A Conceptual Framework for Enhancing Workforce Readiness for Smart Manufacturing in the AI Era

Synthesis: Smith and colleagues (2026) argue that AI, the Industrial Internet of Things, cyber-physical systems, and advanced robotics are reshaping manufacturing faster than engineering curricula can adapt, widening the gap between shop-floor competencies and what universities teach. To close it, they propose a Workforce Readiness Level (WRL) framework that adapts the Technology Readiness Level scale into nine progressive competency stages scored across four pillars — digital and AI literacy, cyber-physical systems fluency, human-machine collaboration, and data-driven decision making — aggregated under a "no-thin-pillar" rule. Instantiated in a university smart-manufacturing teaching laboratory over 89 sponsored capstone projects, the framework surfaced hidden cyber-physical and data-driven gaps and showed that advancement to the highest stages was gated by industry-embedded experience rather than additional coursework, offering educators and accreditation bodies a common instrument for diagnosing workforce readiness.

Key Findings

1. The WRL framework adapts the Technology Readiness Level (TRL) scale into nine behaviorally anchored competency stages grouped into three bands — awareness (WRL 1–3), applied practice (WRL 4–6), and autonomous leadership (WRL 7–9) — with stackable-credential articulation points at WRL 3, 5, and 7.
2. Readiness is scored across four competency pillars (digital & AI literacy, cyber-physical systems fluency, human-machine collaboration, and data-driven decision making) under a "no-thin-pillar" rule that blocks certification at a stage if any single pillar falls below the rubric floor, even when the composite score looks strong.
3. Across four in-depth case studies drawn from 89 sponsored capstone projects, cohort workforce-readiness indices ranged from 5.2 to 6.4, and the no-thin-pillar rule repeatedly surfaced cyber-physical (P2) and data-driven-decision (P4) gaps concealed behind otherwise strong analytics profiles — binding certification in one case.
4. Advancement to the highest readiness stages (the WRL 6→7 transition) was gated by industry-embedded experience (co-ops and Manufacturing Extension Partnership projects), not by additional coursework.
5. The four pillars map directly onto the ABET student outcomes, so a WRL transcript doubles as accreditation evidence and the cohort-level index (WRI) supports ABET continuous-improvement reporting.

The framework: from technology readiness to workforce readiness

The paper's central move is to transplant a familiar assessment logic — the Technology Readiness Level (TRL) scale used across engineering — into the realm of human competency. The resulting Workforce Readiness Level framework defines nine progressive stages of workplace capability, each aggregating four pillars:

1. Digital and AI literacy — the foundational ability to work with data, automation, and AI tools, including computational thinking, the Machine Learning lifecycle, model interpretability, and AI ethics.
2. Cyber-physical systems fluency — understanding how physical processes connect to networked computation, spanning industrial controls, IIoT sensing, and the OT/IT bridge role.
3. Human-machine collaboration — working effectively alongside automation and AI agents, including cobot safety, ergonomics, and human-in-the-loop inspection.
4. Data-driven decision making — using evidence to make operational choices, via KPIs, statistical process control, design of experiments, and Lean/Six Sigma.

Three design commitments ground the construct: competency must be demonstrated through observable performance on artifacts rather than asserted through coursework or self-report (an artifact-anchored, performance-based stance); progression must be stage-gated and monotonic (once a stage is reached it is not lost), so employers and accreditors can rely on the scale without re-arguing evidence; and the construct must be multidimensional, because AI-era manufacturing roles inherently span computing, controls, human factors, and analytics. The nine-stage scale is a deliberate choice inherited from TRL/MRL to stay compatible with the technology-maturity language industry already uses, and the awareness/lab/production banding mirrors the Dreyfus novice-through-expert trajectory.

The four competency pillars

The four pillars span the information-, machine-, human-, and decision-sides of AI-era manufacturing work, and were derived from the Industry 4.0 competency literature (e.g., Hernandez-de-Menendez, Tortorella, Maisiri) foregrounded with the AI-specific competencies highlighted by the World Economic Forum's demand projections. Each pillar lists representative competencies:

• P1 Digital & AI Literacy covers computational thinking, data structures, applied statistics, the machine-learning lifecycle (data → features → training → validation → deployment → monitoring), foundation-model and prompt literacy, model interpretability (SHAP, LIME), and AI ethics including bias, privacy, and human oversight.
• P2 Cyber-Physical Systems Fluency covers industrial controls (IEC 61131-3 PLCs), SCADA and HMI design, IIoT sensing and commissioning, industrial network protocols (OPC-UA, MQTT, PROFINET), OT/IT security, and digital-twin construction — mapping to the OT/IT bridge role identified as the most acute Industry 4.0 skill shortage.
• P3 Human-Machine Collaboration covers collaborative-robot safety (ISO 10218 / ISO/TS 15066), augmented- and virtual-reality-assisted work instruction, human-in-the-loop AI for inspection and decision support, ergonomics, and teaming in mixed human-cobot cells. Including it as a peer of P1/P2 reflects evidence that human-factors competence predicts successful cobot deployment.
• P4 Data-Driven Decision Making covers KPI definition and instrumentation (OEE, first-pass yield, cycle time), statistical process control, design of experiments, root-cause analytics (5-Why, fishbone), and Lean/Six Sigma DMAIC — the pillar closest to the no-thin-pillar floor and most likely to trigger remediation.

These pillars are intended to be conceptually distinct but related in practice, so the rubric scores them separately and the no-thin-pillar rule can detect lopsided profiles — the strong coder who cannot read an SPC chart, or the seasoned operator who cannot interpret a model output. This encodes the "T-shaped versus I-shaped" idea that AI-era competency has several dimensions that cannot be collapsed into one.

Evidence from a teaching laboratory

The framework is instantiated in a university smart-manufacturing teaching lab — the Innovation, Design, and Engineering Education Laboratory (IDEELab) at Mississippi State University — drawing on 89 sponsored capstone projects delivered over four semesters, with four analyzed in depth. The lab's four reconfigurable cells (Robotics & Assembly, Process & Control, Additive & Subtractive, and Digital-Twin & Analytics) publish to a shared OPC-UA/MQTT data backbone, so a single project can exercise the machine, human, and data facets of a task in one continuous workflow. Findings include:

• Cohort workforce-readiness indices ranged from 5.2 to 6.4 across the highlighted cohorts.
• The no-thin-pillar rule was diagnostically informative in three of four cases and the binding certification constraint in one, repeatedly exposing cyber-physical and data-driven-decision gaps that were concealed behind otherwise strong analytics profiles.
• The four cases spanned four sectors — aluminum smelting, heavy-truck manufacturing, commercial refrigeration, and defense/naval research — without change to the pillar rubric, and faculty-mentored research capstones and industry-mentored capstones yielded broadly comparable pillar profiles.
• Crucially, advancement to the highest readiness stages was gated by industry-embedded experience, not additional coursework — the WRL 6→7 transition was reached only after co-op placements or on-site production installs, a finding with direct implications for how workforce preparation is designed.

Assessment and accreditation

The WRL evaluation model is deliberately conservative: a single weak pillar can block certification even if the other three are strong, protecting downstream stakeholders from credential inflation. Each pillar is scored 0–3 against a behaviorally anchored rubric (144 anchors across 9 stages × 4 pillars), and a learner is certified at stage k only if the composite score clears a threshold and every pillar meets the floor. The cohort-level workforce-readiness index (WRI) summarizes the typical level across a cohort and is treated as a comparative Benchmark rather than a competency location on the nine-point scale.

Because the four pillars map onto the ABET student outcomes (SO1–SO7), a WRL transcript doubles as a source of accreditation evidence: P1 and P4 provide strong evidence for complex problem solving, experimentation/data analysis, and applying new knowledge; P2 anchors engineering design; and P3 carries ethical/professional responsibility and teamwork via functional-safety practice. Every ABET outcome receives at least supporting evidence from the pillar set, making WRI a defensible quantitative artifact for accreditation and continuous-improvement files.

What this means for practice

• Instructors. Score competency on demonstrated artifacts rather than coursework or self-report. WRL certifies a stage only when every pillar clears the rubric floor (no pillar below 2) and the composite reaches τ = 2.25, so a strong average — or a stack of completed courses — cannot carry a weak pillar.
• Instructors. Read the four-pillar profile before certifying. The no-thin-pillar rule exposed cyber-physical (P2) and data-driven-decision (P4) gaps hidden behind strong analytics profiles, was diagnostically informative in three of four cases, and became the binding certification constraint in one.
• Administrators. Do not expect coursework to reach the top stages. Advancement from WRL 6 to 7 was gated by industry-embedded experience — co-ops and Manufacturing Extension Partnership projects — so build work-integrated learning into the degree rather than adding another technical elective.
• Administrators. Use the four pillars as accreditation evidence. They map onto the ABET student outcomes SO1–SO7 and the cohort-level WRI summarizes typical standing, so one assessment instrument can serve both certification and continuous-improvement reporting.
• Designers. Weight credentials by the performance they demand. A hands-on skills test should contribute more to a stage than a knowledge-only multiple-choice exam, because passing a PLC exam does not show that a technician can wire, commission, and troubleshoot a live cell.

Limitations

• Single-institution pilot: the framework was exercised at one teaching lab (IDEELab, Mississippi State University) across 89 sponsored capstone projects, of which only four were analyzed in depth.
• Rubric scores were assigned retrospectively by the CDI instructional team, and the two-rater Cohen's κ protocol (target κ ≥ 0.75) was not operationally in force during the case-study period — no κ value is reported.
• The four-pillar partition and the 0–3 anchor scale are design assumptions chosen for rater workflow and single-page reporting, not a validated factor structure; construct-validity and reliability studies are explicitly future work.
• Only the WRL 4–7 band was exercised: the awareness stages (WRL 1–3) and the supervisory and innovation stages (WRL 8–9) were untested by this undergraduate capstone pilot, and replication beyond one site is pending.

Citation

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.