This chapter examines the role of a common skills language as a foundational component of skills-first labour markets. It explains how well-designed skill taxonomies enable skills to be clearly defined and consistently classified across occupations, qualifications and training systems, thereby reducing information frictions for workers, employers and policymakers. Drawing on international experience, the chapter reviews different strategies to developing and updating skill taxonomies and assesses how they are embedded within occupational standards, qualification registries, and labour market information systems. It also highlights persistent challenges related to interoperability.
A Skills‑First Labour Market
Abstract
A common skills language is foundational to train and hire based on skills
A common skills language is key for skills-first labour markets because it allows workers, employers and training providers to communicate clearly about skills, thereby improving hiring, guidance and policy decisions. Skill taxonomies form the core of this language, providing structured definitions of skills and systematic links to occupations and qualifications.
Countries pursue different strategies to develop skill taxonomies, from simply adapting existing frameworks to creating new ones from scratch. In practice, these strategies are often combined and may include stakeholder consultations, structured surveys, expert judgement and data-driven approaches leveraging online job postings and AI. For example, Australia is developing its National Skills Taxonomy by combining extensive stakeholder consultation, expert-led conceptual work, and data- and AI-driven features to build a new, interoperable skills architecture intended to underpin a national shift toward skills-first policies.
Embedding skills into qualification registries and occupational standards strengthens transparency and allows for greater job mobility. Yet, progress on interoperability remains uneven, particularly in higher education, in federal systems, and in the development of user-friendly platforms that make explicit links between skills, learning and jobs. Such fragmentation limits the effectiveness of skills taxonomies, creating frictions for learners, employers and training providers and weakening the scalability of a common skills language.
Firms can use a common skills language internally to redesign roles, support reskilling and enable more flexible internal labour markets, but recruitment and progression practices must align with skills-first principles to realise these benefits.
Skills assessment and anticipation (SAA) exercises are most effective when they focus explicitly on skills rather than occupations alone, using a shared skills language to detect emerging needs, guide training provision and support smoother labour market transitions. Singapore’s Skills Demand for the Future Economy report is an example of how SAA exercises can be designed explicitly around skills. Drawing upon large‑scale job-posting data to track changes in job requirements, the report identifies priority skills and skill clusters across growing sectors, such as the digital, green and care economies.
The transition towards skills-first labour markets requires a common operational framework for defining and classifying skills. Actors must be able to communicate using a shared skills language: a common set of skill concepts and definitions that are interpreted consistently across workers, employers, training providers and public authorities. This shared language should serve not only as a medium for human communication, but also as a dictionary of skills and related concepts that can be understood by electronic systems, enabling search, matching and mapping across platforms and datasets. Without a common reference framework, skills remain difficult to signal, recognise and compare. Workers face challenges when describing their capabilities in ways that employers can interpret reliably; employers continue to rely on indirect proxies such as job titles, years of experience or credentials to infer what a person can do; and education and training providers face persistent barriers when attempting to make learning outcomes legible and comparable across programmes and institutions.
A central challenge is that skills information is currently generated and used in fragmented ways. Different actors may describe similar skills using different terms or use identical labels to refer to distinct underlying competencies. These inconsistencies generate disconnected understandings that impede alignment between hiring, training and guidance in a rapidly changing labour market. Fragmentation also limits interoperability between occupational frameworks, qualification registries and labour market intelligence sources (Figure 2.1). Where skills cannot be expressed as standardised, discrete concepts that travel across these systems, information about what work requires and what learning delivers remains difficult to connect. Policy interventions addressing emerging skills needs then become less targeted and more difficult to implement at scale.
Source: Authors’ elaboration.
Mitigating these frictions requires more than compiling a list of skills. A shared skills language must be sufficiently well-defined and structured to be usable across diverse contexts, while remaining flexible enough to evolve as jobs and tasks change. This creates practical design and implementation trade‑offs: deciding the appropriate level of granularity, avoiding duplication and ambiguity in concepts, aligning with existing occupational and qualification frameworks, and establishing governance arrangements for continuous updating. Finally, the effectiveness of a common language depends on its uptake: the language must be embedded into institutional and organisational practices – such as curricula, learning outcomes, job descriptions, talent management systems, and guidance tools – if it is to shift behaviour away from credential- and title‑based systems toward skills-first approaches.
The operationalisation of skills-first approaches requires some sort of skills language: that is, a lexicon of clearly defined skill concepts. While, in practice, organisations can decide to create or adopt whatever skills language they want, using a skills language that is shared among local labour market actors – employers, workers, training providers, governments and researchers – allows them to refer to mutually comprehensible terms in describing job requirements, learning content and individuals’ capabilities. The development of this language requires a workable definition of what constitutes a skill, a standardised inventory of discrete skill concepts, and an organising structure that makes the lexicon navigable and usable across employment, education and policy domains.
Defining skills
Skills represent the core analytical unit within a skills language, though the variety of human activities and abilities that can be labelled as “skills” makes it challenging to establish a precise and universally accepted definition. Conceptual disagreement persists across academic and policy communities. One influential reference point is the knowledge, skills and abilities (KSAs) model advanced by Stevens and Campion (1994[1]) and subsequently incorporated into numerous applied systems, including the Occupational Information Network (O*NET) developed under the United States Department of Labor. Within this framework, abilities refer to innate or acquired qualities that enable an individual to perform specific tasks, knowledge denotes the understanding and information accumulated through learning or experience, and skills capture the practical application of knowledge and abilities in the effective execution of tasks. A comparable tripartite structure appears in the 2018 Council of the European Union Recommendation on Key Competences for Lifelong Learning, which distinguishes attitudes as dispositions shaping responses to situations, knowledge as established concepts and theories underpinning understanding, and skills as the capacity to carry out processes and apply knowledge to achieve defined outcomes (Council of the European Union, 2018[2]).
At the national level, definitions continue to vary in scope and emphasis. For example, Jobs and Skills Australia describes a skill as a valued and purpose‑driven human ability that is acquired or refined through learning and practice, shaped by knowledge, experience and personal attributes, and influenced by context and environmental demands (Jobs and Skills Australia, 2025[3]). By contrast, the United Kingdom Standard Skills Classification adopts a broader formulation, defining a skill as “a capability enabling the competent performance of a task”, thereby framing skills as means to an end rather than outcomes in themselves (United Kingdom Department for Education, 2023[4]).
Despite differences in terminology, most definitions converge on several core elements. Skills are generally treated as individual attributes that are developed through learning or experience and are closely connected to the execution of tasks in real world environments. In policy discourse, the term “skills” is also frequently used as an umbrella label encompassing a broader range of related constructs such as abilities, attitudes and knowledge, or in some cases as shorthand for the tasks themselves. Indeed, the concept of skills is often subdivided into various components; common distinctions separate knowledge and technical information (e.g. speaking a language), practical competencies (e.g. operating machinery), and behaviours or orientations that shape performance and interpersonal interaction (e.g. teamwork).
For instance, the OECD Skills Profiling Tool distinguishes between occupation-specific skills (including hard technical skills and some occupation-specific soft skills), foundational information-processing skills such as literacy, numeracy and digital skills, and transversal non-cognitive skills associated with personality and attitudes (Tuccio et al., 2023[5]). Other examples of such decomposition include Skills England’s occupational profiles, which define occupations in terms of knowledge, skills and behaviours. Canada’s Skills and Competencies Taxonomy identifies skills alongside knowledge, abilities and other concepts, including personal attributes (Government of Canada, n.d.[6]), and LinkedIn’s Development Data Partnership distinguishes transversal, tech, business, disruptive and specialised industry skill groups (OECD, 2025[7]). Although labels and logics may differ, the shared objective is to define components with sufficiently clear conceptual boundaries that they can express, in a consistent way, the requirements of jobs and the learning outcomes of education and training programmes.
Collecting skills information
High-quality information on skills and skill requirements is a necessary input for any skills language. Three main methods are commonly used to build or expand a skills language, and they are typically combined rather than applied in isolation. Consultative processes remain central in systems built around national qualifications frameworks or established occupational classifications. The Australian National Skills Taxonomy (NST), for example, is rooted in extensive engagement with individuals, unions, employers, industry bodies, tertiary education providers and government authorities. Consultations with these stakeholders are used to clarify use cases, test appropriate levels of granularity, decide alignment with existing standards such as the Occupation Standard Classification for Australia (OSCA) and the Australian Qualifications Framework (AQF), and validate initial skill lists (Jobs and Skills Australia, 2025[3]; 2024[8]).
Structured surveys and expert ratings provide a more systematised approach that supports statistical representativeness and comparability across occupations. The United States’ O*NET illustrates this model through detailed questionnaires capturing the level and importance of hundreds of skills, knowledge areas, abilities and tasks. These surveys are administered to a statistically representative sample of firms and combined with structured ratings produced by trained occupational analysts. This produces a standardised dataset that can be aggregated into a hierarchical taxonomy (United States Department of Labor, n.d.[9]). While resource‑intensive, this approach relies upon representative survey instruments and can provide stable time series and consistent occupational coverage.
Machine learning and AI tools are also increasingly used to build and update skills languages (see, for example, Gallagher et al. (2022[10]), Popov, Snelson and Baily (2022[11]), and Keevy et al. (2026[12])). Indeed, advances in analytical methods (such as large language models) have enabled data mining of online sources – including job postings, course descriptions, training packages, professional standards and CVs – to extract skills information. An advantage of this inductive process is that these methods do not rely on predetermined lists of skills and can therefore detect new skills that had previously been unobserved. In this way, such approaches can respond more quickly to emerging skills than traditional consultation – or survey-based methods, but they also present risks: they can amplify noise and redundancy and over-represent skills frequently mentioned in vacancies that are not central to performance. Expert validation remains necessary to check sources, design analytical procedures and validate outputs.
Many existing systems across the OECD combine these approaches, using each method to address limitations in the others. Skills England’s development of the UK Standard Skills Classification, for example, combines and standardises skill and task statements from multiple sources, validates the resulting inventory against vacancy data, and then creates mappings between tasks, skills and knowledge concepts linked to occupations, courses and qualifications. Stakeholders are involved throughout to review results and improve usability (Skills England, 2025[13]; United Kingdom Department for Education, 2023[4]). Similarly, the European Skills, Competences, Qualifications and Occupations classification (ESCO) has introduced AI methods such as semantic similarity models to detect duplicate or obsolete concepts, while retaining human expert review and validation by member states’ working group processes (European Commission, 2024[14]). Hybrid approaches enhance coverage and responsiveness, while preserving the governance and validation needed for public legitimacy and operational uptake.
Some countries have simply adapted existing frameworks to establish their shared skills language. For instance, Canada’s Skills and Competencies Taxonomy draws on internal and international competency-based frameworks, including O*NET (Government of Canada, n.d.[6]). Japan’s Ministry of Health, Labour and Welfare created a “job tag” website based on data from the Japan Institute for Labour Policy and Training, largely adapted from the O*NET’s skills taxonomies (OECD, 2022[15]). These approaches have the potential to reduce development costs and accelerate implementation, but they still require efforts to ensure that skill concepts, occupational mappings and governance arrangements fit domestic institutions and policy objectives.
Structuring a skills language
An organised structure makes a skill language navigable, often through a hierarchy that groups skills at different levels of specificity. In many cases, the structure of a skill language is simply inherited from occupational standards or industrial classifications, which have their own hierarchies and which define occupations in terms of their skill requirements. For example, France’s ROME 4.0 is primarily an occupational framework organised around professional domains, but it also includes an embedded skill classification (“arborescence des compétences”). This skill classification is organised into a four‑level hierarchy comprising 6 competency domains, 32 strategic skill challenges, 84 competency objectives and 507 macro-competencies, which are linked to detailed inventories of practical skills, knowledge and work-related behavioural competencies, as well as standardised classifications of work contexts (France Travail, 2025[16]). Similarly, the Singapore Skills Framework classifies skills according to roughly 1 900 job roles in 39 main industrial sectors (SkillsFuture Singapore, 2025[17]). These designs emphasise usability for public employment services and related actors by nesting skill concepts within recognisable occupational structures.
By contrast, other countries construct standalone skills taxonomies that do not inherit their structure from occupational classifications, while remaining mappable to occupations and learning programmes. Examples include the United Kingdom’s Standard Skills Classification or Australia’s National Skills Taxonomy, which is currently under development. The World Economic Forum is also establishing a Global Skills Taxonomy, aimed at serving as a common skills language for its network of partners and external users, whereas the ILO and OECD are carrying out a similar exercise for G20 countries.
Firms may describe their job profiles using any internal skills terminology. However, at the national level, systematic linkage between occupations and skills typically relies on occupational standards. An occupational standard is a formal account of an occupation that specifies core duties together with the skills required for competent performance. Occupational standards are technically distinct from skills taxonomies: while skills taxonomies are structured frameworks that classify and organise skills, occupational standards are descriptions of occupations. It is when occupational standards are themselves described in terms of their constituent skills that similarities between occupational standards and skills taxonomies emerge.
Indeed, structuring an occupational standard around clearly defined skill requirements increases the utility of the standard. To be useful infrastructure for skills matching and labour market intelligence, occupational standards need to express requirements through structured skill information rather than narrative descriptions of tasks and responsibilities. Whereas tasks tend to describe what a worker does in a specific job context (for example, “prepare monthly management accounts” or “operate a milling machine”), skills describe transferable capabilities that can be applied across contexts (for example “financial reporting” or “operate specialised machinery”). Structured approaches should indeed emphasise portable capabilities expressed through a consistent vocabulary and stable identifiers (OECD, 2024[18]).
Table 2.1 presents selected OECD Member countries whose occupational standards express skill requirements through structured formats. France offers a particularly illustrative example. Its occupational standard system ROME – managed by the Public Employment Service, France Travail – has undergone several revisions over time, culminating in the release of ROME 4.0 in 2023. This latest iteration reorganised occupational descriptions around a more explicit skills structure and hierarchy intended to strengthen job matching and labour market intelligence. The reform moved away from using tasks as proxies of skills, instead adopting stand-alone definitions of skills and establishing a clearer distinction between context-specific activities and more broadly applicable skills. Another relevant example comes from the United Kingdom, presented in Box 2.1.
Note: Links accessed on 06 March 2026.
Usability of standards also improves when they specify the relative importance of skills within each occupation. For example, Flanders’ Competent framework differentiates “essential” from “optional” skills and knowledge, thereby signalling typical employer expectations while accommodating variation (OECD, 2024[18]). This tiered approach is easy to interpret and to maintain over time. O*NET takes a more formalised, systematic approach, circulating questionnaires among industry representatives who explicitly score the importance of pre‑identified skills associated with each occupation (United States Department of Labor, n.d.[9]).
Occupational standards deliver the most value when they can be easily linked or mapped to other systems, including training standards and qualifications. In French-speaking Belgium, the SFMQ (Service Francophone des Métiers et des Qualifications) provides a shared reference that translates occupational requirements into learning outcomes, units and assessment profiles (OECD, 2024[18]). Additional mappings between national occupational standards and widely used international frameworks – such as O*NET or ESCO – can further improve their use, especially among researchers and labour market analysts. Even within countries, occupational standards sit alongside other classification systems used in administrative data, vacancy statistics, and education systems. Without reliable crosswalks, analysts struggle to connect skills data to labour market outcomes; interoperability depends on maintaining accurate linkages across systems.
Occupational standards must be kept up to date to reflect evolving labour markets and technological change. Responsibility for updates and crosswalks should be explicitly embedded within governance arrangements. For instance, Flanders’ public employment service reviews standards four times per year, updating them when necessary. In French-speaking Belgium, stakeholders can request updates, with simplified procedures available for minor modifications.
The UK Standard Skills Classification Explorer illustrates the representation of occupations through structured skills information. More specifically, occupations are described in terms of their tasks, skill requirements, knowledge, relevant qualifications and even related occupations. For example, these are the descriptors for “Paper and wood machine operatives”:
Tasks, ranked by importance: e.g. “Monitor paper and wood processing operations to detect malfunctions and maintain standards”, “Inspect logs or timber to determine optimal cutting plans”, etc.
Skills, ranked by importance: e.g. “Read job sheets to determine setup and material requirements”, “Review technical specifications and drawings”, etc.
Knowledge areas: e.g. woodworking, machine tools, etc.
Qualification pathways: e.g. apprenticeships as Wood machinist, apprenticeships as Wood product manufacturing operative, etc.
Related occupations: Textile process operatives, print finishing and binding workers, Cabinet makers, etc.
Conceptually, the distinction between tasks and skills can at times be difficult to disentangle. Within the UK Standard Skills Classification, tasks are concrete work activities performed within a specific job context, whereas skills are defined as transferable capabilities that enable the performance of tasks across multiple roles and settings: tasks describe what is done in a particular workflow, while skills capture the underlying capability that enables not just a particular task, but also related activities in other contexts.
Firms aiming at adopting a skills-first approach for hiring and talent development also need to start the process by establishing an internal skill taxonomy. This process involves deconstructing job roles into clearly specified skills and communicating them explicitly to managers and employees. For instance, companies such as IBM have undertaken systematic reviews of job descriptions to identify the knowledge and skills required for each role before rewriting descriptions to emphasise skills rather than qualifications or tenure. This enabled the removal of degree requirements for many positions and created clearer pathways for workers to progress based on demonstrated skills (Burning Glass Institute, 2022[19]).
Other organisations have used skills intelligence tools to compare the skill profiles of declining and emerging roles and to design reskilling pathways based on overlapping skill sets. Unilever and Walmart, for example, have experimented with data-driven platforms that map roles to underlying skills and identify where targeted upskilling can support transitions (Accenture, 2021[20]). These exercises show that many workers already possess a substantial share of the skills required for new roles and can move into them with limited additional training. Firms that adopt such approaches gain more flexible internal labour markets and reduce recruitment costs.
A skills language can be applied at both the system and firm levels. When firms articulate job roles in skills terms, they help create demand for standardised skills descriptions across the wider education and labour-market system. Conversely, when formal qualifications, occupational frameworks and career guidance tools use a shared skills taxonomy, firms can integrate this information directly into their workforce planning, job design and talent management processes. Though a system- or national-level skills language may be broader than the specific descriptions used by individual firms, applying firm-level skills-first language in a way that maintains coherence with commonly used skills languages has the potential to unlock the network effects that make skill taxonomies valuable and expand the range of viable career pathways.
As with any language, the value of a skills language is strongly shaped by network effects. When use is fragmented or confined to isolated parts of the labour market, individuals and employers face frictions in interpreting skill requirements, and policymakers face reduced visibility over emerging skill needs. Therefore, although many skills languages originate from occupational standards, their integration with qualification registries is essential to maximise their relevance. Indeed, when qualification registries are fully articulated through consistent skills descriptors, users can see clearly how specific learning programmes build the skills required in the labour market. For learners, this means gaining clarity over the competencies acquired and the professional trajectories associated with their training; for employers, this implies a more granular understanding of the capabilities of prospective workers.
Across the OECD, there are several examples of countries whose national qualifications – at least for vocational and professional credentials – are explicitly linked to structured skill information (Table 2.2). For example, Skills England’s Occupational Maps rely on its Standard Skills Classification to chart out the skills required for each occupation, the technical qualifications that align with it, and the career progression routes available to those already in the occupation. The maps also reveal where skills overlap across occupations, enabling individuals and guidance practitioners to identify adjacent roles requiring similar skills. Users of Italy’s National Catalogue of Qualifications (Repertorio Nazionale della Qualificazioni) can use the Atlas of Occupations and Qualifications (Atlante del Lavoro e delle Qualificazioni) to examine how qualifications are described in terms of skills and associated work activities or learning outcomes. Presenting qualifications in this structured way simplifies the comparison between the qualifications defined by Italy’s regions and provinces.
Note: Links accessed on 06 March 2026.
Some registries also direct users to training institutions that deliver programmes associated with specific qualifications. Spain’s Catalogue of Professional Certificates (Catálogo de Certificados Profesionales), Australia’s National Training Register, and Skills England’s Apprenticeship Finder include search functions that allow prospective learners to locate nearby providers offering the qualification in question.
A shared skills language also offers a basis for describing non-formal and informal adult learning. To do so credibly, training providers and validation bodies should derive learning outcomes and skills descriptors from training and occupational standards rather than using provider-specific labels. In this way, learning outcomes can be bundled into units that can be combined, recognised and assessed consistently, allowing to stack towards larger qualifications (OECD, 2024[18]). Describing the learning outcomes of non-formal and informal adult learning using a common skills language can reduce fragmentation and make skills development more transparent for workers and employers.
Despite progress, systems that aim to link qualifications, occupations and skills continue to face structural constraints. One limitation arises from uneven sectoral coverage. Many systems focus on vocational education and training (VET), but do not incorporate qualifications in higher education or non-formal adult learning. This is, for example, the case with Portugal’s National Catalogue of Qualifications (Catálogo Nacional de Qualificações) and Spain’s Catalogue of Professional Certificates (Catálogo de Certificados Profesionales). Higher education can be difficult to describe in terms of learning outcomes, as universities typically retain discipline‑based curricula and maintain considerable institutional autonomy. Moreover, outside certain fields of study (such as engineering or health), university degrees tend to be less closely linked to specific occupations. A second challenge concerns the development of integrated digital platforms that present clear and navigable connections between qualifications, skills and occupations. Even where mappings exist, fragmented interfaces reduce usability and limit the capacity of learners, employers and advisers to interpret complex relationships. An effective platform should enable users to explore all relevant linkages within a unified environment.
Skills intelligence has become a critical enabling condition for the adoption and diffusion of skills-first approaches in hiring, talent management and training systems. Employers, education providers and public authorities require granular and timely evidence on evolving skill needs in order to redesign recruitment practices, modernise workforce development strategies and align learning provision with emerging demand. High quality skills assessment and anticipation (SAA) exercises therefore function not only as analytical tools but also as institutional infrastructure that supports the transition from credential-based to skills-based labour market practices.
In particular, skills assessment and anticipation exercises are relevant to skills-first approaches only when they generate granular and timely intelligence on the skills required in the labour market rather than on occupations or industries alone. Their value lies in translating labour market signals into operational insights that employers, training providers and workforce planners can use to redesign recruitment criteria, competency frameworks and learning pathways. Exercises that integrate administrative data, vacancy analysis and structured employer inputs are better positioned to identify emerging skill clusters (that is, groups of interrelated skills that tend to co‑occur in certain occupations or training pathways), transferable skills, and areas where targeted upskilling can replace formal credential requirements.
Methodological design also determines whether SAAs can sustain skills-first implementation. Mixed models that combine quantitative forecasting with qualitative expert consultation and sector-based foresight may be more effective in capturing rapidly evolving technical and transversal skills (OECD, 2023[21]). Continuous data collection, iterative validation with industry actors and alignment with skills taxonomies allow findings to be translated directly into talent management systems and training standards. Without these features, outputs remain descriptive rather than operational and provide limited guidance for organisations seeking to adopt skills-first practices.
Even if dissemination practices are generally strong, with two‑thirds of countries maintaining dedicated labour-market information platforms that present SAA findings in accessible formats to a wide range of users, current approaches often fall short because analysis remains anchored in occupational classifications and qualification structures that mask within-job skill variation and delay the identification of emerging skills (Table 2.3). Fragmented taxonomies and inconsistent terminology across data systems reduce interoperability and make it difficult to connect labour market intelligence with recruitment tools or curriculum design. As a result, assessments often identify labour shortages without specifying the underlying skills that drive them, which constrains their usefulness for skills-first hiring, workforce planning and skill-based education reform.
Note: Rows plot countries according to their responses to the question, “What is the main focus of your skills anticipation exercises?”; columns plot countries according to their responses to the question, “To what extent are findings from SAAs disseminated to relevant stakeholders?”. OECD countries and jurisdictions with available data, as well as Singapore, are shown.
Source: OECD Trends in Adult Learning Policy Questionnaire 2025.
Singapore’s Skills Demand for the Future Economy (SDFE) report illustrates how SAA exercises can be designed explicitly around skills. The analysis – which draws on large‑scale job-posting data to track changes in job requirements – focuses on identifying priority skills and skill clusters across growing sectors, such as the digital, green and care economies, rather than solely projecting occupational demand. The report highlights the most in-demand skills and identifies transversal “critical core skills” that are relevant across occupations. The findings are publicly disseminated through the SDFE report and interactive dashboards on the Jobs-Skills portal, allowing individuals, employers and training providers to explore labour-market insights and plan training pathways (SkillsFuture Singapore, 2025[22]). By combining vacancy analytics with sectoral analysis and presenting results in terms of specific skills and transferable skill clusters, the SDFE functions as a skills-focused SAA system that supports a skills-first approach to workforce planning and lifelong learning.
The development of a common skills language is not a straightforward task, but a complex and long-term process. A central challenge is the persistent fragmentation of existing frameworks, which limits the scalability and comprehensiveness required for a fully functional skills taxonomy. Many countries have sought to merge or align pre‑existing frameworks as a first step, but differences in definitions, units of analysis and levels of granularity often limit the benefits of this approach. Strengthening interoperability within national systems should therefore be a priority.
Federal systems inherently have another layer of complexity. The Canadian case illustrates the challenges of working across multiple jurisdictions. Canada uses the National Occupational Classification (NOC) framework (complemented by an associated Skills and Competencies Taxonomy), but many occupations within this framework have distinct regulatory requirements between provinces, reducing the clarity of skill expectations at the national level (Forum of Labour Market Ministers, 2022[23]). On the education side, Canada has a pan-Canadian Degree Qualifications Framework focused on quality assurance in higher education, alongside several provincial qualification frameworks (for example, Ontario’s Qualifications Framework). Links between provincial qualifications frameworks are not always precisely articulated, nor are they linked to the Skills and Competencies Taxonomy (Canadian Information Centre for International Credentials, n.d.[24]).
Beyond national boundaries, increasing the international comparability of skills has become critical to support worker mobility, recognition processes and alignment of industrial strategies. Supranational initiatives such as the EU’s Skills Portability Initiative and the ongoing OECD-ILO Global Skill Taxonomy effort signal a growing recognition that cross-border consistency is required to facilitate worker mobility and streamline qualification recognition (see Box 2.2).
The Skills Portability Initiative (SPI) is a forthcoming European Commission proposal designed to make workers’ skills and qualifications more easily understood and recognised across all EU member states. It aims to reduce barriers that hinder labour mobility and the functioning of the single market by improving the transparency, comparability and digital usability of credentials, particularly for employers (including SMEs) that currently struggle to interpret qualifications obtained abroad.
SPI is structured around three interlinked actions:
1. Enhancing transparency and digitalisation of skills and qualifications so they are more portable and comprehensible across occupations and borders, building on interoperable tools such as qualifications registries referenced to the European Qualifications Framework (EQF), skills descriptions linked to ESCO, and digital sharing channels such as the EU Digital Identity Wallet.
2. Modernising and potentially expanding recognition procedures for regulated professions under existing EU frameworks to make cross-border recognition faster, more predictable and less administratively burdensome.
3. Simplifying qualification recognition for non-EU nationals, including by reducing fragmentation in assessment processes and improving the transparency of how foreign qualifications and skills are evaluated.
The initiative is being developed through a broad evidence‑gathering and consultation process involving employers, workers, national and regional authorities and civil society, including targeted interviews and expert workshops. If successfully implemented, SPI could reduce procedural friction in cross-border hiring, support career progression across member states, and help mitigate skill mismatches by making skills signals clearer and more usable in recruitment and recognition processes.
Technological innovation offers potential instruments to address interoperability gaps. For example, Leveraging AI for Skills Extraction & Research (LAiSER) is an open-source initiative led by George Washington University in the United States, that applies large language models to extract skill statements from job adverts and educational content, link them to a dynamic skills database, and update the taxonomy as terminology evolves (Learn & Work Ecosystem Library, 2025[25]).
References
[20] Accenture (2021), Future Skills Pilot Report, https://www.accenture.com/content/dam/accenture/final/accenture-com/document/Future-Skills-Pilot-Report.pdf (accessed on 1 April 2026).
[19] Burning Glass Institute (2022), The Emerging Degree Reset, Harvard Business School, https://www.burningglassinstitute.org/research/the-emerging-degree-reset (accessed on 1 April 2026).
[24] Canadian Information Centre for International Credentials (n.d.), Provincial and territorial qualifications frameworks, https://www.cicic.ca/1287/provincial_and_territorial_qualifications_frameworks.canada (accessed on 17 December 2025).
[2] Council of the European Union (2018), Council Recommendation of 22 May on key competences for lifelong learning, https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=oj:JOC_2018_189_R_0001 (accessed on 1 April 2026).
[14] European Commission (2024), ESCO v1.2 is live!, https://esco.ec.europa.eu/en/news/esco-v12-live (accessed on 17 December 2025).
[23] Forum of Labour Market Ministers (2022), A Pan-Canadian Framework for the Assessment and Recognition of Foreign Qualifications, Employment and Social Development Canada, https://www.canada.ca/en/employment-social-development/programs/foreign-credential-recognition/funding-framework.html (accessed on 17 December 2025).
[16] France Travail (2025), Répertoire Opérationnel des Métiers et Emplois, https://www.francetravail.org/opendata/repertoire-operationnel-des-meti.html?type=article (accessed on 17 December 2025).
[10] Gallagher, E. et al. (2022), A New Approach to Building a Skills Taxonomy, Economic Statistics Centre of Excellence & UK Office for National Statistics, https://escoe-website.s3.amazonaws.com/wp-content/uploads/2022/05/11111940/ESCoE-TR-16.pdf (accessed on 10 April 2026).
[6] Government of Canada (n.d.), Skills and Competencies Taxonomy, https://noc.esdc.gc.ca/SkillsTaxonomy/SkillsTaxonomyWelcome (accessed on 17 December 2025).
[3] Jobs and Skills Australia (2025), NST Update: building a system that puts people and skills first, Australian Government, https://www.jobsandskills.gov.au/sites/default/files/2025-10/nst_update_-_building_a_system_that_puts_people_and_skills_first.pdf (accessed on 10 April 2026).
[8] Jobs and Skills Australia (2024), National Skills Taxonomy Discussion Paper: To inform the design of a National Skills Taxonomy, https://www.jobsandskills.gov.au/sites/default/files/2024-06/national_skills_taxonomy_discussion_paper.pdf (accessed on 10 April 2026).
[12] Keevy, J. et al. (2026), “Beyond Qualifications Frameworks: Large Language Models and the Future of Global Skills Recognition”, European Journal of Open, Distance and E-Learning, Vol. 28/1, https://doi.org/10.65043/eurodl.162.
[25] Learn & Work Ecosystem Library (2025), Leveraging AI for Skills Extraction & Research (LAiSER) – George Washington University, https://learnworkecosystemlibrary.com/initiatives/leveraging-ai-for-skills-extraction-research-laiser-george-washington-university/ (accessed on 17 December 2025).
[7] OECD (2025), Empowering the Workforce in the Context of a Skills-First Approach, OECD Skills Studies, OECD Publishing, Paris, https://doi.org/10.1787/345b6528-en.
[18] OECD (2024), Agile Occupational and Training Standards for Responsive Skills Policies, Getting Skills Right, OECD Publishing, Paris, https://doi.org/10.1787/bacb5e4a-en.
[21] OECD (2023), Assessing and Anticipating Skills for the Green Transition: Unlocking Talent for a Sustainable Future, Getting Skills Right, OECD Publishing, Paris, https://doi.org/10.1787/28fa0bb5-en.
[15] OECD (2022), The New Workplace in Japan: Skills for a Strong Recovery, Getting Skills Right, OECD Publishing, Paris, https://doi.org/10.1787/7c897f52-en.
[11] Popov, D., S. Snelson and T. Baily (2022), Review of skills taxonomies: Report prepared for the Skills and Productivity Board, Frontier Economics, https://assets.publishing.service.gov.uk/media/628cd6988fa8f55622a9c92a/Review_of_skills_taxonomies_report_prepared_for_the_SPB_May_2022.pdf (accessed on 10 April 2026).
[13] Skills England (2025), Research and analysis: Summary of methodology, https://www.gov.uk/government/publications/uk-standard-skills-classification-interim-development-report/summary-of-methodology (accessed on 16 December 2025).
[22] SkillsFuture Singapore (2025), Skills Demand for the Future Economy Report, https://jobsandskills.skillsfuture.gov.sg/insights/sdfe (accessed on 4 March 2026).
[17] SkillsFuture Singapore (2025), Skills Frameworks (SFw), https://jobsandskills.skillsfuture.gov.sg/frameworks/skills-frameworks (accessed on 17 December 2025).
[1] Stevens, M. and M. Campion (1994), “The knowledge, skill, and ability requirements for teamwork: Implications for human resource management”, Journal of Management, Vol. 20/2, pp. 503-530, https://doi.org/10.1016/0149-2063(94)90025-6.
[5] Tuccio, M. et al. (2023), The OECD Skills Profiling Tool: A new instrument to improve career decisions, OECD Publishing, Paris, https://doi.org/10.1787/598ff539-en.
[4] United Kingdom Department for Education (2023), A skills classification for the UK: Plans for development and maintenance, https://assets.publishing.service.gov.uk/media/652fdb9d92895c0010dcb9a5/A_skills_classification_for_the_UK.pdf (accessed on 1 June 2026).
[9] United States Department of Labor (n.d.), O*NET Data Collection Program, https://www.dol.gov/agencies/eta/onet/data-collection (accessed on 16 December 2025).