AI Training Doubled. The Time to Take It Did Not

Your daily brief on leadership and business

Companies have discovered AI training. Employees have discovered they have no time to take it. That, in short, is the finding of a new report based on a July 2026 survey of 1,000 full-time salaried professionals at US organizations with 5,000 or more employees.

The progress is real. The share of companies offering AI-specific training jumped from 25% in 2025 to 58% in 2026, and the share of employees using AI tools several times a week rose from 40% to 68%. Eighty percent of respondents say their company is on track for an AI-enabled future, up from 67%.

The missing ingredient is time

Yet 56% of employees say they have no time set aside during work hours to build skills, and 43% say relevant training materials are lacking. Most (84%) spend five hours or less per week on skill development. Nearly 60% say AI skills are weighted the same as other skills, or not considered at all, when assignments and promotions are handed out. Learning AI, in other words, often counts as a hobby rather than a career move.

"The fix belongs to the employer," says Taylor Sullivan, VP of product and assessment at Workera. Development, she argues, should be part of the role: resourced, rewarded and something managers are held accountable for. CEO Kian Katanforoosh adds that leaders who give people time to build AI skills will see the biggest returns.

Invisible skills, missed promotions

The career consequences show up clearly. Forty-three percent of respondents say they were passed over for an opportunity or promotion because someone misjudged their skills. Coaching is patchy: half say they get enough to advance, 27% receive regular coaching from a manager or executive, and 22% never receive individual coaching at all. Employees still trust human managers more than AI to evaluate them, though 41% would opt in to continuous skills measurement if they controlled their own data. Employees are also fairly confident about their place next to the machines: 76.6% say AI could not do their job better than they do, and 78.1% think it could handle 30% or less of their work today. The report comes from a company that sells skills assessment software, so its recommendations deserve a pinch of salt, but the time problem will sound familiar to almost anyone with a calendar.