Why AI training needs domain knowledge and human judgment

The AI skills gap extends into domain knowledge and integration

An AI training budget can miss much of the adoption problem if it focuses on operating the tools. In CompTIA’s Workforce and Learning Trends 2026 study, companies identified insufficient AI-use skills and insufficient core domain skills as major adoption challenges at the same rate: 24% each. Insufficient integration skills followed at 21%. Employees therefore need to learn how to use AI, understand the work it supports, and know how it connects to existing systems.

These findings point to a broader business problem: organizations are introducing AI faster than employees develop the capabilities to work with it. In the World Economic Forum’s (WEF) 2025 Future of Jobs Survey, 63% of employers identified skills gaps as a major barrier to business transformation. Employers planned several responses, with 85% intending to upskill their workforce, 70% planning recruitment, and 50% expecting to move employees from shrinking roles into expanding ones. These approaches can work together.

Training is one response, and the organizations reporting on it have different commercial interests. CompTIA’s training and certification business benefits from demand for skills development; TalentLMS, whose workplace-learning research also informs this discussion, sells learning management software and benefits from companies investing in training. Their findings help distinguish the programs employers provide from the support employees experience. Managers need to examine both when deciding how to develop AI capabilities.

Training investment needs to translate into usable support

Employers are already committing resources to workforce development. CompTIA reports that 48% of businesses identify training as one of their largest planned investment areas. Its research also finds that 57% of respondents say their company is investing in a new training program covering all types of skills, while 50% report entirely new programs dedicated to AI skills. Another 48% plan to introduce AI modules into existing programs, and 11% report taking no related action at present.

But providing programs does not establish that employees are ready to use AI. Only 43% of respondents in CompTIA’s research believe their workforce is enthusiastic about adopting it. CompTIA identifies low confidence in developing skills, at 39%, and uncertainty about where to begin, at 35%, as factors behind limited enthusiasm. Employees therefore need a manageable starting point and a reason to believe they can make progress.

Getting started also requires time to learn. In TalentLMS’s research on learning and development in 2025, almost half of employees and human resources (HR) managers agree that their companies regard training time as time taken away from actual work. Yet 64% of companies intend to upskill and reskill employees: deepen their current capabilities or prepare them for different work. Those plans create conflicting expectations when the time needed to carry them out is treated as an interruption.

Employees already acquire skills through work, giving employers a practical way to connect learning with job requirements. In the TalentLMS research, 86% say they learn new skills through hands-on, on-the-job training. The finding concerns skills generally, so managers still need to identify which workplace activities develop the AI capabilities each role requires. Practical experience becomes more useful when employees understand what they should learn and can get help while applying it.

Whether employees receive that help is a separate management question. TalentLMS found that 83% of HR managers believe their company actively helps employees learn to use AI, compared with 64% of employees. This gap makes employee feedback important when assessing training provision. Managers should ask whether people understand their options, can access relevant learning, and know how it connects to their responsibilities.

Answering those questions takes more than counting available courses. CompTIA’s findings describe program decisions and perceived readiness, while TalentLMS’s findings address learning practices and how different respondents assess support. Each gives managers a reason to examine how training works inside their organization. The starting point is the capabilities employees need in their roles.

AI learning depends on roles, security needs, and human judgment

Role requirements take AI training beyond a shared introductory course. CompTIA identifies demand across technology and non-technology work, including customer-facing, operational, and professional responsibilities. The figures below use CompTIA’s groupings to show which roles drive AI-training demand. Leaders can use this range to start identifying needs across the organization.

Learning across these roles must connect to the work each group performs. A sales team, a legal team, and a software development team need capabilities that reflect their responsibilities, even when they share introductory material. Leaders should therefore establish the skills each affected role requires before assigning courses. Participation becomes useful when learning prepares people for the tasks and decisions they face.

Role requirements also sit within company-wide priorities, which vary with organizational size. CompTIA reports different priority orders for small and medium-sized enterprises (SMEs), large companies, and very large organizations. The rankings below place AI training alongside security, support, management, and integration within each category, giving leaders another basis for deciding which capabilities to develop first.

Security ranks highly because AI changes both defensive capabilities and exposure to attacks. CompTIA explains that AI expands what security operations can do while increasing the scale of threat vectors, the routes through which attacks can reach an organization. Cybersecurity skills become more important as a result. Employers also have a reason to build security awareness across the workforce alongside deeper specialist expertise.

Security awareness is part of the judgment employees need when working with AI. Professionals overseeing AI-generated outputs must assess whether the results are correct and meet the task’s requirements. Using a tool and evaluating its results require different capabilities, so learning and development (L&D) teams should include analytical reasoning and other human skills when preparing employees for AI-supported work.

WEF’s broader skills-demand findings show how much employers value these capabilities. It reports analytical thinking at 68%; resilience, flexibility, and agility at 67%; leadership and social influence at 61%; creative thinking at 57%; and motivation and self-awareness at 52%. These skills matter to AI-supported work because employees must assess results, adapt to changed responsibilities, and coordinate decisions with other people. Technical instruction needs to prepare them for this work alongside teaching tool operation.

As companies put AI to work, CompTIA also describes business attention shifting toward integration with technology systems and workflows, with learning moving beyond introductory concepts toward deeper AI fluency and emerging disciplines. AI fluency means understanding enough about the technology to use it effectively and judge its results in context. As implementation expands, training should give greater attention to interpersonal skills and judgment relative to technical instruction. Employees need to understand how AI changes their work and which decisions remain their responsibility.

Connect learning to work and career development

Those responsibilities provide the starting point for a skills-first strategy: organize development around the capabilities a changing role requires. Managers should establish role requirements, identify missing capabilities, and then select or design the learning. Each training decision then has a purpose. Starting with the course catalogue makes it harder to explain why a particular employee needs a particular course.

A clear purpose also helps employees connect present learning with future career opportunities. A career roadmap should link current responsibilities with the skills needed as AI changes the role, including opportunities to learn on the job. Employees can then see how practical experience contributes to their development. To build that roadmap, organizations need to consider AI’s longer-term effects on career growth when setting training priorities.

On-the-job learning makes the roadmap practical because employees already acquire skills through daily work. Building learning into workflows can encourage participation and improve outcomes, particularly when a capability takes practice to develop. Managers still need to specify what employees should learn and how the activity develops it. An ordinary work task becomes a deliberate learning opportunity when its objective, expected capability, and available support are clear.

Clear objectives also make it easier to treat learning time as part of the job. When a company asks employees to develop capabilities while treating training as a distraction, it gives them conflicting instructions about how to use their time. Leaders should recognize relevant learning as work that prepares people to perform their responsibilities. This matters especially for employees who need practical experience, a clear entry point, and support as they build confidence.

The organization must make this path understandable. Employees can take an active part in their development, but future role requirements and expectations about learning time depend on decisions across the company. Managers need to connect individual effort with those decisions so each employee can explain what they need to learn, why it matters, and how learning fits into their work.

HR, IT, and L&D must check how employees experience support

Checking that understanding means asking employees about their experience. The gap between HR managers’ and employees’ assessments of AI-learning support makes this a necessary part of evaluating what a company provides. Employees’ concerns about falling behind, lacking a clear starting point, and preparing for future roles should inform training decisions alongside assessments of required skills. Feedback can show where a program’s purpose is unclear to the people expected to use it.

Responding to those concerns requires HR, IT executives, and L&D teams to coordinate their work. IT needs to communicate changing technology requirements, while HR and L&D need to connect those requirements with role development and learning opportunities. Employee feedback should guide this work so programs stay relevant as AI adoption advances. Each function helps by making the next learning step clear and possible to take.

Shared responsibility also requires clear expectations for what training should contribute to AI strategy. A plan should identify the capabilities employees need to develop, explain how they will practise them, and provide a way to assess their application as roles change. Integrating learning into workflows can support that development; managers should assess its contribution through the skills employees acquire and use. Before approving a program, they should also establish who will check progress and who will act when employees remain uncertain or need further support.

Main highlights

• Broaden the skills assessment: Companies report gaps in AI-use skills and core domain skills at the same rate, 24%, with integration skills close behind at 21%. HR and IT need to assess all three when setting training priorities.
• Check whether support reaches employees: While 83% of HR managers say their company actively supports AI learning, 64% of employees agree. Employee feedback can establish whether people understand their options, can access relevant training, and have time to learn.
• Match training to role requirements: L&D teams need to connect AI learning to each role’s tasks, security responsibilities, and decisions. Include practice in assessing AI-generated results alongside instruction in using the tools.
• Build learning into career development: Managers need to link current responsibilities, future skill requirements, and practical learning opportunities in a clear career roadmap. Define learning objectives for workplace tasks and recognize the time spent developing those capabilities as work.
• Assign ownership for learning outcomes: HR, IT, and L&D need shared expectations for the capabilities training will develop and how employees will practise them. Assign responsibility for checking application on the job and addressing gaps in employee support.