India's Biggest AI Lie
From Bengaluru startup pitches to government "Digital India AI" schemes — the industry is selling automation and delivering a glorified autocomplete. Here's how the scam works.
Walk into any startup event in Bengaluru, Mumbai, or Hyderabad right now. Within ten minutes, someone will tell you their AI product will "eliminate manual work," "replace repetitive tasks," or "free up your team." The slide will show a graph going up. The demo will be smooth. The founder will be confident.
And most of it is, to put it plainly, jhooth.
Not outright lying — that would be too easy to call out. It's more sophisticated than that. It's the art of letting you believe one thing while selling you something else entirely. And India, with its massive IT workforce, its lakhs of BPO employees, its government digitisation push, and its desperate hunger to be seen as an "AI-first nation" — is the perfect target market for this trick.
To understand exactly what's happening, you need to understand two words the industry uses as if they mean the same thing. They don't.
What Automation Actually Means
Automation means the human is removed from the task completely. The machine does it. The person is no longer needed in that loop.
Think of the STD booth operator who used to connect trunk calls. Gone — automated by direct dialling. Think of the data entry operator manually typing bank transactions. Gone — automated by core banking systems. Think of the factory worker tightening bolts on an assembly line. Gone — automated by robotic arms.
These people didn't get redeployed to easier work. The work disappeared, and so did their jobs. Automation is honest about this. It doesn't promise you'll be more productive — it promises the human position will no longer exist.
What Augmentation Actually Means
Augmentation is different. The human stays. The tool makes them better at what they do.
A CA using accounting software isn't automated — they're augmented. The software doesn't file your returns; it helps the CA work faster, catch errors, and handle more clients. A doctor using an AI diagnostic tool isn't replaced — they review the suggestions, apply clinical judgment, and make the final call. The tool raises their capacity. It doesn't replace their presence.
Augmentation assumes your judgment, your context, your relationships still matter. The AI just helps you show up better with those things.
- Human removed from the task
- Machine makes the decision
- Goal: eliminate the headcount
- Success = position no longer needed
- Example: auto-processing insurance claims
- Human stays in the loop
- Human makes the final call
- Goal: raise human output
- Success = same team does more, better
- Example: AI drafts; manager approves
Where India Gets Scammed Specifically
Here's the pattern. An Indian company — a mid-size IT firm, a bank, a logistics player, a government department — buys an "AI automation" product. The pitch was automation. The price was automation. The board approved it as automation, expecting to reduce headcount or reassign people.
Six months later, the team is still there. They're just doing different work: reviewing AI outputs, correcting its mistakes, handling exceptions it can't process, and explaining to management why the numbers look weird this quarter. The tool is augmentation. The contract said automation. The gap between those two words is where the money went.
This is not hypothetical. It's happening in every large IT services company running "AI transformation" projects for clients. The client buys automation. The vendor delivers a product that still requires human oversight. The vendor collects implementation fees, support fees, and licensing fees — while the client quietly keeps the same number of employees, just with different job descriptions.
(One job. AI did half. Rest is yours. Full bill arrived.)
The Indian IT Industry's Specific Problem
India has roughly 5 million IT professionals, many of whom work in services roles — writing code, managing systems, testing software, processing data — for foreign clients. For twenty years, the pitch to those clients was: "cheaper skilled labour." AI suddenly threatens that entire model.
So what happened? Indian IT companies pivoted. Overnight, everyone became an "AI company." TCS has an AI platform. Infosys has a generative AI unit. Wipro is "AI-first." Every mid-size company has an "AI Centre of Excellence" staffed by people who, six months ago, were doing something else.
The product they're selling their foreign clients is automation — "we'll use AI to do more with fewer people." But internally, they're terrified to actually automate their own workforce, because that's also their revenue. So they sell automation to clients while running augmentation internally — hoping no one notices that the headcount barely moved.
The tell: Ask any Indian IT company how many employees they actually reduced after deploying their own "AI automation" tools internally. Then ask how much they charged clients for the same tools.
The Government Angle: "AI for India" as a Budget Exercise
The scam has a government flavour too. IndiaAI Mission. National AI Portal. State-level AI policies from Maharashtra, Telangana, Tamil Nadu. Crores being allocated for "AI in governance" — AI for agriculture, AI for health, AI for education.
Almost none of it is automation. A farmer using an AI crop advisory app still has to read the advice, judge whether it fits his field, and decide whether to follow it. A government officer using an AI document summariser still has to review the summary, catch errors, and sign off. These are augmentation tools — good ones, potentially — but they're being funded, announced, and celebrated as if they're removing humans from government processes entirely.
The result is that crores get spent, consultants get paid, press releases get issued, and the actual government process is unchanged — except now there's an app nobody uses sitting between the citizen and the clerk who was always going to make the final decision anyway.
The Startup Pitch Deck Translation Guide
If you sit through enough AI startup pitches in India, certain phrases start to have very predictable translations. Here's a quick guide:
"Reduces manual effort by 70%" → 70% of the boring part; the tricky 30% still needs a human
"No more X" → Less X, reviewed by a human who knows what X should look like
"AI-powered decision making" → AI gives options; a human still decides and signs
"Scales without adding headcount" → Scales until something goes wrong, then needs headcount urgently
Why Indian Companies Are Especially Vulnerable
Three things make India a particularly easy market for this confusion.
First, the aspirational pressure is enormous. No Indian business leader wants to be seen as behind on AI. Saying "we evaluated it and it's augmentation, not automation" sounds like you didn't get it. Saying "we're fully AI-automated" sounds modern. So companies buy the story even when the product doesn't match it.
Second, the cost-cutting mandate is real. Indian businesses, particularly post-2022, are under intense pressure to cut operational costs. Automation justifies that. Augmentation doesn't — it just means your existing team gets slightly better tools. So finance teams approve automation budgets that actually buy augmentation products, and nobody wants to be the one who admits the headcount didn't move.
Third, accountability is diffused. In large Indian organisations — both corporate and government — when an AI system produces a wrong output, the question of who is responsible becomes a long, unresolved argument between the vendor, the IT team, the business unit, and senior leadership. This diffusion is actually why AI vendors love selling into these structures. The human who should catch the error exists, but nobody officially assigned them that role. So errors persist, and the vendor is insulated.
What Honest AI Sales in India Would Look Like
An honest pitch to an Indian company would sound like this: "This tool will make your existing team significantly more productive on these specific tasks. You will still need those people. Their jobs will change — they'll spend less time on X and more time on Y. Measure success by output quality and throughput, not by headcount reduction."
That's true. That's valuable. Many Indian companies would — and should — pay for that. But it doesn't give the CEO a talking point about "AI-driven efficiency" in the annual report. It doesn't let the board claim they're running a leaner operation. It doesn't excite investors the same way.
So the honest pitch doesn't get made. The automation pitch gets made instead. And somewhere downstream, a team of people who were told AI would replace them are instead spending their days supervising, correcting, and cleaning up after the AI that was supposed to make their jobs disappear.
How to Protect Yourself and Your Organisation
The simplest question to ask any AI vendor — Indian or foreign — is this: what does a human have to do after your product runs? If the answer involves any reviewing, approving, correcting, exception-handling, or oversight — you are buying augmentation. Price it accordingly. Staff for it accordingly. Measure it accordingly.
Do not fire the team that will need to supervise the AI before the AI has proven it doesn't need supervising. That is the most expensive mistake Indian companies are making right now — reducing the workforce based on automation promises, then scrambling to rehire or reskill when the augmentation reality becomes undeniable.
And if a vendor cannot answer the question — "what does the human do after your AI runs?" — that is not a sign of a sophisticated product. It is a sign that they have not thought clearly about where their product ends and your problem begins.
(First ask: who does the rest after AI? That person is your real cost.)
The Honest Bottom Line
AI tools — honestly used — can be genuinely transformative for Indian businesses. Augmentation, done well, means a team of ten can do what previously needed thirty. That's real. That matters. India's talent pool, combined with genuinely good AI tools, is a serious competitive advantage.
But that advantage only materialises if you're honest about what you're buying. Augmentation requires investment in the humans who will use the tool — training, role redesign, supervision structures. Automation requires replacing those humans. Confusing the two means you neither invest in your people properly nor successfully remove them — you just end up with an expensive product, a confused team, and a gap between your board presentation and your actual operations.
The AI industry — global and Indian — will keep selling automation because automation is what gets approved. It is your job, as a buyer, to demand the honest answer about what the human still does.
Because somewhere in your organisation right now, there is a person doing the job the AI was supposed to eliminate — and they are also doing the new job of cleaning up after the AI. They are exhausted. They did not get a raise. And the vendor already cashed the cheque.
Comments
Post a Comment