Should We Still Chase Competitive Exams in the AI Era?

An AI model can now score in the top percentile of most entrance exams. That fact sits uncomfortably with the Indian middle-class blueprint: coaching from class 8, a rank, a seat, a secure life. If a machine can ace the exam, what exactly is the exam certifying? Parents deserve a straight answer — not slogans from the coaching industry ('nothing has changed, enrol now') and not slogans from the AI-hype industry ('degrees are dead, learn prompting'). The truth is in between, and it has real consequences for how your family spends the next five years.

Start With Why the Blueprint Existed

The exam-rank-seat pipeline made sense for a specific economy: scarce college seats, scarcer good jobs, and employers who had no better signal than the name of your institution. For two generations it genuinely worked — an IIT or AIIMS seat reliably transformed a family's trajectory. The system's costs (childhoods spent in coaching, mental health tolls, a single-number identity) were accepted as the price of that transformation.

AI does not abolish this pipeline. Colleges still admit by rank; the gate is still real. What AI changes is the value of what happens on either side of the gate.

What Exams Still Do Well

• Gatekeeping still exists: top engineering and medical colleges admit through ranks, and those campuses still concentrate opportunity — peers, alumni networks, recruiters. The exam's signalling value has weakened, not vanished.
• Deep work capacity: preparing seriously for two years builds discipline, delayed gratification, and frustration tolerance. These transfer to every field, including entrepreneurship.
• Foundation depth: physics, chemistry, and mathematics learned properly — not pattern-matched — remain the base for understanding and building AI itself. The irony is that real JEE-level maths is excellent preparation for the AI economy; rote JEE coaching is not.
• Medicine remains a licensed profession: NEET is not optional for a doctor, and healthcare demand in India is only growing. The calculus for medicine differs from engineering.

What Has Quietly Changed

1. The Exam Gets You In; It No Longer Carries You After

Employers increasingly test what you can build and how you think with tools, not what you memorised. Hiring processes now routinely include AI-assisted tasks. A rank-holder who cannot use AI fluently loses to a mid-ranker who can. The rank opens the door; capability keeps you in the room.

2. The Marginal Cost of Cramming Has Collapsed in Value

There is a difference between the first 80% of preparation (learning concepts deeply, solving hard problems) and the last 20% (grinding pattern-recognition on previous years' papers, sacrificing sleep, sport, and friendships for two extra percentile points). The first 80% builds a mind. The last 20% builds a slightly better test-taker — and pattern-matching under time pressure is precisely the skill AI has commoditised. That final, most expensive stretch of the race is the part whose payoff has fallen the most.

3. The Backup Careers Moved

The old logic said: even if the rank doesn't come, the preparation leads to a decent engineering college and a safe IT job. But routine IT services work — the historical safety net of Indian engineering — is exactly the layer AI automation is compressing. The 'safe fallback' is no longer safe by default, which changes the risk maths of an all-in exam strategy.

A Saner Middle Path

1. 1Prepare seriously, but cap the cost. No dropping years by default. No cutting off every hobby, sport, and friendship for a rank. Decide the ceiling in advance, together.
2. 2Protect one 'building' activity through the prep years — coding projects, writing, robotics, music, a small venture. This is the portfolio that matters after admission, and it is also insurance if the rank doesn't come.
3. 3Choose coaching that teaches concepts, not just patterns. Ask any institute: 'Show me how you teach a topic, not how you drill previous years' papers.' The difference is visible in one demo class.
4. 4Watch the child, not just the scores. Chronic anxiety, sleep loss, and a fading spark of curiosity are signs the trade has gone bad — no seat is worth a broken sixteen-year-old.
5. 5Have a genuine Plan B and say it out loud. A child who knows one exam will not define their life sleeps better and, paradoxically, usually performs better on that exam.

Questions Worth Asking as a Family

• Is this exam the child's ambition or our anxiety wearing the child's uniform?
• If we spend ₹5–8 lakh on coaching, what else could that capital do — and are we choosing coaching deliberately or by social default?
• What is our definition of 'it worked'? A rank? A capability? A child who still loves learning at 18?