better.

I believe, people who will build their own AI workflows will outpace those who do not.

PS: I am planning to make more repeatable *skills* to buy back my time, for more productive personal & professional work.

The question is not whether to learn this. The question is whether you want to be ahead of the curve or behind it.

Happy to discuss and help out anyone looking to learn.

I write AI related content, to keep you up to speed

Share SSwrites