Employers want applied AI skills. Candidates still don't know how to prove them

Demand for AI capability is rising across technical and non-technical roles. But for candidates, the harder question is which skills, frameworks and credentials will improve their prospects. 

AI skills are becoming a more visible feature of the job market. 

In Canada, the proportion of job postings mentioning artificial intelligence nearly doubled during 2025, reaching 5.9% by the end of November. References remained particularly common in data and software roles, but also appeared in 16% of marketing postings, 10% of business and finance postings and 7% of management postings. At the same time, 29% of Canadian workers reported using AI at work several times per week. 

The same pattern can be seen internationally. According to the World Economic Forum’s Future of Jobs Report 2025, 86% of surveyed employers expect AI and information-processing technologies to transform their businesses by 2030. AI and big data were also identified as the fastest-growing area of skills, while two-thirds of employers said they expected to hire people with specific AI capabilities. 

That creates an obvious pressure for job seekers to develop some level of AI knowledge. The difficulty is knowing what employers mean when they ask for it. 



In some roles, AI capability may simply mean being able to use generative AI tools effectively. In others, it may involve technical development, data management, risk assessment or responsibility for how AI is governed within an organization. 

Those are very different expectations, but they are often grouped together under the same broad language. 

Candidates are then left to choose between courses, certificates, professional certifications, technical qualifications and governance frameworks without always knowing which of them relates to the work they want to perform. 

The shift is becoming particularly visible at the beginning of people’s careers. 

A 2026 survey by the National Association of Colleges and Employers found that more than one-third of the entry-level jobs represented by responding employers required some form of AI skill. That was nearly three times the proportion recorded in the organization’s autumn 2025 survey. 

Among the 185 employers that took part in the spring update, 28% said they were looking for early-career candidates who could use AI in their work. Almost 60% were also giving interns projects involving AI tools or related skills. 

The survey suggests that employers are not necessarily asking graduates to become technical AI specialists. More often, they want people who can use the technology within an existing role and understand how it affects the work they are already doing. 

That makes AI capability a much broader requirement than technical development alone. Its value increasingly depends on how well someone can connect their AI knowledge to the demands of their profession. 

What counts as useful AI capability will therefore look different from one role to another.

Research is beginning to show that visible AI capability can affect how employers assess candidates. 

A 2026 experimental working paper involving 1,700 recruiters in the United Kingdom and United States asked participants to evaluate hypothetical applicants for software engineering, office administration and graphic design roles. 

Candidates whose profiles included AI skills were between 8 and 15 percentage points more likely to receive an interview invitation. The results varied by occupation and recruiter, but the findings indicate that AI knowledge is already influencing hiring decisions across more than just technical roles. 

A separate 2025 study examined approximately 11 million UK job vacancies and found that demand for AI roles rose by 21% between 2018 and 2023. Over the same period, references to university education requirements in those roles fell by 15%. 

The researchers also identified an estimated wage premium of 23% for AI skills, which was higher than the premium associated with most degree levels. 

Formal education and professional credentials still have value, but the research points towards a labour market in which employers are placing greater emphasis on what candidates can demonstrate.

There is no shortage of material available to someone who wants to learn about AI. The difficulty is working out which type of learning is relevant and how it connects to a particular career. 

Candidates are faced with university programmes, short courses, vendor training, professional certifications and a growing amount of free content. Much of it may be useful, but it is designed for very different purposes. 

A person searching for an AI qualification could be trying to improve how they use the technology in their current role, move into technical development or build a career in governance and risk. Without a clear target, it is easy to complete several courses while still struggling to explain what practical capability they have developed. 

Employers often add to that uncertainty by asking for AI literacy, fluency or experience without defining what those terms mean in the context of the role. Candidates are then left to judge what level of knowledge is expected and how they can show that it extends beyond course completion. 

The lack of clarity is not limited to the recruitment process. Training within workplaces has also struggled to keep pace with the speed of adoption. 

Research published by the Future Skills Centre surveyed more than 5,800 Canadians and found that almost three in ten employed respondents were already using AI tools at work. Among those users, 44% had received no formal training. 

Just over half of respondents also felt that their employers were not providing enough support with new technologies. Many were learning independently or beginning to use AI without structured guidance. 

The same pressure is visible globally. The World Economic Forum estimates that 59% of workers will require training by 2030, while 63% of surveyed employers identified skills gaps as a major barrier to business transformation. 

Candidates are being told that AI capability will improve their prospects, but the systems around them are still working out what that capability should involve and how people should develop it. 

Credentials can give candidates structure and a credible way to communicate what they have learned. They can be particularly useful for someone moving into a field where they do not yet have professional experience. 

Their value, however, depends on how closely the learning relates to the work the candidate wants to perform. 

A general AI course will not necessarily prepare someone to manage AI risk, just as studying a governance framework does not automatically mean they can apply it inside an organization. A credential may show that someone has completed training or passed an assessment, but employers will still want to understand what that person can do with the knowledge. 

This is why candidates are better served by starting with the role rather than the qualification. Once they understand the kind of work they want to pursue, they can judge whether a course develops relevant knowledge and gives them opportunities to apply it. 

In AI governance, that may involve assessing how an AI system is being used, identifying risks and responsibilities, or creating documentation such as a risk register or impact assessment. The aim is not simply to recognize governance terminology, but to understand how it shapes decisions inside a real organization. 

ISO/IEC 42001 can provide a useful structure for developing that understanding. The standard sets out requirements for establishing, implementing, maintaining and continually improving an Artificial Intelligence Management System, helping organizations manage AI through defined governance and management processes. 

For candidates interested in AI governance, implementation, compliance or audit, knowledge of the standard can therefore be valuable. But its usefulness comes from being able to connect its requirements to real responsibilities, controls and evidence rather than simply knowing that the framework exists. 

The most appropriate credential is not necessarily the one with the broadest title or the highest level. It is the one that helps a candidate become more capable of doing the work they are aiming for.

Candidates cannot make informed learning decisions if employers describe their requirements only as “AI literacy” or “AI experience.” 

A job posting becomes much more useful when it explains how AI relates to the role. Asking someone to support governance documentation, assess AI use cases or communicate risk gives candidates a clearer idea of the knowledge they will need. 

Training providers also need to be precise about what their programmes offer. Learners should be able to tell who a course is intended for, the level of knowledge it develops and the kind of credential they will receive. 

Greater clarity on both sides would make it easier for employers to identify suitable candidates and reduce the risk of learners investing in training that does not support their intended career. 

AI knowledge is becoming more valuable, but employers are not looking for exactly the same capability in every role. 

For candidates, the challenge is turning a broad interest in AI into knowledge they can explain and use. That starts with understanding the work they want to pursue, then identifying the skills and frameworks that relate to it. 

Completing more courses will not necessarily make that connection clearer. What carries greater weight is being able to show how the learning applies to a real role and how it would shape decisions inside an organization. 

For those interested in AI governance, ISO/IEC 42001 provides one useful route into that understanding. It gives candidates a structured way to explore how organizations assign responsibility, manage risk and establish oversight around AI. 

Our ISO/IEC 42001 resource hub brings together guides, videos and practical learning materials for anyone building their understanding of the standard and its place within AI governance.