AI Cuts Through the Skills-Gap Hype and Finds a Digital Weak Spot

Universities are routinely accused of producing graduates whose qualifications do not match the realities of work. Employers complain that new hires lack practical skills, policymakers warn of widening labour shortages, and higher education institutions face growing pressure to redesign programmes around fast-changing demand. However, a new study suggests that part of the apparent mismatch may be created by the way skills are measured.

The study, "Bridging the Education–Employment Gap: Linking Labour Market Needs with University Programmes Through AI-Assisted Insights and Micro-Credentials," published in the journal Education Sciences, is based on approximately 30,000 Latvian job advertisements, professional standards and university learning outcomes. The researchers developed an AI-assisted framework to test how closely education aligns with employment.

The findings complicate the familiar narrative of universities falling hopelessly behind industry. When skills were compared through strict one-to-one matching, the programmes examined appeared to meet only 40–53% of employer demand. When the researchers accounted for competencies that were functionally equivalent but described differently, alignment rose to 73–85%.

Digital competencies remained consistently under-represented, while employers identified a deeper weakness that no curriculum audit can fully capture: graduates may formally possess knowledge but still struggle to apply it with confidence in real professional settings.

The Biggest Skills Gap May Be Hiding in the Measurement

The study examined three occupations connected to Riga Technical University programmes: personnel specialist, finance manager and organisation manager. These represented short-cycle, bachelor-level and master-level education, allowing the researchers to compare skill alignment across different qualification levels.

The analytical model linked three separate layers:

• Job advertisements represented current employer demand.
• Professional standards captured officially defined occupational expectations.
• Programme learning outcomes represented what the university formally promised to teach.

At first, the results looked troubling. Exact matching suggested that the personnel-specialist programme covered 50% of the skills found in relevant job advertisements. Finance management reached 53%, while organisation management covered only 40%. But those figures depended on an assumption that the same skill would be expressed in the same way across all three systems. In practice, employers, regulators and universities use very different vocabularies.

A job advertisement may ask for "management skills", "working with data" or "customer orientation". A university may divide those broad expectations into planning, information analysis, team coordination and decision-making. Professional standards may describe the same competence at an even higher level of abstraction.

When the researchers applied composite matching, recognising that several specific learning outcomes could collectively cover a broader employer requirement, the picture changed sharply. Coverage rose to 73% for personnel specialists, 78% for finance managers and 85% for organisation managers. The strongest shift occurred in the organisation-manager programme. Strict matching initially identified 51 demanded skills as missing. After accounting for functional equivalence, only four were classified as genuine unmet gaps.

This is more than a technical correction. Skills-gap diagnoses can influence accreditation, public funding and programme closures. If institutions rely on narrow keyword matching, they may mistake differences in language for evidence of educational failure. The study thus offers an important warning about AI-assisted policy tools. Algorithms can process thousands of documents quickly, but a fast comparison is not necessarily an accurate one. Without contextual interpretation, automation can turn semantic differences into inflated claims of institutional underperformance.

The researchers avoid that trap by combining AI extraction with expert review. The model assists with scale; humans decide whether a proposed match makes conceptual sense. The balance is crucial if universities are to use labour-market analytics without allowing black-box systems to dictate curriculum policy.

Digital Skills Are the Gap That Refuses to Disappear

Once differences in terminology were taken into account, many apparent deficits narrowed. Digital competencies did not. Skills classified as "working with computers" accounted for just 5.66% of the personnel-specialist programme's formal competence profile, 5.56% in finance management and only 2.78% in organisation management.

The pattern was independently reinforced by industry interviews. Employers pointed to gaps in data analysis, Power BI, automation, artificial intelligence tools, digital recruitment systems and technology-supported decision-making. Universities may be better aligned with employers than conventional audits suggest, but they remain vulnerable in areas where technological change is moving fastest.

Part of the shortfall may be documentary. One programme director reported that AI-related capabilities were embedded across courses rather than listed as separate learning outcomes. In that case, the curriculum may contain more digital content than formal programme documents reveal. But that explanation only goes so far. Employers were not simply asking for better wording in university handbooks. They reported weaknesses in graduates' applied digital ability.

Revising learning outcomes can make existing content more visible, but it cannot substitute for stronger instruction, practice and assessment. Universities need to determine whether students merely encounter digital tools or learn to use them independently in realistic tasks.

For personnel specialists, this may mean using AI in recruitment, workforce analytics and employee-data management. For finance managers, it involves business intelligence, applied data analysis and automated reporting. For senior managers, digital competence increasingly includes understanding how AI and automation reshape organisational strategy rather than merely knowing how to operate software.

The implications extend beyond Latvia. Higher education systems worldwide face the same structural challenge: curriculum approval moves slowly, while workplace technologies change rapidly. Universities cannot rebuild degree programmes every time a new tool emerges, but neither can they treat digital capabilities as optional additions.

A more sustainable response would combine durable competencies, data reasoning, technological judgement, ethical awareness and adaptability, with regularly updated tools and applications. The goal should not be to train graduates for one software package, but to ensure they can work intelligently in technology-rich environments.

The Real Problem Begins After a Skill Appears in the Curriculum

The study's most revealing insight comes from employers' descriptions of what happens when graduates enter the workplace. Industry respondents identified what the authors call a pedagogical transfer gap: a competence may appear in a curriculum and even be formally assessed, yet graduates may struggle to use it under real conditions.

Employers reported shortcomings in initiative, confidence and applied judgement. Graduates might understand a concept but hesitate to make recommendations, solve unfamiliar problems or take responsibility for decisions. This exposes a limitation shared by nearly every curriculum-alignment system. Document analysis can show whether a programme mentions data analysis, strategic planning or communication. It cannot establish whether students have mastered those abilities.

A university can claim that graduates will "analyse information" or "manage teams", but the practical meaning of those statements depends on teaching and assessment. Were students required to work with incomplete data? Did they defend decisions in front of practitioners? Did they manage conflict, revise a failed plan or take responsibility for outcomes?

If not, a formally aligned curriculum may still produce underprepared graduates. This finding shifts the debate away from endless arguments over whether universities teach the "right" skills. The more important question may be whether students are required to apply those skills in demanding, authentic contexts.

Work-integrated learning, employer-led projects, simulations and problem-based assessments may therefore matter as much as curriculum content. Universities need mechanisms that test competence under uncertainty rather than rewarding only theoretical recall.

The study also notes that student backgrounds can complicate any institutional diagnosis. The organisation-manager programme showed the highest alignment and was viewed positively by its programme director, but many students were already practising managers. Their professional experience may compensate for weaknesses in formal teaching.

Micro-Credentials Can Fill Gaps, But They Cannot Repair Weak Degrees

The study presents micro-credentials as one possible response to fast-emerging technical gaps, but its conclusions are deliberately cautious. Employers were broadly supportive when micro-credentials focused on specific, measurable competencies. Suggested areas included AI-assisted recruitment, data analysis, business intelligence, management accounting and digital-transformation strategy.

They were far less enthusiastic about generic credentials in leadership, communication or other loosely defined soft skills. Employers also stressed that micro-credentials require external quality standards and credible assessment. A university-issued digital badge carries little labour-market value if employers cannot determine what the learner actually demonstrated.

The study rejects the idea that micro-credentials should replace degrees. Their strongest role is in continuing professional development, specialist upskilling and targeted responses to new technical demand. Micro-credentials can help universities react more quickly than conventional programme-review cycles. They can offer working professionals a focused route into new technologies without requiring another full qualification.

However, they also carry risks. Breaking education into narrow skill units can fragment learning, weaken disciplinary understanding and encourage universities to chase short-lived hiring trends. A labour market that demands one tool today may demand another within two years. The best use of micro-credentials is therefore additive rather than substitutive. Degrees should continue to provide intellectual foundations, professional judgement and broad adaptability. Short credentials can then address specific gaps that emerge as industries change.

The study cautions policymakers against treating any single skills-gap estimate as definitive. Universities should adopt continuous curriculum monitoring instead of waiting for periodic crises to trigger reform, while employers need to engage consistently in programme design, teaching and assessment rather than limiting their role to complaints about graduate readiness.