By Govind Rammurthy, CEO & Managing Director, eScan
Every week, I see another advertisement for a new course: “Master ChatGPT in 7 Days,” “Become an AI Expert,” “Learn Prompt Engineering.” Last month, a recent graduate applied to eScan with a resume listing five AI certifications from various online platforms. When I asked him to explain how he would use AI to solve a real security problem we were facing, he froze. He could recite definitions. He knew how to phrase prompts. But he had no idea how to think through a problem.
This is the crisis facing India’s IT workforce, and most people are looking in the wrong direction.
The edge we’re losing
For two decades, India rode a simple advantage: capable developers at a fraction of Western costs, with English fluency that China and other competitors lacked. That combination made us the world’s back office. But every advantage we built is eroding simultaneously.
Vietnam now offers comparable development costs with improving English proficiency. GitHub Copilot and Claude Code write code that often needs less debugging than what junior developers produce. Companies that once needed 100 developers can now accomplish the same work with 30 developers and AI tools. The math is brutal and it’s accelerating.
Then came the security incidents. When Marks & Spencer lost £300 million from a cyberattack traced to a contractor, when Jaguar Land Rover’s production stopped for weeks causing £1.9 billion in damage, when Clorox sued Cognizant for $380 million over a helpdesk breach – each incident reinforced a growing perception: Indian IT services may be cheap, but can you trust them with sensitive systems.
But the biggest threat isn’t competition or reputation damage. It’s that we’re producing hundreds of thousands of graduates every year whose skills are becoming obsolete before they even enter the workforce.
The certificate scam
Let me be blunt: knowing how to write a ChatGPT prompt doesn’t make you an AI expert any more than knowing how to use Google makes you a search engine expert. Yet we’ve created an entire industry selling this illusion.
These courses promise to make you “AI-ready” in days or weeks. They teach you to phrase questions, explain different AI models, maybe show you how to use a few tools. Then they hand you a certificate that’s worth less than the paper it’s printed on – because employers who actually work with AI can spot the difference between prompt typing and genuine capability within minutes of conversation.
I’ve interviewed many candidates with impressive-looking AI certifications. Most can’t explain how large language models (LLM) actually work. They can’t identify when AI is giving them wrong information. They can’t think critically about whether AI is the right tool for a problem. They’ve learned to use AI tools the way someone learns to use a calculator – they can push buttons but don’t understand the mathematics.
This isn’t preparing people for an AI-powered future. It’s creating a generation that thinks they’re skilled when they’re actually just trained to follow scripts that will themselves be automated.
What actually matters
At eScan, we’re increasingly researching with AI for our threat detection systems. When we hire for these roles, we’re not looking for people who can write clever prompts. We’re looking for people who can:
Think systematically about problems. When our AI flags something as suspicious, does the person understand enough about both security and AI to evaluate whether it’s a genuine threat or a false positive? Can they trace back through the logic to understand why the AI reached that conclusion.
Understand AI’s limitations. AI makes mistakes. It hallucinates. It has biases based on training data. The people we need aren’t those who trust AI blindly but those who know when to question it, when to verify, when to override it.
Adapt to changing systems. The AI tools we use today will be obsolete in two years. We don’t need people trained on specific tools – we need people who understand fundamental concepts well enough to learn new tools quickly.
Combine domain expertise with AI capability. A cybersecurity expert who learns to use AI effectively is far more valuable than an “AI expert” who doesn’t understand security. The domain knowledge is what AI can’t replace.
This last point is crucial. Every company I know that’s experimenting with AI in production environments makes the same observation: domain expertise becomes more valuable, not less, when combined with AI. The AI handles routine work, but understanding what questions to ask, how to interpret results, whether the AI’s recommendation makes sense in context, requires deep knowledge of the actual field.
What education should focus on
India’s National Education Policy talks about AI training, but most programs focus on using AI tools rather than thinking critically about problems. That’s backwards.
We need to teach:
Problem decomposition — breaking complex problems into components you can reason about. This matters whether you’re using AI or not, but it’s essential for knowing how to apply AI effectively.
Statistical thinking — understanding probability, recognising patterns, evaluating evidence. AI makes predictions based on patterns in data. If you don’t understand statistics, you can’t evaluate whether those predictions make sense.
Systems thinking — understanding how components interact, how changes propagate, where dependencies exist. Modern software systems are complex, and AI adds another layer of complexity.
Critical evaluation — questioning assumptions, identifying biases, verifying claims. AI will confidently give you wrong answers. You need to know how to catch that.
These aren’t “AI skills.” These are thinking skills that become more important when AI handles routine cognitive work.
The path forward
India won’t regain its IT services advantage by producing more developers who can write code – AI is already doing that. We won’t succeed by certifying people in prompt engineering – that’s not a sustainable skill.
We’ll succeed if we produce people who can think clearly about complex problems, understand systems deeply, and use AI as a tool while recognising its limitations. That means fundamentally rethinking education from rote learning and test scores to developing actual analytical capability.
The uncomfortable reality is that most of what Indian IT education focuses on today – memorising syntax, learning specific frameworks, following established patterns – is exactly what AI handles best. We’re training people for jobs that are disappearing while ignoring the skills that will actually matter.
The companies that will thrive, and the professionals who will succeed, are those who understand that AI isn’t replacing human thinking – it’s making shallow thinking obsolete while making deep thinking more valuable. India’s youth need to be on the right side of that divide.
The certificates won’t save them. Only genuine capability will.