Rahul Arora
Founder, Automate Basics
The AI skills gap, in numbers
Adoption is near-universal. Training isn't, and the gap is widening rather than closing. What the research actually says.
The story most coverage tells is about capability, what the models can do this month. The more useful story is about distribution: who has been taught to use them, and who hasn't.
Every figure below is from a named third party, with the source published in full on our claims and sources page. None of it is our data.
Adoption is effectively universal
McKinsey's State of AI survey reports that 88% of organisations now use AI in at least one business function, with 72% reporting generative AI use, up from 33% the prior year.
Translated into something useful: the tools are already on your desk. Whatever the argument about adoption was two years ago, it's over.
Training didn't follow
This is where it gets interesting.
Jobs for the Future surveyed more than 3,000 workers and found only 36% say they have the training and resources they need to use AI in their jobs: down from 45% in 2024.
Note the direction. Not "still catching up." Widening.
Resume Now's Bring Your Own AI report (n=1,020, May 2026) puts a sharper edge on it: 41% of workers say their employer has provided nothing: no tools, no training, no guidance. Only 19% report comprehensive AI training. More than three in four have used AI tools they found and signed up for themselves.
The Conference Board found in August 2026 that while 55% of workers surveyed regularly use AI, only one in three has had employer-provided AI training in the past six months.
And the training that exists mostly doesn't help
Docebo surveyed 2,000 employees and found 85% say the AI training they receive doesn't help them use AI in their actual role. One in five have had none at all.
That's the most damning figure in the set, because it isn't about budget. Organisations are spending on training that doesn't transfer. Generic tool overviews don't survive contact with a real job.
What the market pays for the gap
PwC's Global AI Jobs Barometer, built on analysis of more than a billion job advertisements across 27 countries, reports a 62% wage premium for roles requiring AI skills: up from 57% the previous year, and more than double the 25% measured two years before that.
The same analysis finds jobs requiring AI skills growing at 69% against 9% for the overall jobs market, roughly eight times faster.
Read those as a description of the market, not a promise about you. What they tell you is where demand is concentrating, not what any individual should expect.
The finding that matters most if you're starting from zero
A randomised experiment with 453 professionals on writing tasks measured a 40% reduction in task-completion time and an 18% improvement in output quality as judged by independent evaluators, with the quality gains concentrated among workers in the bottom half of the initial skill distribution.
Brynjolfsson and colleagues found the same shape studying customer-support agents: a 14–15% average productivity gain, rising to 34% for novices.
The pattern repeats across studies. The largest gains land on the people who were furthest behind.
That's an unusual and genuinely good property. It means the people with the most to gain from learning this are the people who currently feel least equipped to.
Every figure above, with its publisher, method and the date we last verified it, is published on our claims and sources page. These are third-party findings describing the market as those publishers measured it. They are not our results, not a description of our learners, and not a prediction about any individual.
Goes deeper in AIW-200
AI at Work
Your inbox, your meetings, your documents, your spreadsheets. The unglamorous eighty percent of a working week, rebuilt so it holds up when someone else depends on it.