That is partly because nobody has handed us much else. What companies are handing us is a chatbot inside every app, each one built to keep us inside.

The problem: a walled garden in every app

Many companies add a chat icon in the corner and call it an AI strategy. It waits for you to type, and it works only inside that one app. Ten apps means ten assistants, and none of them knows the others exist. Your fridge does not know your calendar, and your calendar does not know your cab.

Here is the test I would apply to any AI feature:

Did this give the user time back, or take it away?

An assistant that keeps me stuck in one app takes time away. The better idea is to build for the assistant I already carry.

The better way: open your product to my assistant

Every phone now has an assistant, whether Siri, Gemini or something else. Suppose I say, “Hey Siri, book me a cab with Uber from here to the airport.” Today, it may just open the Uber app and leave the rest to me. In future, the app should share what it can do, such as check fares, book and track a ride, through APIs or open standards like MCP. The assistant then does the work, with my permission, and says, “Your cab arrives in four minutes.”

This has already begun. Swiggy lets people order food and groceries through ChatGPT, Claude and Gemini. Android’s AppFunctions and Apple’s App Intents are ways for apps to plug into phone assistants. It is early and uneven, but the direction is clear.

Companies can keep their own in app assistants. The problem is only when that is the sole door. Open the front door as well.

Home and food: the fridge that thinks before you shop

Imagine your refrigerator checks its own contents on the device, without sending photos of your kitchen anywhere. It knows the eggs are finished and the milk expires tomorrow. It asks: “Do you want these delivered, or would you rather go and pick them up?”

If you choose delivery, it builds the basket on Instamart or any app you prefer, and you approve it in one tap. If you prefer to go in person, the assistant becomes a planner:

• With your partner: it checks their calendar, only if they have agreed, and shows just the free slots, not the meeting details.

• With the kids: it adds a kid friendly cafe or park near the store.

• Your choice: you can ask for the nearest organic store or the best rated one, even if farther. Your neighbourhood kirana should show up too, not only the big chains.

The AI does not decide for you. It removes the searching and the back and forth, and leaves the choice with you.

What changes for companies: grocery apps and fridge makers should let assistants check stock, build a basket and place an order, and a fridge should work with any grocery app the user picks.

Health: helpful and humble

Apollo, PharmEasy and MedPlus hold a lot of useful context that mostly goes unused. An AI native version could:

• Refill before you run out: “Your BP tablets have 4 days left. Shall I order now?” It should reorder only what a valid prescription allows.

• Flag interactions: if two doctors prescribe different medicines, it tells you to check with your chemist or doctor before taking both.

• Look after parents from a distance: a family member is told about a missed dose only if the parent has agreed, and it must be easy to switch off.

The rule is simple: AI informs, professionals decide. A wrong grocery order is an annoyance. A wrong medical suggestion can be dangerous. So AI should never diagnose or change doses on its own.

What changes for companies: pharmacies should let assistants read refill status, order and track delivery, with strict consent, prescription checks and the tightest permissions of all.

Getting around

Uber, Ola and Rapido know your routine, yet they still wait for you to open the app and type a destination.

• Daily commute: it pre books your ride and warns you, “It is raining and there is a cricket match near the stadium. Leave 15 minutes earlier.”

• Cab sharing: it can suggest people travelling your route, but only with verified profiles, women only options and clear opt in, within local rules.

• Other cities and countries: tell it “I need to be in Bengaluru next Friday,” and it compares flight, train, bus and cab by door to door time and cost, books the airport cab and adjusts your meetings if the flight is delayed. For a trip abroad, it prepares a checklist for visa, forex and connectivity, with links to official sources.

A morning that works. You say, “Book me a cab tomorrow at 9 AM.” Next morning, you oversleep. Instead of a panic, the assistant rings a louder alarm that you had allowed, sees a meeting the moment you reach office, books a cab for right now and offers to tell the attendees you will be a few minutes late. You just get up and go.

What changes for companies: cab, airline, railway and bus apps should let assistants get fares, book, reschedule, cancel and share live status.

The glue: messages, mail and calendar

Gmail can summarise a long thread and pull out the deadline. WhatsApp can turn a 300 message group chat into three lines: what was decided, what is pending and who owes what. Messages can warn you about scam links and fake bank alerts before you tap, which in India alone would save real money. Calendar can add buffers and reschedule when something changes.

What changes for companies: offer clear, permission based actions such as “read this thread” or “send this reply after I approve,” and let users decide what the assistant can see and do, app by app.

The new deal: first mile and last mile

Here is the model I would like companies to think about.

The assistant does the first mile: finding, comparing and helping the user decide. Your app does the last mile: payment, tracking, loyalty and support.

If your price or service is better, the assistant can say so and take the user straight to a simple pay screen in your app. You do not need to trap anyone. You only need to win the recommendation.

This works only if a few things are true:

• Prices are honest. The comparison shows the final price with all fees.

• The handoff is light. No banners and upsells between the user and the payment step.

• Payment is smooth. Today, orders placed through AI tools on Swiggy support only cash on delivery, so this is the hardest step. A secure handoff to UPI in your own app is a real solution.

The uncomfortable parts

Money. Apps earn more when we browse longer, and comparison can squeeze margins. But trust and loyalty last longer than impulse orders.

New gatekeepers. The OS company can become the new walled garden if it favours its own apps or takes a hidden cut. Users should be able to choose their assistant, and recommendations should be explained.

Privacy and security. These ideas need deep access to data, so keep processing on the device where possible, follow India’s DPDP Act, and ask for confirmation before money, health or sharing data. Every open door is also a target for scammers.

Time and access. Integrating thousands of apps will take years, which is why it is worth starting now. Voice and regional languages are essential for India, and small shops need an easy way to be found without building their own APIs.

A fair counterargument. Creative AI is not always slop, and it lets many people make things they never could. My point is about balance. Everyday usefulness is still deeply underexplored.

A note to companies, builders and OS makers

This is not a demand. It is a hope.

The real keys are with the companies that make our apps, the builders who ship them and the OS makers who decide what every app is allowed to do. I hope you will think along these lines on your own, rather than waiting for a regulator or a competitor to push you.

• Companies, open your products to assistants through APIs or MCP, with proper permissions.

• Compete on the outcome, not on lock in.

• OS makers, be the bridge, not the toll booth. Use open standards and let users choose their assistant.

• Keep recommendations fair, and explain why something is suggested.

• Measure time saved, not time spent.

• Build for Bharat, with voice, regional languages and modest phones.

Here is a small exercise. Pick one product you own or use, list the manual steps for one job, and ask which of them an assistant could do if you opened the door. Then open one this week.

The choice in front of us

We can spend these tools on memes, or on giving people back their time, health and peace of mind. My hope is that companies choose to open the door, not lock it.

🧠 Human driven, AI assisted. The perspectives shared in this newsletter are my own, refined and structured with the help of Large Language Models.

I have been thinking about this pattern for a while, and AI is just the latest example of it. It is not a new story. Same story, new player.

The Original Race: Distribution Over Everything

Go back to the early days of Silicon Valley, and the story is more interesting than just who launched first. Xerox actually built the first proper graphical computer, the Xerox Alto, way back in the 1970s. Apple came next with the Macintosh in 1984, built on some of those same ideas. Microsoft was late to this party. Windows 1.0 launched in 1985, and it was seen as a rough copycat, not a leader. Apple even took Microsoft to court over how similar it looked.

So Windows did not win by being first. It won because of what Microsoft did after being late. Instead of keeping Windows locked to one type of hardware like Apple did with the Mac, Microsoft licensed Windows to every PC maker willing to pay for it, IBM, Compaq, Dell, and dozens of others. That meant Windows ended up running on cheap, mass produced machines everywhere, while Apple stayed limited to its own expensive hardware. Businesses built their software around Windows because it was already everywhere, and once that happened, switching away became too costly for most companies. IBM and Xerox both had strong ideas at that time too, but neither one built that same kind of wide distribution, and both lost ground because of it.

Then there is the Microsoft and Apple story from a bit later, one of my favourite examples. In the 80s and 90s, Apple had good hardware but weak software. Microsoft had no hardware at all, but strong software. So they teamed up. Microsoft software started running inside Apple machines. On paper, these two were rivals. But in reality, both cared more about locking in users through wide reach than fighting each other. Get your product everywhere first, sort out the rest later.

In those days, new features came slowly, maybe once a year or once in six months. Companies needed a lot of skilled engineers, hired fresh from colleges and universities, working for months to build the next version. It was slow compared to today, but the thinking behind it was the same as what we see now, get your product into as many hands as possible, then build everything else around that reach.

Then Facebook Did The Same Thing, Just Faster

Fast forward to 2008. A product called The Facebook exploded the same way PCs did in the 90s, just much faster this time.

I still remember our local newspaper, Eenadu, carrying a half page story about Mark Zuckerberg and this platform. That is how big it felt, even reaching a regional newspaper in India.

What Facebook did was simple. It gave people one single place to share everything about their life. What they like, who they married, what they watched last weekend. It made staying connected effortless, and once people joined, they never really left.

Facebook did not stop at being just a place to connect. It kept adding more and more inside itself. A marketplace to sell things. Subscriptions so people could support their favourite creators. An events page so fans could register for things happening around them. Slowly Facebook stopped being just one app and became a whole world by itself, with everything a user might need packed inside it.

It also went after anything that looked like competition. It bought WhatsApp. It bought Instagram. Both are now such a big part of daily life that people do not even think of alternatives anymore.

Google tried to fight this with Google Plus. It failed. Hangouts and Allo tried to fight WhatsApp. They also failed. Twitter built a decent user base, but nowhere close to what Facebook had.

The lesson is simple. Reach the market first, expand fast, then absorb anything that comes close to competing with you. More users mean more profit, more profit means higher value, and that higher value funds even faster expansion. It just keeps growing on itself.

Now AI Is Running The Same Playbook, Just Faster Still

I want to be clear here, I am not against AI. I think it is one of the most useful things to happen to how we work in a long time. But it is worth pointing out clearly what is happening at the industry level.

Tools like Claude, Codex, and Gemini are giving big tech companies the fastest speed to market they have ever had. What used to take a year to build can now be built in weeks. This is a real and useful jump. But it is also creating pressure. Companies that move slower are watching others move faster, and the fear of missing out is real, even for big companies with a lot of money.

At the same time, companies are automating a lot of work that earlier needed big teams, and getting the same or better results with fewer people. This is genuinely happening, not just hype.

Giving The Skeptics Their Due

This kind of change never comes without people questioning it. Windows had its critics. Facebook had its doubters too. AI has its own set of doubts today, and it is only fair to look at them honestly instead of brushing them aside.

AI does make mistakes. It gets things confidently wrong sometimes. The security around these systems is still being worked on, and running these large models is genuinely expensive. These are not excuses made up by people who dislike AI. These are real problems that engineers are actively trying to fix right now.

At the same time, most people who raise these concerns also quietly know that a large part of repetitive, manual work can now be done reliably by AI. The problems are real, but they are the kind of problems that keep shrinking with time, not the kind that stop the whole shift. Early computers from IBM had flaws too. The first version of Windows was clunky. Facebook had privacy troubles from its very first year. None of that stopped the pattern from playing out. AI’s current problems will not stop it either.

The Race Between AI Labs And The Consulting World

Every big AI company, OpenAI, Anthropic, Google DeepMind, xAI, wants to be seen as the leader in this space, in terms of money, users, and influence. At the same time, the big IT consulting firms, TCS, Wipro, Accenture, Persistent Systems etc want to be the ones who bring AI into every company and earn from doing that work.

Neither side wants to lose out, so instead of fighting each other, they are teaming up. AI companies are signing deals with big enterprise and financial firms to get their tools deployed quickly, and the consulting firms get to stay relevant as the ones who actually set it up. It is the same Microsoft and Apple story again, just wearing a 2026 outfit.

Look at how AI companies build their products today. Claude, for example, now has plugins that do the same job as products already sold by other companies with their own decent user base. The difference is scale. A big AI company that already has millions of users can add these features inside its own app, and use its scale and AI tools to build them faster than smaller standalone companies can. Not every one of these will succeed, but the direction is clear, big platforms are aiming to become more complete, all in one products.

But Let Us Push Back On The “SaaSpocalypse” Idea also

There is an argument going around that AI coding tools mean one person, or a small team of ten people, can now build a real competitor to Salesforce or Microsoft. I do not think that holds up, and here is why.

Building something like Salesforce is not just about writing code. It needs huge compute power, dedicated security teams, proper UX research, compliance across many countries, and support teams that only make sense with hundreds, sometimes over a thousand people. AI tools do not remove that need. They only make the coding part faster.

Where AI actually helps a lot is at a smaller scale. A solo person or a tiny team can now build a mobile app, a browser extension, or a simple website in a fraction of the time it took before. Design work is mostly handled by AI now, and coding is much faster too. Security, integrations, and scaling still need real skill, but even that gap is shrinking every month. I have already seen reports of people running fully automated pipelines to take an idea from start to launch with almost no manual work.

So both things are true together. AI tools do help a new player build something fast and even find some early users. But speed of building is only one part of competing at scale. Big tech has money, compute, distribution, and an existing user base, advantages that take years to build and that a new entrant cannot recreate just by building fast. So when a small team launches something useful using AI, big tech is usually well placed to build something similar using the same tools, or to bring the idea in through a partnership or acquisition, the same way Facebook did with WhatsApp and Instagram. The new entrant gets a genuine head start, but big tech has the resources to catch up quickly if the idea proves itself. That is really the point I was making with the Salesforce example. Small players can absolutely build things faster than ever now, but scale and resources still end up deciding who leads the market.

The Bigger Pattern, Everything Slowly Becoming One

This is why the overall direction tends to stay the same, whether a market is being built by a fresh startup, a small AI powered team, or a giant like Facebook or Microsoft. Big tech generally has the resources to compete for the long run and, in many cases, to bring in or absorb whoever comes up with something new. It happened with search, with social media, with messaging apps, and it is now beginning to happen with the small tools AI has made possible. The players change, but the pattern tends to repeat.

The same pattern is playing out well beyond social media and AI, quietly, for years now.

Local clothing shops, kirana stores, hardware shops, absorbed into Amazon and Flipkart. Coaching centres, tuition classes, colleges, absorbed into Udemy, Coursera, LinkedIn Learning, and now AI tools. Friends meeting in person, community gatherings, absorbed into Google Meet and Zoom calls.

None of these changes got the kind of attention AI is getting today, even though the same thing is happening, offline life slowly moving into a few big digital platforms. AI is not something new in that sense. It is just the latest and loudest chapter in something that has been happening for a long time.

Put all these examples together, Windows, Facebook, WhatsApp, Instagram, kirana stores, coaching centres, and now AI, and the direction is clear. Whoever gets to scale first ends up winning. And winning never means stopping. It means absorbing whatever comes next, until entire parts of daily life live inside one place.

AI is simply the fastest version of this pattern the industry has ever seen. The same handful of companies racing to launch first are also the ones building the infrastructure, the compute, the models, that everything else will eventually depend on. What used to take a whole decade to consolidate is now happening in just a few years.

The small tools and apps people use individually today for work, learning, and daily decisions are the next layer this pattern is moving into. Just like kirana stores folded into Amazon and coaching centres folded into Udemy, these small individual tools will likely fold into a handful of big AI platforms. Not because those platforms are always better, but because that is exactly what has happened at every stage of this industry so far, and there is no clear reason this stage will end any differently.

That is the real pattern here. It started with local shops and coaching centres. It moved through operating systems and social media. It is now reaching the small tools sitting on everyone’s phone and laptop, and this time, it is moving faster than any wave before it. Every era has had its own version of this story, and every era has ended the same way, with everything eventually becoming one platform.

🧠 Human driven, AI assisted. The perspectives shared in this newsletter are my own, refined and structured with the help of Large Language Models.

Now here is the part that should make Google uncomfortable. Every single one of those five tools already has an equivalent sitting inside Google’s own product lineup. Google Sites. Google Pay. YouTube. Blogger. Google Contacts. Google Analytics. Play Store. Gemini. All owned by the same company, all built on the same infrastructure, and all currently operating as unconnected islands instead of one unified path for creators.

This article makes one argument. Google does not need to invent anything new to compete with and eventually overtake Substack. It simply needs to connect what it already owns.

The Core Thesis

Substack succeeded by giving independent writers one simple promise. Publish for free, pay only when you earn. That promise built a business worth talking about, but it also left enormous gaps that Google is uniquely positioned to fill, provided it stops treating its products as separate departments and starts treating them as one creator ecosystem.

No competitor, including Substack itself, can match the breadth Google already owns. The question is not whether Google has the pieces. The question is whether Google will ever bother assembling them.

Substack’s Real Pain Points

Before arguing what Google should do, it is worth being precise about what Substack actually costs its creators today.

The fee stack compounds against you as you grow. Substack takes 10 percent of subscription revenue. On top of that, Stripe charges roughly 2.9 percent plus 30 cents per transaction, along with a small recurring billing fee. Add it all up and creators are typically losing 13 to 16 percent of their gross revenue, with no volume discount as they scale. In fact, the better a creator performs, the more expensive the platform becomes, since there is no loyalty tier or reduced rate for high earners.

The real money adds up fast. A creator earning around 5,000 dollars a month in subscription revenue is paying close to 650 dollars a month in combined fees. A creator who crosses 1,000 paid subscribers can lose roughly 6,000 dollars a year to fees alone. That is not a rounding error. That is a second rent payment, gone every single month, forever, as long as the creator stays on the platform.

There is no room to sell a course. Substack only supports subscriptions. A creator cannot sell a onetime course, a template pack, or a digital product natively on the platform. This forces creators to bolt on third party tools just to diversify income, which fragments the very audience relationship Substack was supposed to simplify.

Trust has taken hits over the years. Substack has faced public criticism over its content moderation policies and has also dealt with security incidents involving user data. None of this is fatal on its own, but it adds to a growing list of reasons creators are willing to consider alternatives if a credible one appears.

Analytics, contact import, and native promotion remain weak. Substack gives creators a subscriber list and some basic metrics, but nothing close to the depth of a proper analytics suite. There is no built in way to pull in an existing contact list from another ecosystem, and there is no native paid promotion engine to help a new newsletter get discovered beyond Substack’s own recommendation network.

The Google Ecosystem Play

Here is where the opportunity becomes obvious. Imagine the same five gaps, but each one filled by a Google product the creator is often already using.

The developer. He hosts his Android app on Play Store, builds a companion site on Google Sites under his own custom domain, sells an advanced development course on YouTube, and links a newsletter on the same site to notify buyers of new releases. One login. One audience. One dashboard.

The photographer. Her portfolio lives inside Google Photos, directly embedded on her Sites page with zero third party hosting required. She sells editing masterclasses on YouTube and print orders through the same connected storefront.

The culinary expert. A recipe blog on Sites or Blogger, cooking classes sold through YouTube, and downloadable meal plans distributed through the same payment gateway that powers the newsletter subscriptions.

The musician. Production and mixing courses on YouTube, a fan newsletter and merchandise store on Sites, all branded under one custom domain purchased through Google itself.

The fitness coach and the language tutor. Workout programs or lesson packages sold through YouTube, progress tracking or practice material delivered via newsletter, and client bookings handled through the same integrated site.

The pattern repeats across every vertical. One login instead of five. One payment gateway instead of a patchwork of processors. One analytics dashboard instead of stitching together data from three different tools. One audience that Google already understands better than anyone else on the internet.

The Infrastructure Advantage

Gemini can make rapid site deployment realistic for non technical creators. Instead of spending days on layout and SEO structure, a creator could describe what they want and have a functional, search optimised site ready in minutes.

Google Pay and Play Billing already run at massive scale. While no official fee number exists for this kind of bundled creator product, it is reasonable to argue that payment rails Google already operates for millions of transactions daily make a lower blended fee plausible, especially when compared against Substack’s stacked cost of 13 to 16 percent.

Domain selling folded into the same flow means a creator can register, connect, and brand their site without ever leaving the Google ecosystem or dealing with a third party registrar.

Contacts, Analytics, and Search ads close the final loop. Instant list building from an existing Google Contacts account, unified tracking across the site, the newsletter, and the YouTube channel, and native paid promotion directly through Google Search, without needing a separate ad platform at all.

Addressing the Skeptic

Any argument this ambitious deserves to be challenged honestly, so let us do that.

“Google kills products, why would this be any different.” It is true that Google has a well known habit of shutting down beloved products, and Google Plus, Google Reader, and Currents are the usual examples people bring up. But look closely at Blogger and Google Sites specifically. Both have survived for well over two decades, largely without major investment, and are still running today with no shutdown announcement in sight. Over 353,000 websites still run on Blogger according to independent tracking data. The honest reading of that survival is not neglect, it is proof that a product does not need constant new features to stay alive if it already works and has a real user base depending on it. Now layer on the argument that matters most here. This proposed integration ties Sites and Blogger directly to YouTube and Play revenue. That gives Google an actual financial reason to protect and maintain the product, unlike Google Plus or Reader, which never had a direct revenue relationship tying them to the rest of the business.

The August 2026 Blogger suspension incident. In early August 2026, an automated security system on Blogger wrongly flagged a number of genuine blogs as violating Google’s malware policy. Most were later restored, though the episode did shake some creators’ confidence in the platform. This is not a sign that Google plans to shut Blogger down. It is a sign that its moderation systems need better oversight. That is exactly why revenue linked integration matters. Once Blogger and Sites are tied to YouTube and Play earnings, Google has real financial reason to fix such issues quickly, the way it already protects Search and Ads.

The antitrust question. Bundling this much of the creator economy into a single company’s ecosystem is not a small ask, and it would understandably invite regulatory scrutiny given Google’s existing market position in search and advertising. This piece is not the place to resolve that debate, but it would be dishonest to leave it out entirely. Any real version of this plan would need to be built with that scrutiny in mind from day one.

The Real Question

Every piece Google needs is already sitting inside its own portfolio. Sites, Blogger, Gemini, YouTube, Analytics, Search, Contacts, Photos, and Play are not five different products waiting for five different roadmaps. They are one obvious ecosystem waiting for someone inside Google to connect the dots.

So the real question is not whether Google can beat Substack. On paper, it clearly can, and probably in a single coordinated move rather than a slow multi year rollout. The real question is whether Google, a company that has historically treated its own products as isolated silos rather than as one connected platform, will ever choose to build it.

If it does, Substack will have a serious problem on its hands. If it does not, this remains one of the more interesting blind spots in the entire creator economy today, sitting in plain sight, fully assembled, and completely unused.

That’s happened to me more than once - pulled in purely on the strength of skills and endorsements sitting on my profile. It’s a small, easy-to-miss data point, but it tells you something important: even a system as flimsy as LinkedIn’s skill endorsements, where anyone can vouch for anyone with a single click, is already good enough to change what recruiters do. It already moves real decisions.

Which raises an obvious question. If an unverified, gameable signal is doing that much work - what happens when the signal is actually true?

The verification LinkedIn already does and where it stops

LinkedIn has spent the last few years quietly building real verification infrastructure. You can confirm you’re a real person by submitting a government ID through a partner like CLEAR or Persona. You can confirm you currently work somewhere by verifying a company email address. More than 100 million members have added at least one of these verifications, and LinkedIn has extended the same trust signals to job postings and company pages.

That’s a meaningful foundation. But look closely at what it actually proves: you exist, and you currently have an email address at Company X. That’s it. It says nothing about whether the ten years of job titles above that current role are real. It says nothing about whether the certification listed under your name was ever issued, whether the patent you claim was ever filed, or whether the people who managed you would say you were any good.

LinkedIn verifies identity. It doesn’t verify careers. And for a platform whose entire value proposition is “this is where your career lives,” that’s a strange place to stop.

What a real verification layer would look like

Here’s the proposal, in concrete terms - not “wouldn’t it be nice,” but an actual mechanism.

Certifications and badges don’t need to be invented from scratch. Digital credentialing infrastructure already exists - platforms like Credly, and the growing number of institutions and companies issuing their own verifiable digital badges. LinkedIn wouldn’t need to build a new credentialing standard; it would need to plug into the ones that already exist and are already digitized.

Employment history and peer endorsement is the harder problem, and it’s where the current system’s biggest weakness - and the fix - both live. LinkedIn already tried structured endorsements once, in the form of skill endorsements, and they became noise: anyone can endorse anyone for anything, with zero cost and zero accountability. The fix isn’t to remove endorsements. It’s to gate them.

Picture this: you worked at Company A for three years, then moved to Company B. Under a real verification layer, the only people eligible to endorse your time at Company A are people who were themselves verified employees of Company A, during the same window you were there - verified the same way LinkedIn already verifies current employment, through a company email address. Not “anyone who knows you.” Not “anyone who clicks a button.” Only people who can prove they were actually in the building with you, at the time.

That single design choice turns endorsements from a popularity contest into something closer to a reference check that happens to be structured, searchable, and built into the platform you’re already on.

Patents and published research are, if anything, the easiest piece - they already sit in searchable public databases at patent offices and journals. The hard part was never finding them. It was connecting them to a person’s profile in a way anyone could trust.

Consent governs all of it. This isn’t an always-on data pipe the way a traditional background-check vendor works, where a company pulls your records once you’ve signed a blanket consent form. It’s opt-in and event-triggered: when you move to a new company and that company kicks off a verification process, you switch on data-sharing for that event. Nothing is exposed by default, and nothing is exposed continuously. You control when the API can see anything at all.

And LinkedIn’s incentive to build this isn’t goodwill - it’s revenue. A cut of transaction value, or a flat fee per verification, turned into a licensable API that companies could use instead of, or alongside, traditional background-check vendors. This isn’t a feature. It’s a product line.

Why now, and not five years ago

This idea isn’t new to me, and it isn’t new to this post. I pitched an early version of it internally at Microsoft during a 2023 hackathon. I reached out to LinkedIn’s leadership directly, laying the idea out. I never heard back from either.

At the time, I don’t think the pitch was wrong so much as early. A huge amount of what makes this feasible - employment records, certifications, government-issued documents - depends on those records existing in digitized, cross-referenceable form somewhere. In 2023, that infrastructure was patchy. It isn’t anymore. Governments at national and local levels have spent the last few years aggressively rolling out digital records for nearly everything. The raw material this proposal depends on now exists in a way it simply didn’t a few years ago.

The market has also caught up to the urgency of the problem. Multiple industry analysts now project that a striking share of job applicants worldwide - some estimates put it near one in four by 2028 - could be entirely or partially fraudulent, driven in large part by how easy AI has made it to generate a convincing resume, or an entire convincing profile, out of nothing. Recent employer surveys back this up from the other direction: the overwhelming majority of companies now run some form of background check before hiring, and a meaningful share of them have had to walk back a job offer after discovering the candidate’s resume didn’t hold up. And this isn’t an abstract, far-off risk - it’s already visible in ordinary reporting out of markets like the U.S. and India, where stories of people running multiple LinkedIn profiles with slightly altered details, purely to game recruiters and hiring processes, are common enough to be unremarkable.

Employers already spend billions of dollars a year trying to solve this problem after the fact, through background-check vendors who verify records that a platform like LinkedIn is, in theory, better positioned to verify at the source.

The part almost nobody’s proposal covers: freelancers

Every version of this idea I’ve seen - and every version I’ve pitched before this one - stops at corporate employees. That’s a mistake, and it’s the part of this proposal I care about most.

A corporate employee has an HR department quietly vouching for them by the simple fact of a paycheck and a company email address. A freelancer or gig worker has none of that. They might do excellent work for a dozen clients and have almost nothing to show for it beyond scattered testimonials, screenshots, and word of mouth. There’s no equivalent of a company email address proving they did the work, and did it well.

Platforms like Topmate already let independent professionals collect structured client feedback. Imagine that feedback flowing directly into the same verified-credibility layer described above - not as a review buried on a third-party site, but as a structured, verified credential sitting on the same profile a recruiter or client already checks. A freelancer builds a track record the same way a corporate employee builds a resume, except the record is portable, verifiable, and theirs, not locked inside whichever platform happened to host the gig.

That’s the piece that turns this from “LinkedIn adds a better background check” into something closer to trust infrastructure for the entire working world - corporate and independent alike.

The obvious objections, answered

“Endorsements already failed once - why would this be different?” Because the failure mode was specific: zero-cost, unverifiable, one-click endorsements from anyone. Gate endorsements behind verified shared employment, and the entire failure mode disappears. It stops being a popularity contest and starts being closer to a structured reference check.

“This creates a much bigger privacy problem.” It’s a legitimate concern, and it deserves a firm answer, not a hand-wave: LinkedIn would need to comply with local data regulations in every market it operates in, without exception. Paired with opt-in, per-event consent - nothing shared by default, nothing shared continuously - this is actually a more privacy-respecting model than how most background-check flows work today, where broad consent is granted once and records get pulled indefinitely afterward.

“Why would LinkedIn bother? What’s in it for them?” A monetizable API, licensed to companies who currently pay third-party background-check vendors for information LinkedIn is structurally better positioned to verify at the source. This isn’t a trust-and-safety cost center. It’s a new product line sitting on top of infrastructure LinkedIn has already built most of.

“Won’t incumbents like Checkr and HireRight just compete this away?” Maybe. But none of them start with 1 billion+ existing professional profiles, and none of them are the platform where the verification claims already live. LinkedIn isn’t entering this market from zero - it’s the only player who could build this without first convincing anyone to create a profile.

Where this goes if LinkedIn doesn’t build it

I don’t think this idea is complicated because the technology is hard. Certifications, employment records, patents, and research are all things that already exist in verifiable form somewhere. The hard part has only ever been who’s willing to do the unglamorous work of connecting them to a real person’s profile, and who has enough existing trust and reach to make that connection matter.

LinkedIn is the only company positioned to do this without asking a single professional to sign up for a new platform first. That’s not a small advantage - it’s the entire advantage. If LinkedIn doesn’t build this, someone eventually will, and they’ll spend years just trying to get people to create yet another profile before they can even start solving the actual problem.

I pitched this once, inside a hackathon, and once, directly to the people who could build it. Neither went anywhere. I still think it’s right. And the pieces required to build it - digitized government records, existing credentialing infrastructure, a workforce more independent than it’s ever been — are more in place today than they were the last time I said any of this out loud.

Somebody’s going to build the trust layer the professional world doesn’t currently have. I’d just rather it were the platform that already has everyone’s career sitting on it.

🧠 Human driven, AI assisted. The perspectives shared in this newsletter are my own, refined and structured with the help of Large Language Models.

Share

That’s when it struck me - why is this person sending people off LinkedIn to prove they are worth paying for? LinkedIn already knows where this person worked, for how long, and what their actual managers said about them. Topmate doesn’t know any of that. It just trusts whatever the person types into their bio.

This one small observation is basically the whole article. LinkedIn has the most credible professional data in the world, sitting completely unused, while a scrappy startup is building a business on data LinkedIn already owns. Let me explain properly.

I say this as someone who has actually worked at Microsoft, LinkedIn’s parent company, so I am not writing this purely as an outsider throwing stones. I have some idea of how these organisations think and prioritise, and that is exactly why this gap frustrates me more.

What Topmate Actually Does Well

If you haven’t come across it, Topmate is a platform where professionals, coaches and creators can offer paid 1:1 calls, webinars, courses, and digital products - all from a single profile link. You set your price, people book a slot, you get paid, done. No back and forth on DMs, no chasing payments.

The numbers are genuinely impressive for a five-year-old, bootstrapped-feeling company. Topmate says it is trusted by over 300,000 creators, and it charges a fairly reasonable commission - around 10% on course sales, letting creators keep the rest. It was founded in 2021 by Ankit Agarwal and Dinesh Singh, and has raised close to $1.13 million in funding so far, which by startup standards is quite lean for the scale it has reached. It sits at a healthy 4.1 out of 5 rating on Trustpilot from over a hundred reviews.

So credit where it’s due - Topmate solved a real UX problem. It made it dead simple for anyone to package their “expertise” and sell it. The trouble is in that same sentence: anyone can package anything as expertise. And that is where my real problem with the platform begins.

The Part Nobody Talks About: Topmate Is an Open Floor for Unverified Claims

Here is my honest, slightly unpopular opinion - Topmate, in its current form, is a platform where anybody can claim proficiency in anything, and if you want to “learn” from them, you pay for it. Most people using it are genuine. But a good chunk are simply exploiting the format.

The example I keep seeing, and it genuinely irritates me: someone gets into a decent IT company through campus placement, works there for a year or two, and suddenly starts selling courses on “how to crack FAANG interviews,” “Data Structures and Algorithms mastery,” or “resume secrets recruiters don’t want you to know.” One or two years of experience, no proof of actually clearing a FAANG interview themselves in many cases, and yet they are charging freshers and juniors for “guidance.”

This is not guidance. This is somebody with barely more experience than the person paying them, monetising a title on their profile rather than actual expertise. And the platform has no real way to stop it, because there is nothing to verify against. A bio is not proof. A profile photo in a blazer is not proof. Testimonials can be requested from friends. None of this is checkable.

I am not saying everyone on Topmate is like this - plenty of genuine, experienced professionals use it well. But the ratio of noise to signal is growing, and in the age of AI-generated content flooding every platform (what people are now calling “AI slop”), we honestly cannot afford more unverifiable claims of expertise floating around unchecked. We need less noise, not more of it.

What LinkedIn Already Has, and Topmate Has to Fake

Now compare this to what LinkedIn already sits on, unused:

• Verified employment history - companies confirm employees, not just self-reported job titles

• Real, mutual, and second-degree connections - you can literally see if this “expert” is connected to people you trust

• Recommendations written by actual former managers and teammates - attributable, dated, checkable

• Skill endorsements tied to real working relationships, not anonymous upvotes

This is not a small advantage. This is the advantage. Topmate’s trust signals can be gamed by anyone with a laptop and a bit of marketing sense. LinkedIn’s trust signals mostly cannot be faked so easily, because they are tied to real employment relationships and real people who put their own name behind a recommendation.

If a career coach on LinkedIn has a genuine recommendation from an ex-manager at a company they actually worked at for four years, that is worth ten times more than a five-star review from an anonymous handle on a booking platform. Anyone availing services deserves that level of assurance, especially when they are paying real money for career advice.

The Current UX Failure: A Feature Nobody Knows Exists

LinkedIn does have a “Services” section on profiles. Have you ever clicked on it deliberately? Most people haven’t, because:

1. It’s buried - tucked away below the About section, easy to miss entirely

2. There is no discovery layer - you cannot search “career coaches who worked at Google” or “PMs my connections have recommended”

3. There’s no native booking or payment flow tied to it - worst case, it just gives you a way to message the person

4. It’s not surfaced in feed or search results the way job posts or content is

Compare this to Topmate, where the entire product is built around discovery and booking. LinkedIn’s version is an afterthought bolted onto a resume page, while Topmate’s version is the entire product.

My Fix: What LinkedIn Should Actually Build

This is not a “LinkedIn should copy Topmate” argument. It is simpler than that - LinkedIn should stop sitting on data it already has and build the layer on top of it.

1. Turn Services into a real discovery surface. Let people search and filter by verified experience, mutual connections, and recommendation quality - not by who paid for visibility.

2. Rank by authenticity signals, not marketing budget. A coach with three verified manager recommendations should outrank a coach with none, regardless of how much they’ve spent promoting themselves.

3. Native booking and payments, so people don’t have to leave LinkedIn (and lose all that verified context) to actually pay someone.

4. Let recommendations function like structured reviews - attributable, dated, tied to a real working relationship, visible right where someone is deciding to pay.

5. Surface it in search and feed, not just as a static tab nobody visits.

6. Pull in certifications directly, with verification. Not just a badge someone uploaded, but something tied back to the issuing body, so a “Certified Scrum Master” or a “Google Data Analytics” certificate actually means something checkable.

7. Attach real project details. Let people showcase specific projects they’ve worked on, with the option to tag the actual collaborators - teammates, clients, cross-functional partners - who can then confirm the contribution themselves. This turns “I led a migration project” from a claim into something with names attached.

8. Let tagged collaborators leave contextual feedback, not just a generic recommendation on the profile as a whole, but feedback tied to that specific project or piece of work. Far more useful for someone deciding whether to book a session with this person.

9. Link out to research papers, publications, or public work. For anyone offering guidance in a technical or academic area, a link to an actual published paper or a patent is worth more than any testimonial.

None of this requires new data collection. LinkedIn already has every ingredient. It just needs to be assembled and pushed forward as a priority, not left to rot as a forgotten tab.

A fair caveat here - not everyone has the luxury of glowing recommendations from managers, colleagues, or employers. Some people work in toxic environments where nobody bothers to write one. Some are early in their career and simply haven’t had the chance to build that kind of documented goodwill yet. Some switched fields and their old recommendations don’t even apply anymore. This is a genuine gap, and I won’t pretend the fix I’m suggesting solves it completely.

But this is exactly why widening the net - certifications, tagged project work, research links, peer feedback on specific work rather than just formal manager recommendations - helps close that gap, at least partially. A freelancer who never had a “manager” in the traditional sense can still show a verified certification, a tagged project with a client who confirms the work, or a published paper. It’s not a perfect solution, but it gives far more people a legitimate way to prove themselves than relying on manager recommendations alone.

Why This Needs to Sit at the Top Surface Level, Right Now

This isn’t a “nice feature for later.” It needs to move from a buried tab to the top of the profile, and the timing is exactly why.

Right now, anyone can generate a polished bio, a confident course description, even fake “results,” using AI in a few minutes. This “AI slop” problem isn’t just about images and lazy posts anymore. It has crept into how people package themselves as experts too. A well-written paragraph proves nothing today. Anyone can sound credible. That’s the real problem.

In a world like this, verified signals aren’t a nice-to-have. They become the only real way to tell a genuine expert apart from someone who just prompted a chatbot for a convincing bio. LinkedIn already has these signals - employment history, recommendations, certifications, tagged project work. They’re just buried where nobody looks.

LinkedIn has over 1.2 billion members worldwide, with India alone at roughly 150-180 million members, the second largest market after the US. That scale is already sitting there, unused. Topmate built its 300,000-creator base from zero, mostly on self-reported trust. LinkedIn doesn’t need to build anything from zero. It just needs to stop hiding what it already has.

So this isn’t optional polish. It’s the single most useful thing LinkedIn could put at the top of a profile, right where people are already looking.

To Be Fair: The Counterpoints

I don’t want to make this sound one-sided, because it genuinely isn’t that simple.

• Verification cuts both ways. The moment LinkedIn starts enabling paid services at scale, it also takes on the responsibility of policing who is “qualified” to charge for what. That’s a moderation headache Topmate largely avoids by staying hands-off.

• Org culture matters. LinkedIn, under Microsoft, has historically been B2B and enterprise-first in its priorities, not creator-first. This isn’t purely a UX oversight - it may reflect a genuine mismatch in what the organisation is built to prioritise.

• Risk of feeling “salesy.” LinkedIn already gets criticised for turning into a feed of humble-brags and engagement bait. Push monetisation too hard, and the platform could tip further into feeling like a marketplace rather than a professional network.

These are real risks. But they are solvable problems, not reasons to avoid the opportunity altogether.

Closing Thought

Topmate built trust from scratch, using self-reported bios and testimonials, and it did a genuinely good job of it. But LinkedIn is sitting on trust it already earned - verified employment, real recommendations, real mutual connections - and simply refusing to spend it.

In an era where AI can generate a polished bio, a glowing testimonial, and a “proven track record” in about thirty seconds, the ability to say “this person’s experience is actually verified“ is not a nice-to-have anymore. It is the single most valuable thing a professional platform can offer. LinkedIn already has it, sitting quietly in the background. It just needs to bring it to the surface, before more people are pulled toward platforms that let anyone claim anything.

Sources referenced: Topmate, Topmate Features, Topmate Funding — Tracxn, Topmate Trustpilot Reviews, LinkedIn Statistics 2026 — Charle Agency, LinkedIn Membership Numbers — John Espirian

Share

🧠 Human driven, AI assisted. The perspectives shared in this newsletter are my own, refined and structured with the help of Large Language Models.

OMR sheets - where you shade circles and hand the paper to an examiner - came in simply as a faster way to process large volumes of answers. Then came fully computer-based exams, where candidates picked answers on a screen and saw scores instantly. That felt like a real leap forward.

As digital exams matured, a whole industry of Learning Management Systems was born - Workday Learning, SAP SuccessFactors, Cornerstone OnDemand, LinkedIn Learning, Docebo, TalentLMS - all promising to train and certify employees at scale. The global e-learning market today is valued at over $400 billion. Certifications became the new currency of professional credibility.

But at the core of every single one of these platforms sits the same thing - a fixed question bank curated by domain experts. The first time a candidate sits for the exam, it genuinely tests their thinking. That is the whole point, right?

Well, not for long.

The System Got Rigged - And Everyone Knows It

It did not take long before people started sharing questions with friends and colleagues. Soon, full question dumps were available online. People started selling these banks, and a cottage industry grew around “cracking” certifications. Candidates earned shiny badges and PDF certificates - kept their employers satisfied - without ever truly mastering the subject.

And then there are those who simply clicked “Next… Next… Next…” and finished a 30-minute course in under 30 seconds.

The deeper problem is this - if a candidate fails 5 important questions across 5 topics but answers the rest correctly, the system cheerfully says “Congratulations, you passed with 80%!” and hands over a sub-128 KB PDF with their name on it. That certificate says nothing about what they actually know. It is, in plain terms, a meaningless piece of digital paper.

What the Fix Actually Looks Like

This is completely fixable. And whoever fixes it first is going to own this entire market.

Here is what a reimagined, AI-native exam platform could look like. The candidate logs in, selects their topics upfront, and chooses their exam duration - 30, 60, or 120 minutes. Then an LLM-powered engine takes over.

Think of it like stress-testing in software development. The AI starts at a base difficulty and keeps pushing harder with every correct answer - mapping the full contour of a candidate’s knowledge, finding where their confidence holds and where it breaks. A candidate who answers 40 questions at difficulty level 7 over 45 minutes, getting 34 right, is telling you something far more meaningful than someone who scored 80% on a recycled question bank.

The final score would reflect that - something like difficulty level × correct answers × time efficiency. And the certificate would not just say “passed.” It would say: “This person answered X questions correctly across Y difficulty levels over Z minutes.” That is a credential actually worth something.

“Just Add More Questions” Is Not the Answer

Some will argue we just need a bigger question bank with harder questions. That is old thinking. We need systems that are AI from the ground up - systems that genuinely cannot be gamed because the questions are never the same twice and the path the exam takes is shaped entirely by the candidate’s own responses.

AI will make already intelligent people even more productive. In the same spirit, a truly AI-powered exam platform will help genuinely knowledgeable people rise above the noise and stand where they actually deserve to stand.

One valid concern - not every candidate is the same. Some experience test anxiety, others have accessibility needs. Any serious platform will need adaptive pacing and accessibility-first design built in from day one. That is not a reason to keep the broken system alive. It is a reason to build the new one more thoughtfully.

The Market Is Still Very Much Alive

Some argue certifications will become irrelevant once AGI arrives. But the numbers tell a different story - the volume of people taking professional tests has grown consistently year on year. Learning and assessment are not going anywhere.

The one who builds this AI-native testing platform before AGI arrives captures the whole market.

Will Microsoft Miss This Bus Too?

One name worth watching is Microsoft Viva, sitting right inside Microsoft Teams with over 320 million monthly active users and deep enterprise reach. The distribution advantage is already there. The question is whether they move boldly - or settle for bolting AI features onto the same old question-bank model.

Microsoft missed search. Missed mobile. Was late to enterprise AI. This might be its most consequential chance yet to get ahead of the curve - rather than spending billions catching up after someone else has already rewritten the rules.

How We Got Here

Most people working in IT today - especially in India - started somewhere around early 2000s. Those days everything was paper-based. Print your resume, carry it to the venue, hand it over. Simple.

Then job portals came. Then big enterprise platforms like iCIMS Workday and SAP SuccessFactors arrived and HR teams got very excited. Big deals were signed. And yes, for the HR side it was useful. For the person actually applying? Same headache, just now on a screen.

Worst part is - these platforms were so painful to update that companies just... stopped updating them. Small changes used to require hours of downtime. So vendors left things as they were. And slowly these systems became fossils.

What Is Actually Happening Right Now

Three openings at the same company. You apply for all three. You fill the same form three times. Same fields, same uploads, same phone number typed out again and again. No shared profile, no multi-role option. Nothing.

And then you reach a field that does not apply to you. There is a “Not Applicable” dropdown. You select it. The form still does not move forward. Because you also have to type something in the text box next to it. Otherwise the system will not let you proceed.

Qualcomm and a few others have figured out better flows. But most big companies are still on these old platforms like nothing changed in the last ten years.

The Candidate Was Never the Customer

These platforms were sold to HR departments, not to job seekers. The vendor gets paid by the enterprise. The candidate just absorbs the pain.

Companies spend so much on employer branding — fancy careers pages, culture videos, big talk about innovation. Then they route every candidate through a portal that looks like it was last updated when Hrithik Roshan was in Kaho Na Pyaar Hai.

Also most of these portals are completely broken on mobile. And the resume parsers inside them are so outdated that a modern resume layout can get you filtered out before any human even sees your name.

The Writing Is Already on the Wall

Platforms like Eightfold AI and Workable already solved this. You upload your resume once, the system reads it and fills everything in for you. Done. What legacy platforms call “just how it works” — these companies fixed it and moved on years ago.

A whole new generation of AI-native recruiting tools is being built from scratch. No legacy baggage. No outdated codebase. Designed around the applicant, not the HR admin console.

Remember Kodak? They actually invented the digital camera. Had the technology in their own hands. But they were so comfortable with their film business, so locked into what was already working, that they could not bring themselves to change. Nokia had the best hardware in the world and still got wiped out because the software world moved and they did not.

These legacy HR platforms are in the same spot right now. Not because they lack resources. Not because the technology does not exist. But because there is no urgency inside a company that is already collecting enterprise cheques and renewing contracts every three years.

The AI-native platforms do not need to fight them directly. They just need to keep building. Every frustrated candidate who drops off mid-application, every good hire a company loses because the portal was too painful to complete — that is the market slowly shifting. Quietly, consistently, without any dramatic announcement.

By the time the legacy vendors feel it, the gap will already be too wide to close.

We are possibly in the final years before a full-blown intelligence explosion — a moment when AI systems dramatically outpace human cognitive output across nearly every domain. Yet if you walk through any tier-2 or tier-3 Indian city today and speak to parents, students, local politicians, or journalists, you will find almost no awareness of what this actually means.

Not because people are unintelligent. But because no one is telling them the truth.

The IT Dream That Became a Trap

Over the last three decades, the IT boom reshaped the aspirations of an entire generation of Indian families. For parents who had spent their lives doing physically demanding work — on shop floors, in warehouses, behind counters — watching their child sit in an air-conditioned office in formal wear, earning a salary they could never have imagined, felt like the ultimate victory.

That dream hardened into dogma: engineering college, software job, FAANG salary. This became the socially accepted definition of success, especially in middle-class households. IIT became the holy grail, and where that wasn’t accessible, a thousand coaching institutes, backdoor consultancies, and referral networks filled the gap.

It worked, for a while. Global Capability Centres opened. Multinational companies set up local branches. State governments competed with one another, offering land, tax holidays, and long-term leases in special economic zones. Real estate boomed around tech parks. Infrastructure followed. Politicians claimed credit. People believed them.

And for at least two decades, it made sense to believe them.

Then Came ChatGPT

The launch of ChatGPT in late 2022 was a watershed moment — not just technologically, but psychologically. For the first time, a broad swath of people witnessed a machine performing knowledge work: drafting, reasoning, coding, summarising, translating — tasks that until recently required years of education and experience.

And the pace hasn’t slowed. Every quarter, every major AI lab releases something more capable than the last. The progress is not linear; it is compounding.

What once took a human days or weeks can now, in many cases, be done in seconds. That isn’t an exaggeration — it’s a product demo anyone can run themselves.

This isn’t just a hunch. In November 2025, the McKinsey Global Institute released one of its most striking findings yet: technologies that already exist today, not ones still in development, could technically automate around 57 percent of all work hours in the US economy, roughly split between software agents handling non-physical tasks and robots handling physical ones. That figure nearly doubled McKinsey’s own 2023 estimate within two years, and the researchers frame it as a ceiling on technical potential, not a prediction that this many jobs vanish overnight. But the direction is unambiguous, and the pace of revision — upward, every year — tells its own story. No industry is truly exempt: not IT, not finance, not law, not medicine, and certainly not government administration.

What Indian Governments Are Actually Saying

Here’s where things get frustrating.

Instead of an honest conversation about displacement, reskilling, and economic restructuring, what we get from most state governments is chest-thumping: “We signed an MOU with Company X.” “Our state will have an AI city.” “Data centres will bring employment.”

This isn’t a hypothetical script. In 2026 alone, Uttar Pradesh approved a Data Centre Policy targeting roughly ₹2 lakh crore in investment and up to 2 gigawatts of new capacity, with the state government’s own estimate putting the resulting indirect employment at around 50,000 jobs. Gujarat signed a ₹25,000 crore MoU with L&T for a 250 MW AI-ready data centre campus in Dholera. Odisha, Telangana, Andhra Pradesh, and Haryana have all signed similar headline-grabbing agreements over the past year, each one covered as a jobs win.

Here’s what rarely makes the headline: internationally, the research on data centre employment is now fairly consistent, and it isn’t flattering. A 2026 Brookings Institution study that tracked counties in the US which received their first large data centre found only a modest 4–5 percent rise in overall private employment over five to six years — and that’s the positive finding. Separate industry workforce benchmarks put permanent staffing at the most automated hyperscale campuses at roughly 20–40 people per 100 megawatts once construction wraps up. Construction itself creates real, if temporary, work — often in the hundreds or low thousands of jobs for 18 to 36 months — but the facility that remains afterward typically runs with a skeleton crew of specialised technicians, security, and site managers, not the mass local employment implied by ribbon-cutting ceremonies.

This isn’t a secret. It’s documented in report after report. Yet no journalist of consequence, no editorial board, and no policy think tank in mainstream Indian media is pushing back on this narrative with the urgency it deserves.

Because it doesn’t affect them yet. And in India, problems only become real once they land at your own door.

The Counter-Argument — And Why It Isn’t Enough

To be fair, the optimistic case deserves a hearing. Every major wave of automation — mechanised farming, computers, the internet — triggered predictions of mass unemployment that didn’t fully materialise. New job categories emerged that no one could have predicted in advance. McKinsey itself is careful to say its 57 percent figure measures technical potential, not an inevitable wave of layoffs, and argues the more likely outcome is humans, software agents, and robots working in partnership rather than humans simply being replaced.

The trouble with leaning on this comfort is threefold. First, previous automation wave, however painful, typically unfolded over multiple decades — giving families, institutions, and labour markets time to adjust. The current pace, by McKinsey’s own account, is compressing timelines that used to run to 2060 into a window centred on the 2030s. Second, past automation mostly displaced physical and routine labour; this wave targets cognitive and knowledge work first — precisely the category India spent thirty years training an entire generation to enter. Third, “new jobs will emerge” is true in the aggregate but says nothing about where and for whom. A software engineer in Vizag whose role is automated is not automatically re-employed by whatever new category emerges in San Francisco or Bengaluru’s AI labs.

Historical precedent is a reason for hope. It is not a plan. And right now, India doesn’t have one of those either.

The Economic Chain Reaction No One Is Modelling

Let’s trace the logical chain that almost no one in public discourse is willing to follow through.

If private-sector IT employment stagnates or contracts — and given the automation trajectory above, that is a real risk — consumer spending contracts with it. The restaurants, malls, real estate, ride-sharing, retail, and hospitality sectors built on the back of tech-sector salaries begin to slow. Tax revenue falls. State governments with bloated payrolls and committed infrastructure spending face a fiscal squeeze.

A government that cannot pay its employees on time is not a stable government, regardless of how permanent those jobs appear on paper.

The money stops rotating. And once that happens, the consequences for regular employees and daily-wage workers are severe and immediate.

This is not a slow, manageable transition we’re discussing. It’s a structural shock for which India has no public preparation plan.

Universal Basic Income(UBI) Is Not the Answer

UBI tends to be the first policy reflex in these conversations. It shouldn’t be.

Distributing income without addressing the structural collapse of productive activity is just deferred chaos. The more useful question governments should be asking is: what happens to food security, water access, and basic human welfare when large-scale unemployment arrives? How do we ensure a worst-case scenario doesn’t trigger civil unrest?

What Should Actually Happen

If chest-thumping MOUs aren’t preparation, what is? A few concrete starting points:

• Mandate independent, published job-impact audits for every state-level AI or data centre MOU — permanent jobs versus construction jobs, stated separately and verified, not bundled into one impressive-sounding number.

• Redirect a portion of the tax incentives currently given to data centres toward vocational reskilling programs aimed specifically at IT and BPO workers whose roles are closest to automation.

• Invest seriously in agricultural R&D, water infrastructure, and decentralised food systems — sectors that are far more labour-intensive and far less exposed to near-term automation than the ones currently receiving the political spotlight.

• Build a genuine public safety net designed around welfare outcomes (food, health, housing) rather than a single cash-transfer number that sounds good in a headline but doesn’t address the underlying collapse in productive employment.

• Create a public, state-wise dashboard tracking AI-linked job displacement and creation, so the debate is grounded in real numbers instead of competing press releases.

None of this requires exotic new policy instruments. It requires treating this as the structural issue it is, rather than a talking point to be managed until after the next election cycle.

Why This Silence Suits Everyone in Power

Politicians benefit from delayed public awareness. The longer people stay distracted — by caste politics, celebrity news, regional rivalries, cricket — the less they demand preparedness.

This isn’t accidental. The art of Indian political communication has, for decades, been to keep voters focused on identity and grievance rather than on structural economic futures. It works, until it doesn’t.

The reckoning will come. It always does. The only question is whether it arrives as a manageable transition or as a full-blown crisis.

People will start asking the hard questions the moment their own livelihood feels the pressure — and by then, the window to prepare will have already closed. We should be asking these questions now, not waiting for a crisis to make them feel urgent.

The storm is coming. Whether we’re ready for it is entirely up to us.

Share

🧠 Human driven, AI assisted. The perspectives shared in this newsletter are my own, refined and structured with the help of Large Language Models.

Look at the internet and social media boom. The adoption rate was staggering. Affordable smartphones, and Jio’s free internet offer in 2016, were game changers. That single move shocked not just Indian telecom players but the entire world. Within months, nearly every household had at least one Jio SIM card — that’s the scale of impact we’re talking about.

The Telecom Boom That Changed Everything

Once cheap data became the norm, tech companies saw their opening. A flood of affordable smartphones and new apps followed, making online communication and collaboration easier than ever. And the cheaper the access, the more time people spent online.

Around the same time came another turning point: demonetisation. It completely reshaped how India made payments. The government’s BHIM app was a simple but brilliant starting point — lightweight, straightforward, and the first exposure many Indians had to digital payments. Soon after, private startups built their own UPI-based apps, and the digital payments ecosystem exploded into what it is today.

The Great Indian App Fragmentation

Here’s the real point: when new technology becomes widely accessible, it hands everyone — startups and governments alike — the ability to build. And that’s exactly what happened in India. Every state government launched its own apps and websites for citizens: one for electricity, one for gas, one for transport, one for industries, and so on. As the app boom continued, most of these services split off into their own standalone mobile apps.

And here’s the amusing part — every time a new government comes to power, a few months later, “new” versions of the same old apps mysteriously appear. Same functionality, a slightly different design, and of course, a fresh photo of the new Chief Minister on the home screen.

The scale of this fragmentation is easy to underestimate until you actually try to use it. Move from Delhi to Bengaluru, and you’ll find yourself downloading an entirely new electricity app, a new gas-booking app, and a new transport app — each with its own login, its own interface, and its own quirks. Multiply that across 28 states and 8 union territories, and you get a citizen experience that resets every time you cross a border.

The Missing Piece: A True Super App

What hasn’t evolved is this: no one — not the central government, not any state government — has attempted to bring these services together into a single, unified “super app.” And I’m not only talking about government services. Imagine one app that handles everything: booking cabs and flights, ordering food and medicine, paying bills, and more.

A few players have tried. Paytm and Tata Neu both made attempts, but neither has fully cracked it. Meanwhile, we keep seeing headlines about state governments banning or restricting ride-sharing apps like Ola, Uber, or Rapido following local protests — with some states even launching their own short-lived alternatives to calm things down. These are patches, not solutions.

So Why Hasn’t This Happened Yet?

It’s worth asking why, given all the pieces are already in place, no one has actually built this. A few reasons stand out:

• No single owner. Government services fall under different ministries and state departments, each with its own budget, vendor, and priorities. There’s no one body mandated to own a unified citizen-facing platform.

• Political incentives run the other way. A state-branded app — complete with the ruling party’s face on the splash screen — is a visibility tool. A shared, neutral, cross-state platform offers no such political payoff.

• Data privacy and trust concerns. Merging identity, location, payments, and service history into one app raises real questions about who controls that data and how it could be misused.

• Interstate friction. Transport and ride-sharing regulations differ by state, and local politics around gig-economy jobs (as seen in the recurring Ola/Uber/Rapido bans) make a single national experience harder to standardize.

None of these are insurmountable, but together they explain why the technology has existed for years while the product hasn’t.

Why Should a State Border Decide Which App I Use?

Think about it from an ordinary user’s perspective: should we really need to install a new app every time we cross into a different state? That makes no sense. Part of why Uber works so well globally is consistency — you can land in any city on earth, open the same app, and it just works. Why can’t India offer that same experience?

We already have the foundation for it: Aadhaar, a universal digital identity that could easily plug into a unified system. Imagine a single app that lets you:

• Book your gas cylinder

• Pay your electricity bill

• Recharge your phone

• Order food or medicine

• Book a flight or train ticket

• Use the metro or local buses

One app. Every service. Anywhere in India. Built in with a secure digital locker for your documents and a wallet for instant payments. Apps like Paytm, PhonePe, CRED, Tata Neu, and even Ola already have much of the underlying infrastructure needed. The real question is: who’s willing to take the leap?

What Malaysia’s Grab Taught Me

Last year, I visited Malaysia to meet a friend, and I was genuinely blown away by the Grab app. It handled everything — rides, deliveries, food, and payments — all from one place. As a first-time visitor, I didn’t need to ask anyone for help navigating the city; everything simply worked.

Coming back to India, I couldn’t stop thinking: why don’t we have something like this yet? Years later, we still haven’t taken a real step in that direction.

this isn’t just a Malaysia story. China’s WeChat folded messaging, payments, ride-hailing, and government services into a single “everything app” years ago. Indonesia’s Gojek did much the same, evolving from a motorbike-taxi app into a full super app for an entire country of over 270 million people. If two very different economies in our own region have managed this, the excuse that India is “too big” or “too complex” starts to look thin.

The Case Against a Super App

To be fair, a single unified app isn’t without risk. Concentrating rides, payments, identity, and government services into one platform creates a single point of failure — an outage takes down everything at once, not just one service. It also raises monopoly concerns: whoever builds and owns this app would hold an enormous amount of power (and data) over how citizens access basic services. And a major breach wouldn’t just leak your chat history — it could expose your identity, your finances, and your location in one shot.

These aren’t reasons to abandon the idea, but they are reasons to get the governance right before the technology — clear data ownership rules, strong security standards, and probably some form of public oversight, rather than a single private company holding all the keys.

The Way Forward

Building a super app of this scale isn’t a job for tech companies alone. Central and state governments need to collaborate too, looking beyond short-term politics toward long-term impact for citizens. Once a unified system exists, details like revenue-sharing and service charges will sort themselves out. What matters most is building one seamless digital experience that makes every Indian’s life simpler, faster, and more connected.

So here’s the question I keep coming back to: if you could design India’s super app from scratch, what would you want it to do first — and who would you trust to build it?

🧠 Human driven, AI assisted. The perspectives shared in this newsletter are my own, refined and structured with the help of Large Language Models.

BHIM (Bharat Interface for Money) is a mobile app developed by the National Payments Corporation of India (NPCI) that facilitates quick, secure, and easy transactions using the Unified Payments Interface (UPI). It was launched by the Government of India to promote digital transactions and enable cashless payments.

What’s the issue with the app now ?

tbh, when compared with other payment apps, BHIM is the most light weight app till date. No big features, notifications and no fuss. But the biggest issue with the app is the UX. I’m clueless about the designers point of view in selecting that colour palette, placement of buttons, features - user journey etc existing in the app.

Places where there is a scope to improve the UX in the app are :-

- App widgets -

Generally most of the apps are utilising the core features/API’s provided by the mobile OS companies, Google and iOS. UPI team didn’t utilise the most important feature for making user journey more seamless. Here they should have added shortcuts to core features - scan the code, send to favourites, approve payment request, check bank account balance, tap the machine.

- Verify the ID/number:

We all know that in India all the mobile phone numbers are 10-digit in length and also all the UPI id’s end with “@upi/@okhdfc/@okybl” etc. So whenever user finishes entering the UPI id’s with such ending or when they enter 10 digit number’s app should check the validity of the number/id automatically. Here also the user should do a additional step to check the id

- Colour palette & dead icons -

This part is the most important zone of the app and this is where the team has did a blunder i feel. The colours for the all the icons here are same which makes it a little difficult to identify the exact thing we are looking for. Even I personally found it very hard to find “approve to pay” icon whenever I am trying to pay money. Team should have opted for colours a little different to each and every icon which would have made a lot of difference in UX. Their intention to emphasise more on UPI lite balance and Credit card zones, as visible in the screenshot, should have been little less and should have been more on core features icons.

- Abandoned features/icons:

Payments to bills is also a vital feature to users. Few may argue that they may not be using it on a daily basis, but still the users will be using them if we do provide them at the proper place which in turn makes their journey easier. They don’t need to put extra effort in using that feature. Also when we swipe to the right of this area we go into a new page where we can see a ton of bills categories that we can utilise. This is similar to how internet gives us a gateway to access data about each and every corner of this world. Even in todays world there are people in tier 2-3 cities and villages where they try to go to a local office or internet centres to pay for bills, where they are charged additional fees to pay bills. If the govt’s intention of bringing services to users on phone is very much core to them then this zone also needs to have a better UX and user should be guided to that zone while onboarding.

- Scan to pay & Tap to pay zones:

even in this area the icons are too small and also they do cover other zone. Even this also needs to be properly redesigned at the first place.

- Grouping of services across the whole app:

all the services available in this entire app needs to be consolidated and put in one single place for simplicity so that user don’t struggle to find a certain service. As I said above all the services may not be used on a daily places but all should give proper visibility to the users.

Hope that NPCI UPI team looks into this kind of UX issues at least in their upcoming releases and make the user experience way more better and seamless.

But today Apple’s plans, as per news articles in response to EU demands is quite shocking to me. Apple always strived to be a company which reflects those 3 items, stated above. But looks like they will not be present in future if Apple implements all the proposed changes.

• Side loading of apps - It clearly states that apps can be installed without an app store - without the app review process. Generally, an app is reviewed by the app store team, before being pushed to the app store, to check whether user data is being collected without their consent and that data is being transferred to any other companies etc. Google’s Android and its ecosystem is the prime example of that. In Android, lots of user data gets leaked through side loading of apps from unsecured sources. I personally worked as an Android apps developer for many years and know how data can be siphoned off without user consent. Later when I came to know about the Apple ecosystem, I personally fell in love with it’s stringent policies for data collection and the way the data should be collected by the apps. These days even Google is trying to make Android OS a safe one step-by-step. It has yet to reach out to that place, but I am quite optimistic about that. And now if Apple starts allowing side loading of apps a huge user data will be at risk and. Here the question is whether Apple or the European Union will be responsible for the data loss that might happen. If it is up to Apple, then there will be a huge pressure mounting on them to figure out a way to handle those leaks and also stop them. There is an article on this - Why sideloaded apps and Is sideloading apps on Apple devices safe? (setapp.com) Another effect of this sideloading of apps is the revenue loss for the developers. Yes, we often find a lot of modified apps on the Internet to install so that we can skip that payment part. This is also a big issue which going to pop soon.

• Alternative app stores - Yes in one way it is correct to say that the app store is a monopoly but in another way, it’s the best thing. 1 question that pops up to me is “Do all the alternative app stores follow all the same rules regarding app reviews, data collection, security, privacy, etc ? “. Proposing alternative app stores on the same device just for 1 reason - app store commission - is not a good idea definitely.

Hope the European Commission and its members look at all aspects and propose a new solution by coordinating with the Apple technical team. Even though I don’t work for Apple still I am a big fan of its app store and its stringent policies for data security, privacy, and safety. I feel I am in a position to speak about this considering my previous technical work experience. Do let me know if I missed out on some critical views/points.

Hide Followers & Friends list -
Most of the teens these days feel insecure to have more people connected in a single account. So many of them are having an alternate account to add only close people. It would be much better if Instagram provides an option to completely make their followers and friends list private and nobody can see them except the account owner.

Post visible for certain people only -
If people really want to share pictures to a few select people, the above point makes more sense. Yes, it is similar to what Facebook implemented years ago, but in current time this custom posting feature would make more sense.

Trending Section topic chooser -
These days the trending section shows content based on your post’s interaction and searches etc. and lot of things. This would make more sense in a relevant ads-based products, but many people are really frustrated about this. For example, if a user searches for popular resorts in a particular ‘X’ city, all the ads and trending section changes after some time. This might be perfectly valid for some time till the user finds what he needed, but that might not be relevant when the user no longer wants to search for resorts. So, I feel the users should be able to choose the topics in trending section on which the content needs to be shown.

Birthday remainder -
This feature used to be too good on the FB since its inception and I personally loved the feature a lot. These days everyone is quite busy and since they are engaging more with Instagram when compared to other platforms, I believe Meta team should bring this feature to Instagram also. In this we way we can help the users to get track of their friend’s birthdays and wish then and also indirectly increase the user engagement.

Avatar quality -
The latest feature introduced by the Instagram team is the Avatar feature, which is a competitor to Bitmoji and Memoji, up to some extent. The primary reason why I said only up to some extent is the Avatar quality. TBH it is of very poor quality. The design of avatar faces etc. itself is not inclusive and does not accurately capture the user’s face into avatar. This feature needs to be redesigned with the tech where the user’s avatar details are captured accurately. If Mark and meta really wants to be successful in next booming tech - metaverse - they really need to focus more on this feature.

Issues/alternatives to fix that issue:-

1) charging: The number of people who prefer apartments over individual houses is very huge. In this segment, people will be puzzled about how to charge their EVs in apartments because they need their own charging point so that they can plug in & charge the battery without any hassle. Also, there can be a charging point in the apartment cellar that automatically recognizes the vehicle owner and sends the electricity bill to that respective owner. Some may argue that users can use public charging points, but why should people go outside and charge their vehicles? Fixing this issue might reduce the burden on govt to arrange public points, and also drive the EV to full-scale adoption.

2) Publicpoints - If the above-mentioned idea is not possible then people should definitely rely on public charging points. Installing public charging points by govt or private companies will take a long time for sure. In this case, as an alternative, Pvt companies and govt offices should install charging points on a full scale and also give some reward points to the employees who own an EV and utilize employers charging points.

3) Swappable- Time is the most precious thing in everybody’s lives. When you are traveling generally you stop at a petrol/diesel bunk & fill your vehicle and start driving immediately. But in the case of EV, you need to recharge the battery which takes a long time and there should be a recharging station at regular distances which is not an easy task to set up. Hence a swappable battery kind of should be present in EV’s.

Once these issues are solved, I believe people will have no fear/doubts about using an EV. Hoping to see these issues solved by EV companies in the coming days...

#EV #ElectricVehicles