Companies are investing heavily in AI upskilling because they know employees need new skills. Yet Docebo’s 2026 AI Readiness Gap research found that 85% of employees say the AI training they receive does not help them use AI in their actual roles. That tells me companies may be starting with the wrong question. Before asking what employees need to learn about AI, leaders need to ask what could keep them from learning it. In my research on curiosity, I identified four factors that can inhibit people from questioning, exploring and challenging what they “know”: fear, assumptions, technology and environment. Those same barriers can interfere with AI upskilling before an employee ever enters a training session. You can have excellent instructors, expensive platforms and impressive technology, but if you ignore what is already getting in the way of learning, employees may finish the training without changing much about how they use AI.

Fear Can Undermine AI Upskilling Before Anyone Experiments

It is difficult to learn something when part of your attention is focused on protecting yourself. AI creates several reasons for employees to feel that way. Employees may worry about looking incompetent, making mistakes, falling behind or even what AI could mean for their future roles. Those fears can follow employees directly into AI training.

Fear changes the way people learn because they become more interested in avoiding mistakes than exploring possibilities. That can make an AI training session look successful from the outside. Employees attend, complete the exercises and say they understand. What leaders cannot see as easily is whether people felt safe enough to admit what they did not understand.

Before AI upskilling begins, ask employees what concerns them about using AI. Give them a way to answer privately if necessary. Find out what mistakes they believe would be costly and what they think AI means for their future roles. Leaders can also help by showing their own learning process. When a senior executive demonstrates an imperfect prompt, questions an AI response or admits that a tool produced something useless, employees get a very different message than when leaders act as though everyone should already know what they are doing.

Assumptions Can Make AI Upskilling Seem Irrelevant

Employees do not arrive at AI training without opinions about AI. They may already believe it is overhyped, too difficult, unreliable, threatening, useful only for technical employees or unrelated to their jobs. Those assumptions affect how much attention people give the training before it even starts.

A company might interpret employees’ hesitation to embrace AI as resistance when employees may simply have never been shown a relevant reason to use it. There can also be assumptions working against learning. An employee who tried an AI tool a year ago and received a terrible response may have decided AI is useless. Someone who watched a coworker produce impressive results may assume that person has technical abilities they do not have. Another employee may believe that becoming more efficient will simply lead to being assigned more work. You cannot correct those assumptions by adding more slides to a training program.

Start by learning what people already believe. Ask where they see AI helping their work, where they do not and what experiences have influenced their opinions. Then connect the training to problems they actually recognize. Instead of beginning with ten things an AI platform can do, you might begin with a task employees complain takes too long, information they struggle to organize or a question they repeatedly have trouble answering. Once people see a connection between the technology and something they care about, curiosity has a reason to take over.

Technology Can Work Against AI Upskilling

Sometimes technology gets in the way of learning about technology. AI makes that especially easy because it can produce an answer almost immediately. An employee can type a question, copy the response and move on without understanding why the response was good, what might be missing or whether a different approach would produce something better. That is usage, but I would not call it learning.

Simply tracking whether employees use AI tells leaders very little about whether people have changed the way they work or think. If training focuses too heavily on prompts, features and shortcuts, employees can learn how to operate the technology without learning how to work well with it.

Before training, leaders need to decide what effective AI use should look like. When should employees use it? When should they question it? What types of information should never be entered? When does a person need to verify a response? Which decisions still require deeper human analysis?

Employees also need enough space to struggle with the technology. If the trainer provides every prompt, example and answer, participants can become passive. Give people a problem and allow them to determine how AI might help. Let them compare different prompts, receive weak responses, ask why they were weak and improve them.

Environment Determines Whether AI Upskilling Reaches The Workplace

You can teach employees everything they need to know about AI on Tuesday and send them back Wednesday to an environment that makes using any of it nearly impossible. Maybe they have no time to practice. Their manager may want immediate results and have little patience for experimentation. Employees may hear executives talking about innovation while watching coworkers get criticized when something does not work. If nobody on the leadership team uses the technology publicly, training cannot compensate for that environment.

If you want employees to experiment with AI, they need time to do it. They need managers who ask what they are learning rather than simply whether they finished the course. They need opportunities to share something that did not work without worrying that failure will be held against them.

Leaders also need to model the behavior they want. If executives announce an AI initiative but continue working exactly as they always have, employees notice. If managers encourage AI use but never ask employees what they are discovering, the training quickly becomes another program people attended and forgot. The environment after training may be more important than the training itself because that is where new behaviors either become part of the work or disappear.

Start AI Upskilling Before The Course Starts

Companies need AI upskilling , and that need is going to grow as the technology changes. Yet the success of that investment depends on more than choosing the right platform or teaching employees the latest tools. Before the training begins, look at what employees fear, what they already assume, how they currently use technology and whether their work environment gives them permission and time to experiment. Those factors influence whether people ask questions, test ideas, admit confusion and apply what they learn. If you want AI upskilling to change performance, start by removing the barriers that can prevent people from becoming curious enough to learn in the first place.