Development of the Workforce Readiness Assessment Protocol (WRAP)

Introduction

The UK's Offshore Renewable Energy Catapult, one of eight research and technology centres managed under Innovate UK, embarked on a series of Workforce Foresighting (WF) studies between 2023 and 2026. These studies aimed to identify the skills and capabilities required to exploit emerging technologies in the renewable energy sector, with a focus on "Horizon 2" technologies (i.e. expected to reach broad-front deployment in around five to seven years). The WF programme sought to shift the industry's approach to workforce development from a lagging to a leading perspective, by analysing the organisational capabilities needed to deploy new technologies and then "rewinding" to seed the necessary education and training now.

The Catapult network did not initially have a remit to look at skills. The WF programme addressed this gap, examining the technologies likely to disrupt or accelerate industrial sectors and the capabilities needed to exploit them. While the quality and quantity of engagement and output varied across studies, the WF approach offered a unique lens: cross-industry and importantly, one based around the needs of tomorrow rather than the gaps of today.

Subsequent to the WF programme, we considered there was no simply presentable or metric output from workforce foresighting. The reports do not inform a reader in simple terms how ready the workforce is to meet a technology, nor the effort required to achieve that readiness. We began to compare with the Technology Readiness Level (TRL) scale used to assess technological maturity - a widely accepted framework for evaluating the readiness of technologies. A corresponding metric for workforce readiness is missing. This gap has significant implications for the effective deployment of emerging technologies and the development of the necessary skills and capabilities. Our subsequent research shows that while the concept of a “workforce readiness level” or WRL is not new, previous efforts have tended to focus on this in specific centres and (and a reductive level) tend to broadly equate to TRL bands, rather than considering the multifactored and dynamic journey of building workforce readiness.

A Missing Metric

Industrial strategy in advanced economies has become largely equivalent to technology strategy. Governments and sectors identify a set of technologies judged critical to future competitiveness (or occasionally to meeting policy commitments) and seek to organise investment, regulation and infrastructure around their deployment. Examples include the UK’s Modern Industrial Strategy, even as it aims to be place-based (Bailey et al., 2026), the CHIPS and Science Act in the US (Hogan, 2025) and similar semiconductor safeguarding approaches in South Korea and Japan, as well as the UK’s Offshore Wind Industrial Growth Plan (RenewableUK, 2024) . The vocabulary of this work is well established with roadmaps, pilot plants, demonstration projects, first-of-a-kind (FoaK) deployments, spin-outs and scale-ups.

The scale commonly used to place technologies along this journey from ideation to commercialisation is the Technology Readiness Level. Originally introduced within NASA in the 1970s and given a nine-level definition (Mankins, 1995), the TRL has become what a “discipline-independent figure of merit” for communicating technological maturity (Mankins, 2009). Its adoption well beyond aerospace, into defence, energy policy, research funding and corporate R&D portfolio management, reflects a genuine need met by the measure. TRLs moved discussions about whether a technology was "ready" from being highly subjective to being structured around key reference points.

Workforce planning has no equivalent. There is a substantial and often sophisticated literature forecasting skills demand, and the UK has produced influential work on the future composition of employment (Bakhshi et al., 2017). Sector bodies, government (local and national) and trade bodies often commission studies estimating how many workers of what type will be needed by a given date. What is absent is a shared, portable scale that allows a workforce to be located on a trajectory in the way a technology can be, and particularly a shared and portable view of how to move along that scale.

The consequence is an imbalance in how deployment risk is assessed. A programme board reviewing a major infrastructure or manufacturing commitment will typically have a view of the technological maturity of but it is much less likely to have a comparably structured view of workforce maturity: Whether the people needed to build, commission, operate and maintain it will exist in sufficient number and with sufficient competence at the point they are required. Where that question is addressed, it tends to be answered with a headcount forecast, which is a necessary input but not a readiness assessment. The forecast tells you how many people you will need. It does not tell you whether the system that produces those people can produce them.

The constraints on workforce supply are rarely reducible to a single number. Training capacity, the time required to reach genuine competence as distinct from formal qualification, the availability of adjacent skills that can be redirected, the responsiveness of curricula, the willingness of employers to invest where trained workers may be poached, and the presence or absence of institutions capable of coordinating across a fragmented supply chain all bear on whether a stated skills gap will actually close. These are heterogeneous factors, some quantifiable and some not, and they interact.

This paper (re)introduces the Workforce Readiness Level (WRL) as a nine-level scale describing the maturity of the workforce system in a given sector, region or organisation . The WRL is generated by a proprietary diagnostic instrument, the Workforce Readiness Compass, which assesses readiness across six dimensions drawn from established research traditions in labour economics, innovation studies, organisational learning and skills policy.

Three points about the construction of the WRL scale are worth highlighting at the outset:

1. First, the WRL is constructed as a companion to the TRL rather than a free-standing metric. Its principal diagnostic output is not the WRL value in isolation but the relationship between the WRL and the TRL of the technology in question. A technology at TRL 8, entering full commercial deployment, supported by a workforce system at WRL 4, where only pilot training provision exists, describes a specific and actionable problem. The same WRL 4 alongside a technology at TRL 4 describes a workforce system that is roughly keeping pace. The gap, not the level, carries the diagnostic weight.

2. Second, the framework does not claim predictive validity. It offers is a structured, theoretically grounded and – a principal aim, a repeatable – way of characterising the current state of a workforce system, of comparing that across sectors or over time, and of focusing attention and argument on the specific constraints most likely to act as a brake on maturity. It is an instrument intended to be diagnostic and to support dialogue and action.

3. Thirdly, other attempts to define workforce readiness level have been made and the term itself is not unique. Models developed at MIT (Aiken et al., 2024) and Mississippi State University (Smith et al., 2026) are considered in the following section. The distinctions we are keen to draw for this WRL tool are (1) that we aim establish whether the right conditions exist (and the difficulty of introducing the right conditions) for the workforce to move along a maturity path, and (2) we seek to ensure applicability across a wide range of sectors, and the ability to assess at macro- and micro-scales.

Existing Definitions of Workforce Readiness Level

Researchers at the Massachusetts Institute of Technology (MIT) defined workforce readiness levels for manufacturing in the US defence sector in the Workforce Readiness Level Deskbook (Aiken et al., 2024). This is a niche, single-sector application of a workforce readiness instrument, but it makes two useful observations. Firstly, by identifying the need to consider workforce readiness alongside technology readiness, and manufacturing readiness levels. Secondly, the Deskbook observes that the shape of the workforce changes as technology matures. Early phases rely on “creators”, later stages require “innovators”, and at high TRL the workforce needs “implementors”. However, a criticism we make of the MIT scale, is that the levels are defined by what the workforce mix looks like. We think this significantly limits the usefulness of the scale: To take an overly critical perspective, the WRL can - more or less - be read over from the TRL.

Another work has been produced more recently: (Smith et al., 2026) considered WRL as a ranked measure similar to TRL, but this time they addressed it to the level of an individual’s capability / comptentence, not of the organisation or sector more broadly. Their method includes a four-pillar rubric. The approach set out by Smith et. al does have some similarity with our approach (compared with the MIT work) in that they use multiple dimensions (four rather than six), and their scoring uses an approach which means that strength in one of the dimensions or pillars can’t hide weakness in another (similar to the geometric averaging that we propose).

We therefore consider that our proposed method, WRAP brings something unique to this discussion: It is broad-sector not single-sector and does not just read across from TRL (differentiating it from the MIT work), and it addresses entities at several levels up to state or sector (differentiating it from the Mississippi State work which considers the individual competency only), and finally – our method assesses the readiness of the workforce system as a whole, to adapt to technology disruptions.

Theoretical Grounding

This section outlines the six dimensions of the WRL, setting out the theoretical grounding for each. The aim has been to ensure a robust diagnostic capability for sectors, regions, and organisations.

Technology Maturity & Disruption (TMD)

The first dimension, Technology Maturity & Disruption (TMD) captures the degree of disruption that an emerging technology is expected to bring to the sector or entity under assessment. In application to workforce foresighting, TMD would be assessed as part of identifying or scoping the particular technology subject to foresighting.

Geels observes that technologies do not function in isolation, but they are embedded in "sociotechnical configurations" that include skills, regulations, and infrastructure (Geels, 2002). This means that transitioning or adapting to a new technology is not a simple substitution, but a "stepwise process of reconfiguration" involving multiple levels (the “Multi-Level Perspective”): Niches (where the innovation or “radical variety” is generated); Regimes (the selection and retention mechanisms), and Landscape (the wider external context). Geels also asserts that "technology, of itself, has no power, does nothing". It only fulfils functions when aligned with human agency and social structures .

This thinking helps us then to connect with Mankins' (1995) definition of TRL as a systematic measurement system for maturity. We suggest that the WRL framework acts as the "workforce development overlay" to the technical scale. As technology moves from TRL 1 ("basic principles observed") to TRL 9 ("flight proven"), the workforce must undergo a parallel development and increase in maturity. The disruption indicated by the TMD dimension as the shift in the sociotechnical regimes and landscape needed to allow innovation to mature from radical or niche, to broad-front deployment with impact across the landscape.

Geels later addressed criticism regarding a "lack of agency" by incorporating power and politics into his model (Geels, 2011). The WRL assessment framework therefore also addresses this, and we borrow from the idea of worker roles (the Creators, Innovators, Implementers) required to break these stabilities (Aiken et al., 2024).

Readiness for technological disruption is determined by how well the workforce can adapt or replace the "lock-in mechanisms" of existing systems - such as sunk investments in old competencies - to innovate and apply new configurations.

Workforce Supply-Demand Gap (WSD)

The Workforce Supply-Demand Gap (WSD) dimension measures the quantitative and qualitative mismatch between current labour availability and the specific tasks required by new technologies. It is grounded in task-based labour economics and skill-biased technological change. The underpinning theory shifts the focus from educational credentials to actual job content (Autor et al., 2002). They argue that computer-based technologies substitute for workers in routine tasks (rules-based activities) while complementing workers in non-routine cognitive tasks (problem-solving and complex communication).

Autor et al. observe that "computer capital (1) substitutes for workers in carrying out a limited and well-defined set of cognitive and manual activities... and (2) complements workers in carrying out activities involving non-routine problem solving". This creates a "demand shift" favouring educated labour, as "repetitive, predictable tasks are readily automated".

The WSD dimension uses this "machine’s-eye view" of work to identify where new technologies will create gaps. For emerging technologies, the WRL assesses whether the supply of "middle-skilled workers" (e.g., engineering technicians) can meet the increased demand for non-routine manual and cognitive tasks that cannot yet be codified.

While Autor et al. focus on computerization, Bakhshi et al. (2017) extend this using machine learning to identify "skill complementarities"—bundles of skills like "originality" and "social perceptiveness" that are increasingly valuable together. They argue that purely qualitative assessments of gaps are subject to "human biases," necessitating the mixed-methods approach integrated into the WRL.

Capability Investment Burden (CIB)

The Capability Investment Burden (CIB) dimension assesses the economic and organisational costs of upskilling. It draws from human capital theory, specifically the distinctions regarding training incentives and the "who-pays" question.

As noted in the WRL Deskbook (Aiken et al., 2024) and supporting literature, emerging technologies require significant "on-the-job training" and "specialized creator" inputs. A core tension exists between general skills (portable across firms) and firm-specific skills. In high-mobility environments, firms may under-invest in training for fear that competitors will "poach" their newly skilled staff, while scientists and engineers in high-skill ecosystems often learn through "seeking and being given cutting-edge technical challenges" rather than formal courses (Finegold, 1999). However, "intensity of effort is critical" to develop effective capacity; as Cohen and Levinthal (1990) note, "the more deeply the material is processed... the better will be the later retrieval".

CIB underpins the WRL by evaluating the "investment cost" associated with each level. For example, moving from WRL 4 to WRL 5 requires shifting from a "laboratory environment" to a "production relevant environment," placing a heavy burden on innovators to learn "cross-cutting skills".

Skills Criticality & Dependency (SCD)

This dimension evaluates how essential specific skills are to the technology's deployment and the degree to which an organisation depends on external knowledge sources. It is grounded in absorptive capacity theory (Cohen & Levinthal, 1990).

Cohen and Levinthal (1990) define absorptive capacity as a firm's ability to "recognize the value of new, external information, assimilate it, and apply it to commercial ends". This capacity is "largely a function of the firm's level of prior related knowledge". They argue that learning is cumulative and "greatest when the object of learning is related to what is already known". Crucially, they warn of "lockout"—a condition where a lack of early investment in a technical area "may foreclose the future development of a technical capability in that area".

SCD underpins the WRL by assessing whether a workforce has the "prior related knowledge" to absorb a TRL-advanced technology. For instance, at WRL 7, the transition from "innovators" to "implementers" depends on whether the technicians have the "cross-function absorptive capacities" to understand designs handed off from R&D.

Education & Training Agility (ETA)

The Education & Training Agility (ETA) dimension measures the responsiveness of the education infrastructure to shifting technological demands. It draws from learning systems and vocational education research. Readying a workforce requires an agile interaction between "producer preference" (what educators teach) and "industry need". Buchanan et al., (2017) describe the challenges of "coordination failures" in vocational education and training (VET) systems, noting that enduring change requires "social coalitions" between employers, unions, and government.

The literature highlights that the most effective learning often occurs through "informal means such as working with others in their networks". However, the MIT Deskbook notes that long-term readiness requires "community colleges and vocational schools" to develop training programs in parallel with technology maturation so that "training will be ready and available when production needs to scale".

ETA evaluates the "agility" of this pipeline. WRL Level 8, for example, is reached when hiring and initial training guidance" is developed, signifying that the education system has successfully "vectored" its output with the technology's deployment.

Institutional & Demand Context (IDC)

The final dimension, Institutional & Demand Context (IDC), assesses the broader ecosystem's health. It is grounded in skills ecosystems theory. Finegold (1999) defines a High-Skill Ecosystem (HSE) as a "geographic cluster of organizations... employing staff with advanced, specialized skills". These ecosystems require four elements: a catalyst, ongoing nourishment, a supportive host environment, and interdependence.

Finegold argues that the "classification of sectors... as either high- or low-skill may itself be misleading" because regions can be trapped in a "low-skill equilibrium"—a self-reinforcing network of institutions that keep skills and wages low. Breaking this requires "institutional coordination" to ensure firms are "ambitious and demanding users of skills".

IDC underpins the WRL by evaluating whether the "market need" and "supply chain" are aligned with workforce development. A mature WRL 10 workforce is only possible in an ecosystem where "education at the post-secondary level... provides relevant exposure" and roles are "standardised".

Compositing Approach

The WRL framework aggregates these six dimensions into a single readiness score using a composite indicator methodology. Following the OECD Handbook (Nardo, Michela et al., 2005) we adopt a geometric mean for aggregation across dimensions rather than a simple arithmetic mean. In the context of workforce readiness, a high score in "Education Agility" cannot compensate for low scores in "Skills Criticality" or "Technology Disruption"; a serious failing in any one dimension is reflected in the overall readiness level, providing a more accurate diagnostic for policy and investment decisions.

WRAP Architecture

Design principles

The WRAP instrument has been designed against four constraints, which are in some tension with one another.

• It had to have a theoretical grounding and be defensible from an academic perspective. A composite instrument whose components are chosen by intuition alone invites the reasonable objection that a different author would choose differently. Each dimension of the WRAP instrument therefore corresponds to an identifiable research tradition that supplies both a rationale for its inclusion and a body of evidence about how the factor behaves.

• We wanted the tool to be deployable within realistic timescales and budgets. An instrument requiring eighteen months of primary data collection might be statistically more robust, but it would also have limited commercial application. WRAP is designed around a combination of secondary labour market and technology foresight data with a programme of semi-structured interviews across the relevant ecosystem, deliverable in a matter of weeks rather than years.

• We want the tool to accommodate qualitative evidence without pretending that it is quantitative. Several of the most important factors in workforce readiness, like the adaptability and agility s of education systems, and the quality of coordination between institutions or regions, do not reduce to available statistics. WRAP scores these using anchored qualitative judgement rather than excluding them, on the view that an informed judgement about an important factor is preferable to precise measurement of an unimportant one.

• We wanted to produce an output that non-specialists could act on.

The six dimensions

WRAP assesses workforce readiness across the six dimensions introduced above. In summary form, these are:

• Technology Maturity and Disruption, concerning the position and trajectory of the technology itself and the degree to which its adoption displaces or reconfigures existing practice. This dimension draws on the multi-level perspective in sociotechnical transitions research, which treats skills systems as part of the incumbent regime rather than as a neutral resource pool responding to technological change.

• Workforce Supply and Demand, concerning the relationship between forecast requirement and available or prospective supply, including the demographic health of the pipeline. Its theoretical grounding lies in the task-based account of technological change, which explains why technological adoption reshapes the composition of demand and not merely its volume.

• Capability Investment Burden, concerning what it costs, in time and money, to move a person from their current position to genuine competence in the new technology. This dimension is grounded in human capital theory, and in particular the distinction between general and firm-specific skills and the investment problems that distinction creates.

• Skills Criticality and Dependency, concerning how far successful deployment actually depends on scarce human expertise, how concentrated that expertise is, and how readily capability can be redirected from adjacent domains. Its anchor is the absorptive capacity literature, which holds that the capacity to assimilate new knowledge is a function of prior related knowledge and is therefore cumulative and path-dependent.

• Education and Training Agility, concerning the responsiveness of the provision system: how quickly curricula, qualifications and training capacity can be reconfigured, and how well providers and employers are connected.

• Institutional and Demand Context, concerning the wider institutional setting within which skills are formed and the quality of employer demand for them. This dimension is grounded in the skills ecosystems tradition, and in the older observation that skills systems can settle into self-reinforcing equilibria in which weak demand suppresses supply and weak supply in turn discourages employers from pursuing higher-skill strategies.

Each dimension resolves into a small number of assessed sub-factors (two or three). The specification of those sub-factors, the descriptive anchors used to score them, and the procedure for reconciling assessor judgements constitute the proprietary methodology of WRAP and are not set out here.

The sixth dimension deserves brief comment because it was not part of the original design. An earlier version of the framework was found to be supply-oriented. It assumed, in effect, that technology deployment generates demand for skilled labour and that the analytical question is whether supply can meet it. However, we viewed this assumption as fundamentally unsafe. Employers may pursue deployment strategies that minimise skill requirements; wages and conditions in emerging roles may be insufficient to attract entrants regardless of training capacity; and the institutional machinery that coordinates skills formation varies enormously between sectors in ways that materially affect outcomes. This led to our addition of a sixth dimension, and adjustment of the existing dimensions.

Scoring and aggregation

Each sub-factor is assessed on a bounded ordinal scale using descriptive anchors, with evidence drawn from both interview testimony and secondary data. Sub-factor scores combine to a dimension score, and the six-dimension scores combine to a composite from which the WRL band is derived.

The aggregation method used in the framework's first iteration was an equally-weighted arithmetic mean. This treated the dimensions as “compensatory”: A high score on one dimension offsets a low score on another, without limit. For workforce readiness this assumption is difficult to defend. An excellent curriculum does not compensate for the absence of anyone to teach it, or a course with no entrants. The dimensions are better understood as partially necessary rather than mutually substitutable.

WRAP therefore aggregates using a geometric mean, which penalises imbalance across dimensions and reduces, without eliminating, the compensability problem. This follows established practice in the composite indicator literature and the well-documented precedent of the UNDP's move from arithmetic to geometric aggregation in the Human Development Index. The change is not noticeable in many cases – especially where there is already some balance. However, in the cases where one dimension is missing or scoring at very low levels, the geometric mean will differ from the simple mean more noticeably.

We supplement the effect of the geometric mean by including a “critical minimum” flag on any single dimension that falls below a defined limit.

A further caution applies to the arithmetic itself. The sub-factor scores are ordinal, generated from descriptive anchors, and treating them as cardinal quantities to be multiplied and rooted involves an assumption that the intervals between anchor points are meaningfully equal. They almost certainly are not. This is a common weakness in composite indicator construction rather than a distinctive flaw of this instrument (Nardo, Michela et al., 2005), but it is a reason to treat the composite as an ordering device rather than a measurement, and to attend more closely to the profile than to the number.

The Workforce Readiness Level Scale

Structure of the scale

The WRL describes the maturity of the workforce system supporting a given technology in a defined context. It runs from 1 to 9, deliberately mirroring the TRL, and describes states rather than scores.

• WRL 1. Workforce implications may not be recognised. No structured assessment of workforce implications have been undertaken.

• WRL 2. Roles and requirements articulated. Emerging roles, indicative skill profiles and approximate volumes have been described, typically through foresight work. No pipeline for supply of capability / talent yet exists.

• WRL 3. “Pioneer” capability can be identified. A small number of practitioners hold relevant capability, usually acquired through research, adjacent practice or externally (to the entity being assessed). There are no structured routes to acquiring capability.

• WRL 4. Pilot provision established. First formal training interventions, curricula or conversion routes are operating at small scale, typically supported by public funding or a small number of committed employers. The gap between supply and requirement remains significant.

• WRL 5. Recognised pathways operating. Qualifications, apprenticeship standards or established conversion routes exist and are producing graduates. Supply is may be below requirement, and institutional coordination is emerging but incomplete.

• WRL 6. Provision at scale. Training capacity has extended across multiple providers. Standards and, where relevant, licensing frameworks are maturing. The gap is narrowing under existing arrangements.

• WRL 7. Requirement is broadly met. Supply approaches requirement across most roles. Residual bottlenecks persist, generally in the most specialised or experience-dependent positions, but the system is functioning without exceptional intervention.

• WRL 8. A mature workforce ecosystem. Supply of both qualified and experienced practitioners is self-sustaining. Career pathways, professional recognition and labour mobility are established.

• WRL 9. Embedded profession. The workforce reproduces through ordinary education, recruitment and career progression. The technology no longer presents a distinct workforce readiness question.

The mapping from a WRAP profile to a WRL band is part of the methodology. At present, this ladder is rather steep, and further work and critique is required to align scores and levels. In general terms, bands are determined by the overall profile rather than by the composite alone, since two sectors with identical composites can occupy different bands where their weaknesses fall on different dimensions. However, rather like TRL, we acknowledge it is likely that the scale will be used intuitively or without robust assessment at times. The descriptors therefore seek to provide some guidance on a standalone basis.

The readiness gap

The scale's principal use is comparative. Setting the WRL against the TRL of the technology produces a readiness gap that is straightforward to communicate and difficult to argue past.

A technology at TRL 8, entering commercial deployment, supported by a workforce at WRL 4 exhibits a four-level gap. That formulation locates the problem without requiring the audience to interpret a composite score, and it makes explicit something that headcount forecasting tends to obscure: the issue is not simply that too few people exist, but that the system responsible for producing them is several stages behind the technology it is meant to serve.

Gaps are not inherently ‘bad’. Some lag is normal and arguably efficient, since building workforce capacity ahead of confirmed technological demand carries its own costs and risks. A modest gap in a technology whose deployment timeline is long may require no intervention at all. What the scale supports is a judgement about whether a particular gap, given a particular deployment timeline, is tolerable. That judgement remains a matter for the parties involved. The instrument informs it rather than making it.

A caution inherited from the TRL

Pairing the WRL with the TRL borrows the latter's credibility, and it also borrows its problems. Reflecting on more than a decade of TRL practice, Mankins catalogued a number of issues in how the scale came to be used: readiness assessed for a component in isolation from the system it must work within, levels asserted rather than evidenced, and a general tendency for a communication device to harden into a procurement gate for which it was never designed (Mankins, 2009).

As with all scales and measures, caution must be used during application and in determining action based on a metric alone. Is a WRL claimed without assessment better or worse than no WRL at all? A WRL assessed for one occupational group and generalised to a whole sector might mislead. A scale intended to structure conversation between industry amd educators might be repurposed as a funding hurdle, which could create incentives to inflate. These risks are not hypothetical, and mitigating them is partly a matter of methodology and partly a matter of how the tool is governed in use.

Application and Use Cases

Levels of application

WRAP is designed to operate at multiple levels of analysis, so we describe it as a ‘fractal’ tool: You can ‘zoom in’ and apply the tool and it will have the same shape and form, or ‘zoom out’ to a higher level with the same shape and form. But low level applications may not be visible (impactful) at the higher level.

At a sector-level application, WRAP characterises the workforce system supporting a technology across a national or sub-national industry. This is the most common application and the one for which the instrument was originally designed. It suits sector skills bodies, industry associations, government departments and public agencies with a convening or funding role.

At a regional level, it might characterise the same question within a defined geography. Skills systems are strongly regional, and a sector that is broadly well provided for nationally may have constraints in other places, where deployment is actually planned.

At an organisational level, WRAP can be used to assess a single employer's readiness, drawing on the same structure but with narrower evidence requirements. Several dimensions, notably those concerning education systems and institutional context, are not obviously assessable at this level, but they remain relevant since an employer faces constraints it cannot resolve alone: In this case the dimensions function as context rather than as levers.

Evidence base

We intend that a full WRAP study combines secondary and primary evidence. Secondary sources include labour market statistics, vacancy and wage data, published foresight and technology roadmapping work, sector strategies and existing skills studies. Primary evidence comes from semi-structured interviews across the relevant ecosystem, deliberately spanning employers, education and training providers, professional bodies, and policy or funding actors, with the specific composition varying by sector.

Longitudinal use

A single assessment describes a state. Repeated assessment describes a trajectory, and the trajectory is arguably the more useful object. Movement between WRL bands provides a concrete milestone against which intervention can be evaluated, in a way that changes in a composite score do not.

Longitudinal application also partially addresses the framework's reliability problem, since a consistent assessment team applying a stable methodology to the same sector over time will produce more comparable results than the same methodology applied by different teams across different sectors. This is a real benefit but a modest one, and it does not substitute for establishing inter-rater reliability directly.

Discussion and Limitations

The framework described here is at an early stage. It has face validity, a defensible theoretical structure and has been ‘bench tested’ on several scenarios. It does not yet have the evidential foundations that would justify strong claims, and this section sets out what is missing.

WRAP is a diagnostic tool and a structured basis for dialogue. It is not a forecast, and it should not be treated as one. The framework can’t predict whether a skills shortage will materialise, when, or with what consequences. It characterises the current condition of a workforce system against dimensions that theory and evidence suggest are important to determine whether future shortfalls in capability will occur.

The determinants of workforce outcomes include labour mobility, migration policy, wage dynamics, macroeconomic conditions and the behaviour of competitor economies, none of which WRAP models and several of which can overwhelm domestic skills system performance entirely. In our view, a framework that claimed to predict outcomes while ignoring these factors would be overreaching.

Reliability

The most serious current gap is the absence of established inter-rater reliability. Scoring depends on structured judgement applied to interview evidence, and structured judgement varies between assessors. Until the framework has been applied independently by multiple assessors to the same case, or potentially to multiple cases and compared, we cannot be confident in the reliability of the rubric tools, and we expect to go through multiple further iterations before finalising these. Establishing reliability is the highest priority for the framework's development, and it requires a volume of application that no single engagement can generate. To that end, we are interested in collaborative deployment at scale.

Interview sampling and biases

Interview-based assessment is only as good as its sample. The ecosystem for any emerging technology contains actors with divergent and sometimes opposed interests, and those most readily accessible to a consultancy engagement are systematically not a random sample. This matters for specific dimensions: Assessments of role attractiveness and of the practical realities of retraining are precisely where employer testimony is least reliable and worker testimony most valuable. Where the sample may be biased towards employers, the framework should be expected to overstate readiness on these points. Ideally, cases for development work should be designed against bias.

Overlap between Dimensions

We acknowledge that several dimensions of WRAP are conceptually similar, and they may not be fully independent. The capacity to redirect skills from adjacent sectors affects more than one factor. If a technology displaces existing skills, then it also impacts the time required to build new skills, which means two dimensions might partially capture one underlying effect.

At this stage it is unclear how problematic “double-counting” might be, but the tool’s design is intended to minimise the effect of complementary measures. Until test or case studies are numerate enough for some statistical / correlation analysis, we cannot rule out this overlap. Future work may be able to better resolve overlap, or potentially reduce the number of dimensions or sub-dimensions / factors.

Weightings

The framework currently aggregates dimensions with equal weight, which is a starting point rather than the final design. We suspect the six dimensions do not contribute equally to workforce readiness, and it is likely that relative importance varies by sector.

Initial bench-test applications suggest that the framework's conclusions are reasonably robust in terms of balanced weightings, in that the rank ordering of the three applied cases held across the weighting schemes tested. This is reassuring but should not be overstated, since it was tested on three cases with a small number of schemes. Systematic sensitivity analysis remains to be done across a larger sample of ‘real world’ applications.

Omitted Factors

The framework cannot practically consider all potential impacts on workforce readiness, so it is necessary to focus on a ‘vital few’ that have the most effect. Some potential factors are consciously omitted (e.g. not practical to assess), others may have fallen outside our consideration.

“Worker agency” is partially represented, i.e. whether emerging roles are attractive enough to draw entrants, but the framework is based on an assessment of systems rather than input from the the people within them. There is nothing in the framework for example that addresses how workers experience transition (the UK’s “Just Transition” from work in offshore oil and gas, to work in offshore wind is a good example).

Labour mobility and labour migration are also largely outside scope. For several technologies, international recruitment is a material part of how a workforce gap is being addressed (expert immigration to Silicon Valley in the US, or overseas medics into the UK’s NHS, for example). The level at which mobility is considered, is always one or more levels above the WRAP assessment – for example, national or sector application may need to consider international migration; organisational assessment may consider ‘poaching’ and recruitment from rivals.

Diversity of the workforce is not assessed. There are some grounds to think that narrowly recruited workforces are more fragile and less adaptable, which would make this a readiness question rather than solely an equity one, but the framework does not currently address this.

Validation

The framework has not been tested retrospectively in several ‘bench test’ applications combining author’s knowledge with the use of AI to simulate responses based on web-search results and model knowledge. Applying WRAP to a historical case where the workforce outcome is now known, and asking whether the assessment would have produced outputs corresponding to what did happen might provide the closest available approximation to validation. Several candidate cases exist within the UK energy and manufacturing sectors over the past two decades. This work remains to be done.

Conclusion and Future Development

The argument of this paper is narrow. Technology readiness is assessed with structure, shared vocabulary and reasonable rigour. Workforce readiness generally is not assessed in the same seriousness, despite being a comparable determinant of whether broad-front technological deployment succeeds. Workforce Readiness Level is proposed as a means of closing that gap, and the Workforce Readiness Assessment Protocol (WRAP) as the instrument that generates it.

Most of what WRAP assesses has been examined somewhere in the existing literature. What has been missing is an instrument that brings these considerations together in a form that can be applied consistently across sectors, that produces an output legible to decision-makers, and that is anchored in theory rather than assembled from readily available, or easily-synthesised data.

Three features of the WRAP tool have proved more consequential in application than anticipated. The distinction between formal qualification and operational competence, and the time required to bridge them, turns out to be a significant constraint in technologies where few operational sites exist to learn on, and one that training-volume metrics miss entirely. The institutional dimension, added late in development, has discriminated between sectors that other dimensions ranked similarly. The framing of readiness as a gap against technology maturity, rather than as an absolute state, may communicate more effectively with technical and “non-skills” audiences than composite scores.

Against this, the framework is currently lacking in testing and evidencing. We need to further test and then refine for reliability, interactions between dimensions, and whether the default equal-weighting approach is reasonable or needs to be adjusted (potentially, on a per-sector basis). None of these gaps are unusual for a diagnostic instrument at this stage of development, but they help to set the agenda for the next stages of development.

That agenda is not achievable through commercial engagement alone. Establishing reliability requires independent parallel assessment. Testing overlap between dimensions (independence) needs a larger sample of tests, retrospective validation requires access to historical cases and the labour market data. Extending the framework beyond its current calibration requires application in sectors and countries where its assumptions may not hold. We are therefore interested in collaboration to further develop and test this framework, specifically in three kinds, and enquiries are welcome:

1. We believe trade associations and public agencies may be able to provide access to a substantial pool of organisations for assessment. Deployment at scale within a single sector would generate the volume of cases needed to test reliability and construct independence, while producing a benchmarked picture of that sector which would be of considerable value to the body itself.

2. Academic partners with interests in skills policy, labour economics or transitions research. The methodological questions raised in Section 7 are research questions, and they would be better addressed by researchers than by practitioners working alone.

3. Organisations in sectors unlike those considerd so far, particularly regulated professional contexts and settings outside the United Kingdom, where the framework's implicit assumptions can be tested rather.

In closing, the value of the Technology Readiness Level is not (and never was) that it measures technology maturity precisely. Its levels are qualitative descriptions, and level assignment often relies on judgement and intuition rather than analysis. But, what TRL did was give people who need to make decisions a common way of describing where a technology stands. This has turned out to be more useful than any of the more precise metrics available at the time or since. In our view, if we can drive the concept of Workforce Readiness Level achieves to achieve some similar for the workforce question, and if WRAP can add some additional rigour, then this will be a significant step forward.

---

References

Aiken, J., Moore, E., Fedewa, A., Jordan, T., Field, F., Unger, E., & Kirchain, R. (2024). Workforce Readiness Level (WRL) Deskbook (V1.01). Conducere-MIT Collaboratory (CMC). https://dspace.mit.edu/entities/publication/d27087c3-9d61-4370-ae9a-721f63257b16

Autor, D. H., Levy, F., & Murnane, R. (2002, September). The Skill Content of Recent Technological Change: An Empirical Exploration. Department of Economics Working Paper Series.

Bailey, D., De Propris, L., Dimos, C., Fai, F. M., Hardy, S., & Tomlinson, P. R. (2026). A critical review of the UK’s Modern Industrial Strategy: Lessons for ‘place-based’ policy. Regional Studies, 60(1), 2597466. https://doi.org/10.1080/00343404.2025.2597466

Bakhshi, H., Downing, Jonathan M., Osborne, M. A., & Schneider, P. (2017). The Future of Skills Employment in 2030. Pearson and Nesta.

Buchanan, J., Anderson, P., & Power, G. (2017). Chapter 21: Skill Ecosystems. In The Oxford handbook of skills and training. Oxford University Press.

Cohen, W. M., & Levinthal, D. A. (1990). Absorptive Capacity: A New Perspective on Learning and Innovation. Administrative Science Quarterly, 35(1), 128–152.

Finegold, D. (1999). Creating self-sustaining, high-skill ecosystems. Oxford Review of Economic Policy, 15(1), 60–81. https://doi.org/10.1093/oxrep/15.1.60

Geels, F. W. (2002). Technological transitions as evolutionary reconfiguration processes: A multi-level perspective and a case-study. Research Policy, 31(8–9), 1257–1274. https://doi.org/10.1016/S0048-7333(02)00062-8

Geels, F. W. (2011). The multi-level perspective on sustainability transitions: Responses to seven criticisms. Environmental Innovation and Societal Transitions, 1(1), 24–40. https://doi.org/10.1016/j.eist.2011.02.002

Hogan, G. C. H., Megan. (2025, January 15). Industrial policy through the CHIPS and Science Act: A preliminary report | PIIE. https://www.piie.com/publications/piie-briefings/2025/industrial-policy-through-chips-and-science-act-preliminary-report

Mankins, J. C. (1995). Technology Readiness Levels [White Paper]. NASA. http://www.artemisinnovation.com/images/TRL_White_Paper_2004-Edited.pdf

Mankins, J. C. (2009). Technology readiness assessments: A retrospective. Acta Astronautica, 65(9–10), 1216–1223. https://doi.org/10.1016/j.actaastro.2009.03.058

Nardo, Michela, Saisana, Michaela, Saltelli, Andrea, Tarantola, S., Hoffmann, A., & Giovannini, E. (2005). Handbook on Constructing Composite Indicators: Methodology and User Guide (OEC Statistics Working Papers No. 2005/03). OECD Publishing. https://doi.org/10.1787/9789264043466-en

RenewableUK. (2024). 2024 Offshore Wind Industrial Growth Plan: Expanding the Horizon of the UK’s Offshore Wind Supply Chain. Renewable UK, OWIC, Crown Estate Scotland, The Crown Estate.

Smith, D. R., Whittington, W., Martinez, A., Duncan, A., & Li, G. (2026). A Conceptual Framework for Enhancing Workforce Readiness for Smart Manufacturing in the AI Era (arXiv:2608.11540). arXiv. https://doi.org/10.48550/arXiv.2608.11540