AI Just Took My Colleague's Job. Here's What Nobody Is Telling You.

Last Tuesday, my colleague Priya didn't get a meeting invite. By Thursday, her badge stopped working.

The official line was "restructuring to align with our AI-first strategy." That's the phrase. Write it down somewhere. You'll hear it again.

I worked next to Priya for four years. She wrote the internal style guide every junior writer on our team used. She onboarded three of us. She knew which clients hated em-dashes and which ones loved them. Last week, a generative model started doing what she did. Worse, actually — but cheaper. And in the spreadsheet, "worse but cheaper" wins.

The story they sell

You've heard the pitch. AI is a "co-pilot." It "frees you up for higher-value work." It "augments human creativity." Pick your favorite verb. Every one of them was carefully chosen by a comms team that knows exactly what the real story is.

The real story is this: someone at the top of the org chart looked at a line item and asked whether a model could do most of it for a fraction of the price. The answer was yes — not all of it, not well, but enough. The line item moved. The person attached to the line item did not move with it.

The model doesn't have to be good. It has to be cheap enough that "good enough" becomes the new standard.

What "AI-first" actually means

When a company says it is "AI-first," it is not saying it loves technology. Every company already loves technology. What it is really saying is something more specific: it has decided that the cost of a worse output is lower than the cost of a human salary plus benefits plus PTO plus the small inconvenience of a person who occasionally has feelings.

That is a defensible business decision. I'm not naive about it. What I object to is the rebranding. Call it what it is. It isn't a transformation. It's a layoff with a press release attached.

The thing nobody is saying out loud

Here's what I haven't seen on a single panel, podcast, or LinkedIn post about the AI revolution:

The people losing these jobs were good at them. They were not redundant. They were not obsolete. They were not "low-skill." Priya could do something the model genuinely cannot do — hold a brand voice across a three-year relationship with a client who has changed CMOs twice. The model can fake it for a quarter. By then, Priya will be looking for work, and the client won't notice the slide for another six months. By the time they do, the budget will already be allocated.

It's not that AI is taking over because it's better. It's taking over because it's faster to deploy than to fire a person and slower to fire than a person. That's it. That's the whole magic trick.

What I'm doing about it

I started a notebook again. A real one. I've been jotting down the things I do at work that genuinely require a person — the client read between the lines, the late-night judgment call, the email I rewrote four times because the tone wasn't right. I want a record of what I do that a model can't, because the day they tell me to "align with our AI-first strategy," I want to walk into the meeting with receipts.

I don't know if it'll save my job. Probably not. But at least when they hand me the press release, I'll know exactly which lines are lies.

A small request

If your company just announced an "AI-first strategy," ask one question in the next all-hands: "How many roles are we expecting to eliminate in the next twelve months?" Watch them squirm. The answer is the only number that matters, and it's the one number nobody on the keynote slide deck will show you.

Stay angry. Stay analog. Stay employable.