As more organizations adopt AI, 59 percent of leaders that conduct some kind of training at those companies are still seeing significant disparities in their technicians’ AI skill levels. Traditional training programs are at fault: by focusing on prompt engineering, basic programming skills, and data interpretation alone, they treat skill gaps as an individual issue rather than a systemic one. As a result, the new skills that technicians gain often become ineffective.
For instance, these programs don’t always educate IT teams on what to do when the AI’s output is incorrect, which leads to inconsistent solutions and additional work. These initiatives also rarely train technicians on how to effectively hand off these AI outputs across teams.
Meanwhile, recent research revealed that 41 percent of IT executives believe past AI pilots failed because workflows were not ready for AI, while another 32 percent cited insufficient training or change-management support. Companies that layer AI initiatives on top of these bottlenecks only amplify the skills gap they set out to close.
Effective AI upskilling doesn’t just look like providing technicians with technical knowledge. Instead, it looks like building operational judgment and establishing a flexible, long-lasting learning environment to alleviate friction in IT team workflows.
Embedding AI Into Workflows
Broadly, successful AI upskilling initiatives occur when organizations provide their IT professionals with structured and unstructured training time to build their decision-making capabilities around AI. But this starts with leaders who investigate how their technicians work and where work breaks down.
For instance, 46 percent of employees spend more than 20 percent of their day on meta-work tasks, including chasing approvals, switching tools, and rework. These areas present opportunities for AI to step in and alleviate the cognitive load placed on employees.
Generally, a comprehensive map of where work breaks down should inform training programs. These chokepoints include transitions between teams, approval queues, product handoffs between teams like engineering and business stakeholders, and onboarding. Likewise, leadership should pay attention to other time-consuming work, such as ticket triage, escalation handling, and Tier-1 tasks for IT teams.
Building Operational Judgment
Once AI becomes part of their day-to-day work, technicians must learn how to work with AI in a management role. This involves understanding when they need to trust agents, set guidelines, or intervene.
In practice, an upskilled technician reviewing a ticket queue handled by AI is not simply checking the agent’s decisions. They are also stepping in when needed, flagging errors, and improving the system’s output over time. By developing these management skills, IT professionals can pass off time-consuming work to AI.
Designing clear roadmaps for how IT professionals can develop this judgment is also a core part of this training. Technicians should know how to set boundaries for AI, review audit trails, and correct the AI’s logic to improve its future behavior.
Build a Flexible, Ongoing Learning Environment
Embedding time to experiment with AI also makes these initiatives more productive. When technicians are given this extra time beyond a structured curriculum, they learn to take advantage of and trust the benefits that come with AI.
This training can look like simulating incidents, leading peer-led walkthroughs, or establishing an open, communal environment around AI. As opposed to relying only on rigid learning modules, leadership should encourage employees from a variety of teams to share their prompt libraries, ask their peers questions on AI usage, or troubleshoot errors communally.
Experimenting in this way decreases the friction of implementing new tech in an organization. As technicians figure out how they can work with AI, workflow pain points are easier to surface, and AI is more likely to become a natural part of their workflow.
Looking Ahead
The end goal of developing these AI skills should be to help technicians work smarter, not harder, and allow IT professionals to allocate their creative thinking and judgment to higher-impact work. When AI acts as a translation layer, pre-screener, search engine, and routine task automator rather than simply a polish layer, IT professionals get time and energy back. As a result, they begin to see AI as a trusted teammate rather than just another tool to manage.
That trust is foundational for the long-term success of these projects. Ultimately, the organizations that meaningfully decrease the AI skills gap are the ones that familiarize their technicians with new tech without asking them to overhaul their methods for doing work. In this human-centered strategy, technicians become active collaborators with these agents, as AI makes their work easier.
