The quiet divide at work: people who use AI, and people who don't.

When "Daniel" opened his laptop on a Tuesday morning in March, he noticed something strange on his colleague Maya's screen. She wasn't writing the client memo. She was talking to it.

"It was like watching someone use a calculator for the first time while you're still doing long division," Daniel told us. He's 44, a senior associate at a Midwestern consulting firm, and until that morning, he had never opened ChatGPT, Claude, or any of the AI tools that have quietly become standard among a certain subset of his colleagues. "I thought I had a few more years. Turns out I had maybe six months."

Daniel is not unusual. According to the March 2026 St. Louis Fed Survey of Consumer Expectations, only 33% of US knowledge workers report using generative AI tools at work on a daily or near-daily basis. The other 67% are split between occasional users and people who have never opened the tools at all.

What's striking is not the number. It's the shape of the curve underneath it.

The 18 months that quietly rewrote middle-skill work

Between late 2024 and the start of 2026, something happened in professional knowledge-work that did not happen with any prior software wave — not cloud, not mobile, not even spreadsheets. The productivity gap between AI-fluent and non-AI-fluent workers stopped being marginal and started being categorical.

A 2025 study from MIT Sloan compared two groups of marketing analysts of equivalent seniority. The AI-fluent group completed standard analytical briefs in 41% less time, with 18% fewer revision cycles. They were also more likely to be promoted within twelve months — by a factor of 2.3.

"The hardest part isn't that AI is taking jobs. It's that it's invisibly redistributing them. A worker who uses AI well still has a job — they just have a different one, with different leverage, and they're harder to replace." — Dr. Erika Voss, labor economist, Brookings Institution

For workers in the second group — the 67% — the gap is widening without their full awareness. They are not being fired. They are simply being asked to do more, with the same tools they always used, while their AI-fluent peers ship more, faster, with fewer hours. That asymmetry tends to surface in performance reviews, not in headlines.

• 2.3x Likelihood of promotion within 12 months for AI-fluent knowledge workers vs. non-users, controlling for seniority. Source: MIT Sloan, 2025
• 41% Reduction in time-to-completion on standard analytical tasks among AI-fluent professionals. Source: MIT Sloan, 2025
• 67% Of US knowledge workers do not yet use generative AI tools at work on a daily basis. Source: St. Louis Federal Reserve, March 2026
• $3.70 Estimated return per $1 invested in structured AI training programs over 12 months. Source: Iternal Research, 2026

Why most professionals haven't started

Among the people we interviewed for this piece, none of them — not one — said they were ideologically opposed to AI. They weren't worried about ethics. They weren't worried about job loss in the abstract. They were worried about something far smaller and far more human: looking like a beginner.

"I'm 47," one director-level operations leader told us. "I have a team of nine. I do not want to be the person fumbling with prompts in front of a 28-year-old senior analyst who's been doing this since college."

This is the quiet divide. Not a technological one — a confidence one.

What the AI-fluent group is actually doing differently

We asked Maya — Daniel's colleague — what she does on her laptop that he doesn't. Her answer was, in some ways, anticlimactic.

She doesn't use AI for big, dramatic tasks. She uses it for the unglamorous middle layer of her day: summarizing meeting notes she half-listened to, rewriting the third paragraph of a draft that's almost right, asking it to find inconsistencies in a 40-page contract, translating jargon-heavy slides into plain English her client can actually read.

"None of it is impressive in isolation," she said. "But I get an hour back. Sometimes two. And then I use that hour to think about the harder problems my boss wants me to think about. That's the whole game."

The most underrated finding in our interviews was this: the productivity gap is not driven by people using AI for moonshots. It is driven by people using AI to recover small slices of time, hundreds of times a week.

What to do about it (if you're in the 67%)

The good news, if there is good news, is that the divide is not yet permanent. Most labor economists we spoke to said the catch-up window is 6 to 18 weeks, depending on role. The skills involved are not technically difficult — they are habits. The people pulling ahead are not smarter. They are practicing.

If you haven't started, the question is not whether to start. It's how to start without spending six weeks on YouTube watching tutorials that contradict each other.

That's the gap Iternal Academy was built to close. It's the platform 47,000+ professionals use to catch up to — and quietly pass — the AI-fluent half of their team. Short lessons, organized by what you actually do for work, updated weekly. The first step, though, isn't a course. It's knowing where you stand.