Ask the internet whether AI is coming for your job and you get an average for your job title. That is the wrong unit. A title is a label attached to wildly different jobs — two people who both write “Operations Manager” on LinkedIn can share almost no daily work — and replacement doesn't happen to titles anyway. It happens to people whose particular contribution turns out to be easy to reproduce.
So this asks about you instead: how deep your skills go, how current they are, whether you've actually learned anything recently, how far your range extends, and how fluent you are with the tools that are changing your field. Five answers, one score.
What this is not. It cannot tell you how automatable your role is — nobody honestly can, and the tools claiming to are reading an average off your job title. It tells you how replaceable you are within whatever role you hold, which is the part you can actually do something about.
What the five inputs are doing
They aren't weighted equally. In descending order of how much they move your score:
Where the anxiety is, in numbers.In IIM Ahmedabad's 2024 study of Indian white-collar workers, 68% feared their role could be automated within five years — including among those who had already adopted AI tools and been trained on them. A LinkedIn survey published in January 2026 found 84% of Indian professionals felt unprepared to find a new job, while 72% were looking anyway.
Both numbers describe a feeling. Neither tells you where you stand, which is the gap this is for.
Reading your band
Low (0–34). Hard to replace on your own answers. Re-check yearly — this band decays quietly if you stop moving. Medium (35–64). Nothing is exposing you, but nothing is protecting you either. This is the band that becomes High after two uneventful years. High (65+). Most of what you do could be picked up by someone or something else without much of a gap. It is a portfolio problem rather than a verdict on your work, and it is the most fixable of the three.
A note on the arithmetic: the score is built from five inputs that each contribute a maximum, so a genuinely worst-case set of answers lands around 80 rather than 100. Treat it as a position on a scale, not a probability that you'll be replaced.
What to do about a score you don't like
The score names your most exposed area on purpose — one thing, not five. Attack that first; the others move more slowly and matter less. If the answer was AI fluency, the fix is unglamorous and fast: use the tools on real work in your own field until you know where they break, which is the knowledge that actually transfers. If it was depth, the fix is slower and involves saying no to breadth for a while.
Worth being clear about the limits of a five-question tool: this measures the part of your position you control. It says nothing about whether your employer is healthy, whether your pay has kept up, or how quickly you could move if you had to — those are separate questions with separate answers, and they interact. The longer read on checking your own AI risk goes deeper on the reasoning, and the Career Health Score puts replaceability next to the other five.