Sheet Metal Workers
47-2211.00Median wage $61,800/yr119,770 employed (US)Rank #684 of 923 scored · top 74% by substitution
Fabricate, assemble, install, and repair sheet metal products and equipment, such as ducts, control boxes, drainpipes, and furnace casings. Work may involve any of the following: setting up and operating fabricating machines to cut, bend, and straighten sheet metal; shaping metal over anvils, blocks, or forms using hammer; operating soldering and welding equipment to join sheet metal parts; or inspecting, assembling, and smoothing seams and joints of burred surfaces. Includes sheet metal duct installers who install prefabricated sheet metal ducts used for heating, air conditioning, or other purposes.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution20
Exposure14
Augmentation37
Substitution — the headline: capability discounted by cost, barriers and adoption.
Exposure — technical capability alone, regardless of whether anyone deploys it.
Augmentation — how much AI assists without replacing. High here + moderate substitution = a changing job, not a disappearing one.
Tasks on the substitution scale
19 rated tasks, binned by substitution score.
Position among all scored occupations
Distribution of 923 occupation scores; the marker is this occupation.
Tasks with substitution ≥ 70
0%
Run 1.0.0-draft.1 · computed 2026-08-05 · rater panel: claude-sonnet-5, claude-haiku-4-5-20251001 · intervals span rater disagreement.
Why this score
The five weighted dimensions of the composite, averaged across this occupation's tasks (importance-weighted, panel mean). Exact weights and formulas: /api/v1/methodology.
Task automatabilityw 35%15
panel mean rating 1.6/5 → substitution pressure 15/100
Technical feasibility todayw 20%11
panel mean rating 1.5/5 → substitution pressure 11/100
Cost vs. human wagew 15%13
panel mean rating 1.5/5 → substitution pressure 13/100
Adoption barriersw 20%inverted — strong barriers lower the score45
panel mean rating 3.2/5 (barrier strength) → substitution pressure 45/100
Sector adoption velocityw 10%12
panel mean rating 1.5/5 → substitution pressure 12/100
Task breakdown (19 tasks)
Substitution pressure per task, weighted by O*NET importance in the composite. Expand a task for the full rater audit trail — every rating, every model, every rationale.
Inspect individual parts, assemblies, or installations, using measuring instruments, such as calipers, scales, or micrometers.
51CI 35–67 · exposure 45 · augmentation 63 · importance 4.4/5 · click for rater detail
Inspect individual parts, assemblies, or installations, using measuring instruments, such as calipers, scales, or micrometers.
51| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Large automotive and aerospace sheet metal suppliers have adopted vision-based inspection systems, but small and mid-sized job shops lag significantly due to capital constraints and variability in part geometry. Adoption is growing but remains spotty across the broader sheet metal sector. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Metal fabrication is a physically-oriented, lower-digitization sector where automated inspection systems are adopted selectively in high-volume production lines but slowly in job-shop or custom sheet metal contexts. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-driven measurement and defect detection can dramatically accelerate inspection, flagging anomalies and measurements for human review rather than requiring manual measurement of every dimension. This substantially raises inspector productivity and enables focus on complex judgment calls rather than routine dimensional checks. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Digital calipers/micrometers with data logging, and camera-based measurement apps can assist workers by speeding up recording and flagging deviations, improving consistency without removing the human from the loop. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Machine vision systems can measure geometric dimensions autonomously using image analysis and comparison to CAD specifications, achieving both speed and consistency gains over manual measurement. While setup for each part type requires some configuration, the measurement operation itself is highly automatable and could easily exceed the 50% time-saving threshold at equal or superior quality. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical inspection of manufactured parts using hand tools requires manipulation and tactile measurement that current general-purpose AI cannot perform end-to-end; some vision-based automated inspection exists but it's not the same as manual caliper/micrometer use by a worker on the shop floor. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | There are few legal or licensing barriers to automating dimensional inspection; no human signature or authorization is strictly required by law, though quality documentation practices may require review. Organizational friction exists around trust in AI measurement and downstream liability, but these are softer adoption barriers rather than hard regulatory requirements. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement mandates human inspection specifically, but quality/safety liability and the physical nature of handling parts creates some friction against pure automation without human oversight. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Modern vision-based inspection systems have become significantly cheaper than dedicated human inspectors when accounting for labor burden, throughput, and consistency, though initial capital and integration costs must be amortized. For high-volume operations, automated systems cost a fraction of human inspection labor per unit. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Vision-based inspection hardware and integration costs are substantial relative to a worker simply using calipers, and setup/calibration for varied part geometries typically exceeds the marginal cost of human inspection in small-to-mid batch sheet metal work. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Automated optical and machine vision inspection systems are deployed in manufacturing environments, but they typically require careful lighting, fixturing, and calibration for each application. Real-world systems have material limitations on part complexity and occlusion, and integration into existing workflows remains imperfect across diverse sheet metal shops. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Automated optical/dimensional inspection systems exist in high-volume manufacturing, but for sheet metal work with varied parts and manual measuring tools, deployed AI-driven inspection replacing the worker's hands-on task is rare and narrow in scope. |
Convert blueprints into shop drawings to be followed in the construction or assembly of sheet metal products.
37CI 25–49 · exposure 38 · augmentation 63 · importance 3.7/5 · click for rater detail
Convert blueprints into shop drawings to be followed in the construction or assembly of sheet metal products.
37| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Sheet metal work remains a skilled trades sector with relatively low digital adoption compared to tech-forward industries. While CAD use is standard, autonomous conversion tools have seen limited production deployment in real shops; pilots are uncommon and adoption is lagging. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Construction and metal fabrication remain a moderately digitized, project-based sector with slower AI tool adoption compared to information/finance industries, though BIM adoption is growing steadily. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | Current CAD and AI drafting assistants can speed up routine aspects like geometry alignment and dimension extraction, helping human drafters work faster. However, the augmentation is limited to component parts of the task; human expertise in manufacturing feasibility and assembly logic remains essential. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-assisted CAD tools significantly speed up drawing generation, error-checking, and layout optimization, letting drafters focus on complex judgment calls while routine drawing tasks are automated. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can extract some dimensional data from blueprints and generate basic CAD geometry, converting blueprints to detailed shop drawings requires spatial reasoning, material-specific decisions, assembly sequencing, and judgment about constructability that current systems struggle with. The task involves significant domain knowledge and contextual interpretation beyond what off-the-shelf AI can reliably automate end-to-end. |
| Task automatability | claude-sonnet-5 | 3/5 | CAD/BIM software with AI-assisted drafting can convert blueprints into shop drawings for many standard sheet metal parts, but complex or non-standard geometries still require skilled human interpretation and coordination with fabrication constraints. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Legal and liability barriers are substantial: shop drawings are engineering documents that may require professional sign-off or licensure depending on jurisdiction and product criticality. Safety and quality standards in manufacturing mean that automation without qualified human review faces regulatory and organizational resistance. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement specifically for shop drawing creation, but liability for fabrication errors and coordination with site conditions creates moderate friction and reliance on experienced tradespeople for sign-off. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-assisted tools still require significant human oversight, verification, and correction by skilled sheet metal workers. The total cost of integration, human review time, and error correction keeps the cost comparable to or potentially higher than traditional human drafting, especially for complex projects. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Software licenses plus the need for skilled drafters/estimators to verify and adjust AI-generated drawings keep costs comparable to human drafting, though some efficiency gains exist for repetitive standard parts. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some CAD software includes automation features for basic geometry conversion, but no deployed product reliably converts complex blueprints to full, production-ready shop drawings without substantial human intervention. Current AI struggles with ambiguous blueprint notations, non-standard specifications, and the manufacturing expertise required for this conversion. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | CAD tools with parametric and AI-assisted drafting features (e.g., AutoCAD, SolidWorks, BIM sheet metal plug-ins) are deployed in production, but full automated blueprint-to-shop-drawing conversion for varied real-world jobs still needs significant human oversight and correction. |
Determine project requirements, such as scope, assembly sequences, or required methods or materials, using blueprints, drawings, or written or verbal instructions.
28CI 25–30 · exposure 25 · augmentation 50 · importance 3.9/5 · click for rater detail
Determine project requirements, such as scope, assembly sequences, or required methods or materials, using blueprints, drawings, or written or verbal instructions.
28| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Sheet metal work remains a physical, trade-based sector with lower digitization than IT or finance; adoption of AI for planning and specification review is still in pilot phases, with most firms continuing traditional human-led requirement analysis. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Construction and metal fabrication trades are historically slow adopters of AI, with most digitization limited to CAD/BIM software rather than autonomous planning agents. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by automatically extracting and organizing information from blueprints, flagging inconsistencies, and generating preliminary material lists, which would reduce manual review time. However, the human expert must still validate and make final decisions. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help by summarizing blueprints, flagging inconsistencies, or suggesting standard assembly sequences, meaningfully aiding human planners even though the human retains final judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can extract and summarize information from blueprints and technical documents through OCR and vision models, determining project requirements requires integrating multiple information sources, resolving ambiguities, and making judgment calls about feasibility—tasks that typically require human expertise and site context. Current AI cannot reliably do this end-to-end with 50% time savings. |
| Task automatability | claude-sonnet-5 | 2/5 | Interpreting blueprints and determining assembly sequences requires spatial reasoning and physical context that current AI can partially assist with (e.g., reading drawings) but cannot reliably complete end-to-end without human verification. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Project requirements determination typically requires sign-off by licensed professionals and project managers; errors in interpretation can lead to significant rework costs and liability. Organizational workflows embed human authority in requirement-setting, and there is strong preference for human accountability. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No formal licensing requirement for this specific planning task, but liability for incorrect assembly sequencing or material specification creates strong incentive for human verification. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The cost of AI document processing, blueprint analysis, and human oversight for verification approaches the cost of a trained sheet metal worker reviewing specifications themselves, especially accounting for integration and error rates in real-world conditions. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI tools that parse blueprints still require significant human oversight and integration with fabrication planning, so cost savings are modest rather than transformative. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Prototype systems can parse technical drawings and extract basic specifications, but no deployed product reliably determines complete project requirements, assembly sequences, and material selections from mixed input formats in production environments. The integration of visual, written, and verbal inputs with error correction remains limited. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some CAD/BIM-integrated tools can extract specs from drawings, but no deployed product autonomously determines full project scope and material/method requirements in production sheet metal work. |
Select gauges or types of sheet metal or nonmetallic material, according to product specifications.
28CI 23–33 · exposure 20 · augmentation 50 · importance 3.7/5 · click for rater detail
Select gauges or types of sheet metal or nonmetallic material, according to product specifications.
28| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Sheet metal working is a traditional manufacturing trade with relatively low digitization outside large aerospace/automotive suppliers. Most small to mid-sized sheet metal shops operate with manual processes and low digital integration, meaning adoption of AI-based selection tools remains slow and limited. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Sheet metal fabrication is a physical, lower-digitization trade with minimal AI agent deployment in material selection workflows. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could moderately assist by providing quick lookups, cross-referencing specifications against material properties, or flagging mismatches—helping workers verify selections faster. However, the task itself is straightforward enough that augmentation gains are incremental rather than transformative. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI tools (spec databases, material lookup assistants) can help workers quickly cross-reference specifications and standards, improving speed and reducing errors in selection. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Selecting the correct material type and gauge requires interpreting product specifications and matching them to available materials, which involves domain knowledge and nuanced judgment. While AI could assist in matching specifications to a database of materials, the task demands verification against real-world constraints (availability, cost, performance characteristics) that typically requires human oversight and cannot achieve 50% time savings end-to-end today. |
| Task automatability | claude-sonnet-5 | 2/5 | Material/gauge selection requires physical inspection, tactile judgment, and integration with fabrication constraints that current AI cannot fully replicate end-to-end, though spec lookup could be partially assisted.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Some organizational and technical friction exists: workers are accustomed to manual selection, material databases vary by supplier, and errors can be costly (wrong gauge leads to product failure). However, there is no legal licensing requirement or hard regulatory barrier preventing automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement specifically for this selection step, but quality/safety consequences of wrong material choice create moderate organizational caution against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Material selection is typically a quick manual task performed by experienced workers as part of setup, with low marginal human cost. The setup and integration overhead for an AI system to access specifications, material databases, and constraint checks would likely exceed the small time savings, making AI more expensive all-in. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Any AI assistance would still require human verification and physical handling, so cost savings are minimal relative to the skilled worker's wage for this narrow decision task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably performs full material selection from specifications autonomously at scale. While AI can help search databases or suggest candidates, real-world adoption in sheet metal shops remains minimal; most selection still relies on human expertise and manual lookup processes. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously selects sheet metal gauges/types on shop floors; this remains a human judgment call informed by specs and physical handling. |
Trim, file, grind, deburr, buff, or smooth surfaces, seams, or joints of assembled parts, using hand tools or portable power tools.
25CI 15–35 · exposure 13 · augmentation 25 · importance 3.6/5 · click for rater detail
Trim, file, grind, deburr, buff, or smooth surfaces, seams, or joints of assembled parts, using hand tools or portable power tools.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Sheet metal work remains concentrated in mid-sized job shops and small manufacturers with low digital maturity; even large fabrication firms adopt robotic finishing only on high-volume, standard products, not on the variable, custom work typical of the sector. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Sheet metal fabrication is a physical, lower-digitization trade with slow robotics/AI adoption compared to information-sector tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI vision could highlight defects or guide workers to problem areas, but current systems offer limited real-time feedback during active hand-tool use; augmentation potential is narrow because the task is already tactile and iterative, where human judgment and feel dominate. |
| Augmentation potential | claude-sonnet-5 | 2/5 | Power tools and some robotic-assist grinding arms can aid throughput, but general AI systems offer minimal direct assistance to this hands-on finishing task. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI vision systems can detect surface imperfections, the task requires precise hand-tool control, haptic feedback, and adaptive judgment about pressure and angle—capabilities current robotic systems struggle with in unstructured, variable parts. Material removal tasks demand dexterity and real-time adjustment beyond today's automation threshold. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical manual finishing task requiring dexterity, tactile feedback, and tool manipulation on varied metal parts; no off-the-shelf AI system can perform this physical labor end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | No legal licensing or human sign-off is required, but shop-floor safety, quality control oversight, and the difficulty of deploying reliable automation in small-batch or custom work create practical friction against substitution. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but quality/safety tolerances and physical variability of parts create practical friction against automation without significant capital investment. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Specialized finishing robots are capital-intensive and require setup; labor costs for sheet metal workers remain modest per unit, and custom fixturing for automation often exceeds the savings from a few hours of hand finishing per part. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Robotic finishing cells require expensive custom fixturing, programming, and maintenance, often exceeding the cost of a skilled worker for low-volume or variable sheet metal work. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Robotic deburring and finishing systems exist in limited, structured production lines, but they are narrow in scope, require extensive fixturing, and struggle with part variability. No general-purpose deployed solution reliably handles the full range of hand-finishing tasks across different materials and geometries. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | Deployed robotic deburring/grinding systems exist only in narrow, highly controlled industrial cells for repetitive parts, not as general-purpose AI products handling varied sheet metal assemblies. |
Lay out, measure, and mark dimensions and reference lines on material, such as roofing panels, using calculators, scribes, dividers, squares, or rulers.
24CI 19–30 · exposure 20 · augmentation 38 · importance 4.0/5 · click for rater detail
Lay out, measure, and mark dimensions and reference lines on material, such as roofing panels, using calculators, scribes, dividers, squares, or rulers.
24| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Sheet metal work remains largely in small-to-medium firms with lower digital infrastructure; while some large fabrication shops use CAD-to-production software, adoption of physical automation for layout and marking is still in pilot phase, not production-grade displacement. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Construction and metalworking trades show low AI/robotics adoption for physical layout tasks, remaining a manual craft with minimal digitization in this specific step. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-assisted measurement (dimension suggestions, layout optimization from images or CAD) can speed up human layout work and reduce calculation errors, but the human worker must still execute and verify marks on material, so assistance is partial rather than transformative. |
| Augmentation potential | claude-sonnet-5 | 2/5 | CAD/CAM software and digital measurement tools can assist in planning layouts, but AI's role in the physical marking process itself is minimal. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Current AI vision systems can identify and measure physical dimensions from images, but reliable end-to-end automation requires precise 3D spatial understanding, physical marking capability, and real-time adaptation to material variations—none of which are deployable at production speed today. The task involves physical manipulation and marking that AI cannot execute without robotic systems, and even with vision input, the quality control bar is high. |
| Task automatability | claude-sonnet-5 | 2/5 | This requires physical manipulation of material and precise hands-on marking, which current AI systems cannot perform end-to-end; only digital layout planning could be assisted, not the physical marking act itself.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | No legal licensing requirement prevents automation, but organizational barriers exist: workers often customize layouts on-site based on material irregularities and site conditions, quality expectations demand human sign-off, and integration of automated marking into existing workflows requires process redesign and inspection protocols. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing strictly requires a human, but the physical nature of marking materials on-site with tools creates strong practical barriers to automation without robotic hardware. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI vision and measurement software costs less than human labor per unit, but integrating such systems into a workshop workflow, combined with robotic marking hardware, quality assurance, and human oversight, makes the all-in cost comparable to or higher than a skilled sheet metal worker's loaded wage for equivalent output quality. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing the physical marking task, so cost comparison favors the human worker entirely. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Computer vision can measure and analyze 2D layouts in controlled settings, but no mature product reliably performs the full task (layout, measurement, marking, and reference line inscription) on varied physical materials in real shop environments. Partial automation of measurement input exists, but the marking and physical scribing steps remain manual and human-validated. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product physically lays out or marks material; this remains a manual trade skill requiring physical dexterity and tool handling. |
Verify that heating, ventilating, and air conditioning (HVAC) systems are designed, installed, and calibrated in accordance with green certification standards, such as those of Leadership in Energy and Environmental Design (LEED).
23CI 20–25 · exposure 20 · augmentation 50 · importance 3.6/5 · click for rater detail
Verify that heating, ventilating, and air conditioning (HVAC) systems are designed, installed, and calibrated in accordance with green certification standards, such as those of Leadership in Energy and Environmental Design (LEED).
23| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | HVAC construction and retrofitting remains a traditional, geographically distributed sector with fragmented adoption patterns. While some firms use digital tools for documentation, AI-driven compliance verification adoption in production is minimal and limited to early adopters. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Construction and skilled trades are historically slow adopters of AI for physical verification tasks, with adoption concentrated in office/planning functions rather than field inspection. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by organizing LEED requirements, cross-referencing design drawings with standards, and flagging documentation gaps, meaningfully reducing verification prep work. However, the final judgment call on compliance and calibration correctness requires human expertise. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help by summarizing LEED requirements, flagging documentation gaps, and generating checklists, but the actual verification against installed systems still depends on human inspection. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | This task requires domain expertise in HVAC systems, green building standards, and code interpretation, combined with site-specific judgment and stakeholder communication. While AI could assist in documentation review and checklist generation, the verification itself demands qualified human judgment about design compliance and on-site calibration correctness. |
| Task automatability | claude-sonnet-5 | 2/5 | This requires physical inspection, hands-on verification of installation and calibration, and judgment against certification standards—AI can assist with document checks but cannot perform the physical verification itself. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Certification standards such as LEED require documented sign-off by qualified professionals, and verification of system compliance carries liability exposure if done incorrectly. Building codes and green certification frameworks create regulatory and legal barriers to full automation without licensed professional involvement. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Certification compliance often requires sign-off by qualified/certified professionals, and liability for faulty HVAC installations creates strong barriers to full automation of verification. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI assistance would require significant integration with building documentation systems and expert review oversight, making the all-in cost comparable to or exceeding the cost of a skilled sheet metal worker performing the verification task. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI could cheaply assist with documentation review and checklist generation, but the core physical verification still requires a paid skilled worker on-site, limiting overall cost savings. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No mature AI product reliably performs HVAC compliance verification against LEED standards independently. Some tools exist for code reference and documentation parsing, but production systems do not yet perform credible end-to-end verification without expert oversight. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs on-site HVAC installation verification against LEED standards; this remains a field inspection task requiring physical presence and expertise. |
Fabricate ducts for high efficiency heating, ventilating, and air conditioning (HVAC) systems to maximize efficiency of systems.
20CI 10–30 · exposure 13 · augmentation 50 · importance 4.2/5 · click for rater detail
Fabricate ducts for high efficiency heating, ventilating, and air conditioning (HVAC) systems to maximize efficiency of systems.
20| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Sheet metal fabrication, particularly in small-to-medium job shops and on-site fabrication for HVAC, remains labor-intensive and low-digitization in most sectors. Adoption of advanced automation has been slow; most shops use semi-automated equipment with substantial human oversight rather than autonomous systems. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Construction and skilled trades are among the slowest sectors to adopt AI/automation for physical fabrication work, with low digitization of hands-on tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-assisted design optimization (thermal modeling, duct layout generation) and automated nesting software can improve worker productivity in planning phases, and CNC programming assistance reduces setup time. However, the hands-on fabrication and quality control remain largely manual, limiting transformative augmentation. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-enabled CAD/CAM software and duct design optimization tools can assist in planning and sizing ducts for efficiency, improving the design phase even though physical fabrication remains manual. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While CAD design and some layout planning can be automated, the actual fabrication requires physical manipulation of sheet metal, precise tooling adjustments, and real-time quality assessment that current robotics cannot reliably handle end-to-end. The task involves bending, cutting, and joining operations that demand dexterity and contextual problem-solving beyond current deployed automation. |
| Task automatability | claude-sonnet-5 | 1/5 | Physical fabrication of ductwork—cutting, bending, welding, seaming sheet metal—requires manual dexterity and physical manipulation that current AI systems cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | HVAC duct fabrication must meet building codes and performance standards, and poor quality ducts directly affect system efficiency and safety; however, no explicit licensing requirement mandates human sign-off, only that the work meet code—creating some organizational friction and quality accountability but not a hard legal barrier. |
| Adoption barriers | claude-sonnet-5 | 3/5 | While not licensed like some trades, HVAC installation often requires code compliance, inspection sign-off, and safety standards that favor human tradespeople, plus physical presence is inherently required. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Industrial sheet metal cutting and bending equipment requires high capital investment and ongoing maintenance, and still needs skilled operators for setup and oversight. The all-in cost per duct currently remains comparable to or higher than skilled labor due to equipment amortization and integration complexity. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute for physical fabrication, so AI cost is effectively infinite relative to human labor for this task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some specialized CNC machines can cut and bend sheet metal, but they require significant setup and operator oversight. No current system reliably performs the full duct fabrication workflow—design optimization, material selection, cutting, bending, joining, and quality verification—without human intervention and correction. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product fabricates HVAC ductwork; robotic sheet metal fabrication exists only in narrow industrial automation contexts, not as general AI-driven solutions replacing skilled workers. |
Fasten seams or joints together with welds, bolts, cement, rivets, solder, caulks, metal drive clips, or bonds to assemble components into products or to repair sheet metal items.
20CI 10–30 · exposure 13 · augmentation 25 · importance 4.2/5 · click for rater detail
Fasten seams or joints together with welds, bolts, cement, rivets, solder, caulks, metal drive clips, or bonds to assemble components into products or to repair sheet metal items.
20| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of fastening automation is concentrated in high-volume automotive and aerospace manufacturing; small shops, field repair, and maintenance work remain largely manual. Public data shows slow penetration of AI/robotics in general sheet metal sectors, with many firms still relying on skilled human labor. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Construction and metal fabrication trades are among the slowest sectors to adopt AI/robotics due to low digitization, variable job sites, and capital constraints for small firms. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI offers limited assistance for this task; joint assessment and fastening method selection remain largely human-driven, and real-time feedback on weld quality or fit is not yet widely integrated into augmentation tools. Welding simulation and design tools exist but do not meaningfully boost human productivity during the assembly or repair phase itself. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with limited functions like optimizing cut layouts or diagnosing weld defects via imaging, but offers little direct assistance during the physical fastening/joining process itself. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While robotic welding and some fastening automation exist in controlled settings, the task requires assessing fit, selecting appropriate fastening methods, and adapting to irregular seams or repair contexts. Current AI systems cannot autonomously handle the full variability of materials, joint geometries, and quality inspection required to meet the ≥50% time-saving bar across diverse sheet metal work. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical manipulation of metal components with tools like welders, riveters, and fasteners in variable real-world conditions, which current AI systems cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Some regulatory and quality oversight requirements exist in aerospace and automotive (structural integrity verification), but no blanket licensing mandate requires a human to perform fastening. However, customer preference for skilled craftwork and in-situ repair contexts create organizational friction to full substitution. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing typically required for most sheet metal work, but liability for structural/safety integrity of welds and joints, plus physical workspace constraints, create moderate friction against automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Specialized robotic fastening equipment carries high capital and integration costs, while sheet metal work often involves smaller batches or field repairs where labor costs per unit remain competitive. For generalist fastening across the range of methods listed, AI cost per task-equivalent currently exceeds loaded human wage. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Specialized robotic welding cells are costly to program and reconfigure for varied jobs, making them more expensive than a human worker for the flexible, varied tasks described here. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Robotic welding systems are deployed in some manufacturing settings, but they operate in highly structured environments with pre-positioned parts. End-to-end automated fastening that spans welds, bolts, cement, rivets, solder, caulks, and repair work—adapting to damage assessment and variable joint conditions—lacks reliable production systems today. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed general-purpose product performs full sheet metal joining/fastening reliably; welding robots exist only for narrow, pre-programmed, fixed industrial setups, not general repair/assembly tasks. |
Finish parts, using hacksaws or hand, rotary, or squaring shears.
19CI 15–24 · exposure 8 · augmentation 25 · importance 2.9/5 · click for rater detail
Finish parts, using hacksaws or hand, rotary, or squaring shears.
19| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Sheet metal finishing remains largely manual in small to medium shops; adoption of robotic or AI-driven automation is slow and limited to high-volume, standardized production environments, reflecting the sector's low overall digitization and capital constraints. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Sheet metal fabrication is a low-digitization, physical trade sector where AI adoption for hands-on tasks like manual cutting/finishing is minimal to nonexistent. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI offers minimal assistance for the hands-on cutting and finishing work itself, though computer vision for part inspection or optimization of cutting patterns could provide modest support to the worker's productivity. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with generating cut plans, nesting layouts, or work instructions beforehand, but offers little real-time assistance during the physical hand-tool finishing itself. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Current AI cannot physically operate hacksaws or shears to finish parts. While CNC and robotic systems exist for metal finishing, this task requires manual dexterity, tactile feedback, and real-time adjustment in a physical environment—capabilities that today's AI lacks for reliable end-to-end execution without extensive custom engineering. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical manual-dexterity task requiring hand tools to cut and finish metal parts; no current AI system can perform the physical cutting/finishing itself. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | No legal licensing requirement exists for this task, but physical workspace constraints, tooling variability, and the need for human oversight of quality and precision create moderate organizational friction to automation adoption. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement specifically restricts this task, but physical presence, tactile skill, and workplace safety practices create practical barriers to remote or software-based substitution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The capital cost of robots, integration, and maintenance for this task far exceeds the loaded wage of a skilled sheet metal worker, especially for variable or small-batch work that requires frequent tool or technique changes. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this physical task, so any AI-based approach (e.g., specialized robotics) would be far more costly than a skilled worker with hand tools. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product performs manual sheet metal finishing with handheld tools. Robotic metal finishing exists in research and specialized industrial settings, but not as a general-purpose off-the-shelf system that can be applied to the task as stated. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs freeform hand-tool metal finishing; this remains a purely manual craft task with no robotics product in general production use for this specific work. |
Hire, train, or supervise new employees or apprentices.
15CI 5–25 · exposure 13 · augmentation 38 · importance 4.0/5 · click for rater detail
Hire, train, or supervise new employees or apprentices.
15| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Sheet metal work is a skilled trades sector with apprenticeship systems, union participation, and small-to-medium firm structures that adopt new technology slowly. Hiring and training remain largely manual, with limited automation visible in production. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Construction and skilled trades sectors show low AI adoption for management and supervisory functions, remaining highly manual and relationship-based. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist supervisors by drafting job postings, summarizing candidate profiles, generating standardized training materials, and flagging scheduling conflicts, raising efficiency in administrative portions of hiring and onboarding without displacing supervisor judgment. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with scheduling, training materials, or administrative aspects of hiring, but core supervisory and mentorship functions remain unaided. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with candidate screening and training content generation, the task requires live judgment in hiring decisions, relationship-building, performance feedback, and legal compliance that demand human discretion. Even with strong AI assistance, a human supervisor must remain the final decision-maker and mentor. |
| Task automatability | claude-sonnet-5 | 1/5 | Hiring, training, and supervising apprentices requires interpersonal judgment, mentorship, and physical skill assessment that current AI cannot perform end-to-end.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong barriers exist: employment law (hiring discrimination, wage compliance), apprenticeship licensing and union agreements in many jurisdictions, liability for training quality and worker safety, and organizational culture favoring direct supervisor accountability in hiring and development decisions. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Hiring and supervision involve legal responsibilities (labor law, safety compliance, apprenticeship certification) requiring accountable human judgment and authority. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The full task—screening candidates, conducting interviews, providing ongoing mentorship, assessing competency, and managing apprentice development—involves substantial human oversight. Current AI tools reduce clerical work but do not displace the core labor cost of a supervisor or trainer. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI cannot substitute for the human supervisory role, so no meaningful cost comparison favors AI for this task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed AI product reliably performs end-to-end hiring, training, or apprenticeship supervision at scale. Existing HR tools handle narrow pieces (resume parsing, onboarding documents) but lack the interpersonal, contextual, and mentoring judgment this role demands. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously hires, trains, or supervises tradespeople; this remains a human management function. |
Perform building commissioning activities by completing mechanical inspections of a building's water, lighting, or heating, ventilating, and air conditioning (HVAC) systems.
15CI 5–25 · exposure 13 · augmentation 50 · importance 3.4/5 · click for rater detail
Perform building commissioning activities by completing mechanical inspections of a building's water, lighting, or heating, ventilating, and air conditioning (HVAC) systems.
15| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Building and construction sectors have slower AI adoption than information or finance; commissioning remains a specialized, hands-on domain where AI pilots are rare and production deployment is minimal. Most organizations still rely on traditional inspector workflows. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Construction and building trades are among the slowest sectors to adopt AI-driven automation for physical inspection tasks, with adoption largely limited to digital documentation tools. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can meaningfully assist by analyzing sensor logs, flagging anomalies, organizing inspection checklists, and documenting findings, but the human inspector remains essential for physical verification, judgment calls, and sign-off on system readiness. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist with generating inspection checklists, analyzing sensor/IoT data from HVAC systems, and documenting findings, but the physical inspection itself remains human-performed. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Building commissioning requires physical inspection and hands-on troubleshooting of complex mechanical systems. While AI could assist with scheduling, documentation, or data analysis of sensor readings, current systems cannot physically inspect ducts, test water pressure, or diagnose heating system failures in situ without human presence and tactile feedback. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical presence, hands-on inspection of ductwork, piping, and mechanical systems, and manual verification of physical installations that current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Building commissioning typically requires licensed mechanical contractors or engineers to certify system performance and sign off on compliance. Liability for system failures and regulatory building code enforcement create legal and contractual barriers to full automation or oversight-free AI substitution. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Building codes, safety regulations, and often licensed inspector sign-off requirements create strong barriers, plus liability concerns around HVAC and water systems failures. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Deploying AI for building commissioning would require significant capital investment in inspection robots, sensors, and integration with building management systems, making the total cost comparable to or higher than human inspectors per visit, especially for small or routine projects. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI cannot substitute for the physical labor and expert judgment involved, so there is no viable cost comparison—human inspectors remain necessary. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some inspection support exists (e.g., thermal imaging analysis, sensor data interpretation), but no deployed AI system performs end-to-end commissioning. Robotics for duct inspection and autonomous diagnostics remain largely research or narrow-domain solutions, not mature production systems at scale. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs physical building commissioning inspections autonomously; sensor-based monitoring exists but full mechanical inspection remains a research/pilot concept at best. |
Fasten roof panel edges or machine-made moldings to structures by nailing or welding.
15CI 5–25 · exposure 13 · augmentation 25 · importance 3.0/5 · click for rater detail
Fasten roof panel edges or machine-made moldings to structures by nailing or welding.
15| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Sheet metal work remains predominantly on-site, physically variable, and low-digitization construction work. Adoption of robotic fastening systems in production is minimal; the sector remains heavily manual and fragmented. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Construction and skilled trades are among the slowest sectors to adopt AI/robotics for physical field tasks due to variability of job sites, low digitization, and lack of scalable robotic solutions. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI offers minimal augmentation for the core fastening task itself, though assisted design or layout planning could help at preparation stages. The immediate physical task of fastening provides little opportunity for AI-assisted productivity gains. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI provides minimal direct assistance to the physical act of nailing or welding, though some digital tools (measurement apps, project management) offer marginal planning support. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | This task involves physical fastening (nailing or welding) of roof panels to structures, which requires precise manipulation in 3D space, real-time environmental adaptation, and integrated tool control. Current robotic systems can perform welding in controlled, repetitive settings, but adaptation to varied roof geometries, panel alignment, and quality verification remains largely manual. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical trade task requiring dexterity, mobility on roofs/structures, and precise manual welding or nailing that no current AI or robotic system performs autonomously in field conditions. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Safety regulations, worker certification requirements for welding, building code compliance with human inspection, and the need for skilled judgment on structural integrity create meaningful legal and organizational barriers to full automation of this task. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Building codes, safety regulations, and structural liability typically require licensed or certified tradespeople for roofing and structural fastening work, creating strong regulatory and liability barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Specialized robotic systems capable of roof fastening are capital-intensive with high setup and integration costs. The loaded wage of a sheet metal worker is low enough that per-task automation cost remains high relative to direct labor, especially given the variability of job sites. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI/robotic substitute for this task, so any hypothetical automation solution would require expensive custom robotics far exceeding the cost of a skilled sheet metal worker. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While welding robots exist in manufacturing, they operate in controlled factory environments with standardized parts. Deployment on construction sites with variable roof structures, environmental conditions, and safety requirements remains extremely limited in production use. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs on-site fastening of roof panels or moldings; construction robotics remain research-stage or limited to controlled factory settings like panel fabrication, not field installation. |
Maintain equipment, making repairs or modifications when necessary.
13CI 5–21 · exposure 8 · augmentation 38 · importance 4.3/5 · click for rater detail
Maintain equipment, making repairs or modifications when necessary.
13| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Sheet metal manufacturing is capital-equipment-heavy but not yet a fast-mover in AI adoption; most maintenance still relies on human expertise, preventive schedules, and OEM service contracts rather than autonomous or AI-driven systems. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Manufacturing and skilled trades sectors show slow, shallow AI adoption for physical maintenance tasks compared to office/information work. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist with equipment condition monitoring, failure prediction, and maintenance scheduling, helping technicians work more efficiently, though the core hands-on repair and judgment remain human-driven. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with predictive maintenance alerts, manuals lookup, or troubleshooting guidance, but offers limited help with the actual hands-on repair or modification work. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Sheet metal equipment maintenance involves diagnosis of complex mechanical failures, physical manipulation in varied environments, and judgment about when to repair vs. replace—tasks that require embodied problem-solving. AI can assist with diagnostics and documentation but cannot perform end-to-end repairs with the dexterity and real-time adaptive reasoning current systems lack. |
| Task automatability | claude-sonnet-5 | 1/5 | Diagnosing and physically repairing sheet metal fabrication equipment requires hands-on manipulation, disassembly, and mechanical judgment that current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Safety-critical equipment maintenance often requires licensed technicians, manufacturer warranties may mandate human certification, and liability for failed repairs creates strong legal and organizational barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing typically required, but safety liability, physical access needs, and employer risk aversion around equipment damage create moderate friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of AI vision systems, diagnostic tools, and human oversight combined with the remaining need for human technicians to execute repairs makes the total cost comparable to or higher than a skilled sheet metal worker's loaded wage. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute for physical repair labor, so cost comparison favors the human worker entirely; any AI role is only advisory. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product reliably diagnoses and executes physical repairs to sheet metal equipment in production settings; this remains firmly in human or specialized robotics territory that is not yet mature at scale. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously performs physical equipment maintenance and repair; AI-assisted diagnostics exist but the actual repair work remains manual. |
Transport prefabricated parts to construction sites for assembly and installation.
9CI 5–13 · exposure 0 · augmentation 25 · importance 4.2/5 · click for rater detail
Transport prefabricated parts to construction sites for assembly and installation.
9| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Even in digitally advanced sectors, autonomous logistics for construction material transport remains pilot-stage; uptake is slow due to regulatory uncertainty and the physical unpredictability of construction sites. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Construction and skilled trades are among the slowest sectors to adopt AI/automation for physical logistics tasks, with low digitization and no meaningful autonomous transport deployment in this niche. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI can assist with route planning and inventory tracking, but the core task of physically transporting materials to sites offers limited opportunity for AI-human collaboration, as the task is primarily physical movement and placement. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can help with route planning, scheduling, and inventory tracking for parts delivery, but offers minimal direct assistance to the physical transport task itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Transporting physical prefabricated parts to construction sites requires dynamic vehicle operation, navigation, real-world hazard avoidance, and coordination with site logistics—capabilities that current AI systems cannot reliably handle end-to-end in uncontrolled outdoor environments. |
| Task automatability | claude-sonnet-5 | 1/5 | Physical transport of fabricated metal parts to job sites requires driving vehicles and physical logistics that current general-purpose AI cannot perform end-to-end without robotic/vehicle automation far beyond typical deployed systems. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Legal liability for vehicle operation, insurance requirements, workers' compensation frameworks, and site safety regulations create substantial adoption barriers; autonomous vehicles operating on public roads and construction sites face regulatory hurdles that currently require human operators. |
| Adoption barriers | claude-sonnet-5 | 3/5 | Commercial driving requires licensing (CDL for larger loads) and safety regulations, plus site-specific logistics coordination, creating moderate regulatory and liability barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The infrastructure cost of autonomous vehicles capable of construction-site transport, combined with ongoing oversight, liability, and route management, remains far higher than employing human drivers and logistics coordinators. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Human drivers/laborers with trucks remain far cheaper and more flexible than any AI-driven logistics or robotic solution capable of this task today. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed autonomous systems can routinely transport construction materials to job sites and arrange them for assembly without human supervision. While some specialized autonomous vehicles exist in controlled settings, production-grade systems for this general task do not exist. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No mature deployed product autonomously handles pickup, loading, transport, and delivery of sheet metal parts to construction sites; autonomous trucking exists only in narrow pilot corridors, not general job-site logistics. |
Install assemblies, such as flashing, pipes, tubes, heating and air conditioning ducts, furnace casings, rain gutters, or downspouts in supportive frameworks.
7CI 5–10 · exposure 0 · augmentation 25 · importance 4.2/5 · click for rater detail
Install assemblies, such as flashing, pipes, tubes, heating and air conditioning ducts, furnace casings, rain gutters, or downspouts in supportive frameworks.
7| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Sheet metal work is a physical, site-based trade in low-digitization sectors (construction, HVAC) where adoption of advanced automation has been minimal. Pilots of robotic sheet metal installation are rare and not yet in mainstream production use. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Construction and skilled trades are among the least digitized, lowest AI-adoption sectors, with physical installation tasks seeing negligible automation penetration. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist in planning (design optimization, duct routing) or quality verification (image inspection of completed work), but these are peripheral to the core manual installation task. Meaningful augmentation of the hands-on assembly work itself is limited with current technology. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with planning, measurement calculations, or generating installation diagrams beforehand, but offers little real-time assistance during the physical installation itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires physical installation of complex metal assemblies in three-dimensional space with precision fitting. Current AI systems cannot perform end-to-end physical manipulation, alignment, fastening, and verification needed for this hands-on work. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical installation task requiring manual manipulation of metal assemblies in three-dimensional space, cutting, fitting, and fastening—capabilities far beyond current AI systems including robotics. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | The task inherently requires on-site physical presence and skilled human judgment in variable field conditions. Building codes, safety regulations, and liability for structural integrity typically require licensed professionals to perform or certify the work. |
| Adoption barriers | claude-sonnet-5 | 3/5 | While not strictly licensed in all jurisdictions, this trade involves safety codes, building inspections, and physical liability that create real friction, though not a hard professional-licensure requirement everywhere. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Specialized mobile manipulation robots capable of performing sheet metal installation tasks would cost far more than trained sheet metal workers, both in capital and operational expenses, making AI significantly more expensive per task completion. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI/robotic substitute performing this physical installation, so any hypothetical automation would be far more costly than a skilled tradesperson today. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI products can autonomously install metal assemblies into supportive frameworks. The task demands embodied robotics capabilities that remain largely confined to research labs and highly controlled factory environments, not field installation work. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product installs flashing, ducts, or gutters autonomously; construction robotics remains research-stage for such varied, unstructured installation work. |
Maneuver completed roofing units into position for installation.
7CI 5–10 · exposure 0 · augmentation 13 · importance 3.9/5 · click for rater detail
Maneuver completed roofing units into position for installation.
7| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Construction remains a low-digitization, fragmented sector with slow automation adoption; roofing installation specifically involves custom, site-specific work with high variation that discourages rapid AI/robotics deployment. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Construction and roofing trades are among the least digitized, lowest-AI-adoption sectors, with physical, unstructured environments resisting automation and no measurable AI displacement in this niche. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist marginally with planning or visualization of unit positioning before installation, but current systems offer limited real-time assistance for the actual maneuvering task on site. |
| Augmentation potential | claude-sonnet-5 | 1/5 | Current AI tools offer no meaningful real-time assistance for physically maneuvering and positioning heavy roofing units; the task is purely manual/mechanical. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Maneuvering heavy roofing units into precise position requires real-time spatial reasoning, force control, and adaptation to uneven surfaces—capabilities that current AI systems cannot execute in physical form without extensive custom hardware and site-specific engineering. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical materials-handling task requiring lifting, carrying, and precise positioning of heavy roofing units on rooftops or job sites; no off-the-shelf AI or robotic system performs this end-to-end today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Safety liability (fall risk, injury), union labor agreements in many regions, and the requirement for on-site judgment and physical presence create substantial barriers to automation, though not absolute legal prohibition. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing law mandates a human perform this specific step, but safety regulations, site variability, and liability for improper installation create real practical friction against automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Custom robotics or teleoperated systems capable of this task would cost far more than the loaded wage of a skilled sheet metal worker per installation, with high maintenance and poor generalization across sites. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI/robotic substitute for this physical task, so any hypothetical automation (custom robotics, cranes with advanced control) would be far more costly than a human worker or crew doing it directly. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed commercial products reliably perform autonomous positioning of roofing units on job sites; this remains a manual task with no mature AI/robotics solution in production at scale in the construction industry. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed commercial product maneuvers completed roofing units into place; construction robotics remains research-stage and limited to narrow, controlled tasks like layout or fastening, not full unit positioning. |
Fabricate or alter parts at construction sites, using shears, hammers, punches, or drills.
7CI 5–10 · exposure 0 · augmentation 25 · importance 3.9/5 · click for rater detail
Fabricate or alter parts at construction sites, using shears, hammers, punches, or drills.
7| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Sheet metal work in construction is performed by small to mid-sized teams in physically heterogeneous environments with limited digitization. The sector lags in AI adoption; automation remains confined to factory preprocessing rather than site-based work. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Construction trades are among the least digitized and slowest to adopt AI/robotics for hands-on fabrication work, with minimal production deployment of automation for this task. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI offers minimal assistance for the core physical fabrication task. Some potential exists for design optimization or tool-path planning before site work, but such pre-work augmentation is peripheral to the hands-on fabrication itself. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with design specs, cut lists, or measurements via software, but it offers little real-time assistance for the physical fabrication and alteration process itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires physical manipulation of tools (shears, hammers, punches, drills) in unstructured construction site environments. Current AI systems lack the embodied robotics, dexterity, and real-time sensorimotor adaptation needed to perform sheet metal fabrication or alteration independently. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical manipulation of sheet metal with hand and power tools at variable construction sites, a domain far beyond current robotics or AI capability outside narrow controlled settings. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Union representation, apprenticeship/licensing requirements, liability for structural integrity of fabricated parts, and the legal requirement for qualified personnel to certify work create strong adoption barriers. Construction standards and safety regulations typically mandate human sign-off. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No strict licensing mandates a human specifically, but safety codes, liability for structural work, and unpredictable site conditions create meaningful practical barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Acquiring, maintaining, and deploying robots capable of sheet metal work at construction sites would be substantially more expensive than the loaded wage of a skilled sheet metal worker, particularly when accounting for setup, safety, and site variability. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI/robotic substitute performing this physical trade task, so any comparison would favor the human worker by default given absent automation options. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI products reliably perform on-site sheet metal fabrication with hand tools today. While robotic sheet metal cutting exists in controlled factory settings, mobile autonomous systems capable of site-based fabrication work remain research-stage. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs on-site sheet metal fabrication or alteration; even advanced construction robotics remain research-stage or limited to fixed factory operations. |
Shape metal material over anvils, blocks, or other forms, using hand tools.
7CI 5–10 · exposure 0 · augmentation 25 · importance 3.6/5 · click for rater detail
Shape metal material over anvils, blocks, or other forms, using hand tools.
7| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Sheet metal work remains a traditional, physically-intensive trade with low digital adoption. Sectors employing this task (small fabrication shops, artisan work, custom manufacturing) move slowly on automation and lack the infrastructure for AI/robotic deployment. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Manual trades like sheet metal work are a low-digitization, physical craft sector with minimal AI/robotic adoption for this specific freeform task. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI offers minimal assistance to a worker actively shaping metal by hand; computer vision or design software might help with planning, but provides little real-time support during the actual hand-tool shaping process. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could assist with design specs, pattern generation, or instructional guidance, but offers little real-time assistance during the physical hand-shaping process itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Shaping metal by hand over anvils requires fine motor control, real-time tactile feedback, and physical dexterity that current AI systems and robots cannot reliably replicate. This is a fundamentally physical task requiring adaptive force application and spatial judgment that falls far outside current automation capabilities. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical hand-tool metalworking task requiring tactile feedback and manual dexterity that current AI systems cannot perform; robotics for this specific freeform shaping is not deployed at scale. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | This task has substantial barriers: it requires hands-on physical execution, significant safety hazards (hot metal, sharp edges, heavy tools), and union/apprenticeship requirements in many jurisdictions that legally protect skilled worker roles in metalworking. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement blocks automation, but the physical dexterity, variable material behavior, and skilled craftsmanship create substantial practical barriers to substitution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | A fully autonomous system capable of hand-tool metal shaping does not exist commercially; the capital cost of custom robotics and control systems would far exceed the loaded wages of a skilled sheet metal worker performing this task. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI/robotic substitute performing this task at any cost, so AI is not cheaper—it's not a functioning alternative at all. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI or robotic systems reliably perform manual metal shaping with hand tools at production quality. While industrial robots exist for some metalworking, they operate under rigid, pre-programmed conditions and cannot adaptively shape metal as a skilled tradesperson does with hand tools. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No commercial product performs freeform hand-tool metal shaping over anvils/blocks; this remains outside deployed robotic or AI capability. |
Related occupations — Construction & Extraction
How to read this
A high substitution score does not mean this job disappears — it means a large share of its current tasks face replacement pressure, so the mix of tasks is likely to change. High augmentation alongside substitution typically means the occupation reorganizes around the protected tasks. Wide confidence intervals mean the rater panel disagreed: treat those scores as open questions, not verdicts.
What would change this score
New model capabilities (automatability, feasibility), falling inference costs (cost ratio), regulation and licensing shifts (barriers), and measured sector adoption (velocity) all re-enter at every index release. Each release is recomputed, versioned and kept queryable — scores are claims with a date on them, not permanent labels.