AI Took the Parts of Your Job You Didn’t Notice Were Gone

“Which jobs has AI replaced?” is the wrong question. The right one is: which parts of this job has it already taken?

Very few jobs have disappeared outright. Pieces of almost every job have. The routine tasks go first, quietly, and most companies only notice when they’re making a headcount decision about the whole role.

That decision is happening everywhere. According to Challenger, Gray & Christmas, employers have cited AI for 120,136 announced US job cuts in 2026, about 21% of the total. Behind most of those numbers is a leader who had to make the call without a clear view of the work.

The decision nobody has a good tool for

When a founder weighs hiring a person against deploying an agent, the usual inputs are a vendor demo, a LinkedIn thread, and gut feel. None of them looks at what the role actually does all week.

So we built the tool we wanted, and opened it to everyone. You describe a role the way it really runs, and it returns a split, like 70/30 or 52/48, on a scale from Human through Both to Agent. You can see which parts of the week tools can handle and which still need a person.

It’s built for founders deciding on their next hire before they post the job or sign the contract, and for HR and TA leaders who need a defensible answer when the CFO asks why a role isn’t automated, or when the board asks why it was.

Most roles land somewhere in the middle. That’s the most useful answer you can get: hire the person, give them the tools, and know which part of the job is theirs.

Here’s what we won’t pretend: it isn’t finished.

Why we’re building it in the open

The answer to “human or agent” changes constantly. New agents ship every month. Tools that impress in a demo stall in production. A verdict that was right in spring can be wrong by fall.

So we keep refining how the Index weighs the evidence, and we’re doing it in public. A tool like this has no finish line, only better versions.

We want your two cents

If you think about this for a living, whether you invest in AI companies, build agents, lead talent teams, or design GTM systems, we’d like you to stress-test it. Run a role you know well. Tell us where the verdict is right, where it’s wrong, and what it’s missing.

Contributors whose input shapes the next version will be credited in our upcoming methodology piece here on HeadhuntEZ Insights.

Run any role through the Ask HeadhuntEZ at headhuntez.com, before you make the call.