Maintenance and Repair Workers, General

49-9071.00
Median wage $49,590/yr1,529,700 employed (US)Rank #583 of 923 scored · top 63% by substitution

Perform work involving the skills of two or more maintenance or craft occupations to keep machines, mechanical equipment, or the structure of a building in repair. Duties may involve pipe fitting; HVAC maintenance; insulating; welding; machining; carpentry; repairing electrical or mechanical equipment; installing, aligning, and balancing new equipment; and repairing buildings, floors, or stairs.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution24
Exposure18
Augmentation43

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

27 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

7%

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%19

panel mean rating 1.8/5 → substitution pressure 19/100

Technical feasibility todayw 20%15

panel mean rating 1.6/5 → substitution pressure 15/100

Cost vs. human wagew 15%16

panel mean rating 1.6/5 → substitution pressure 16/100

Adoption barriersw 20%inverted — strong barriers lower the score50

panel mean rating 3.0/5 (barrier strength) → substitution pressure 50/100

Sector adoption velocityw 10%15

panel mean rating 1.6/5 → substitution pressure 15/100

Task breakdown (27 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.

Record type and cost of maintenance or repair work.

75

CI 72–77 · exposure 75 · augmentation 75 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Maintenance-heavy sectors (manufacturing, facilities, utilities) are adopting predictive and digital-first maintenance workflows, but adoption remains mixed; small shops and manual work environments lag significantly.
Sector adoption velocityclaude-sonnet-53/5Facilities and maintenance sectors are moderately digitizing with CMMS adoption growing, but many smaller shops still use paper or basic spreadsheets, giving middling overall velocity.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can assist technicians by auto-filling fields from photos or voice notes, suggesting cost codes, and flagging missing or anomalous entries, materially speeding up record completion without removing human judgment.
Augmentation potentialclaude-sonnet-54/5AI-enabled apps can auto-populate cost data, suggest categorization, and transcribe technician notes, meaningfully speeding up record-keeping while the worker still verifies accuracy.
Task automatabilityclaude-haiku-4-5-202510014/5Recording maintenance and repair work type and cost is largely data-entry and classification. Current AI can extract structured information from work orders, receipts, and technician notes with high accuracy, then populate databases or spreadsheets, achieving >50% time savings with minimal human oversight.
Task automatabilityclaude-sonnet-54/5Recording maintenance type and cost is largely structured data entry that can be automated via mobile CMMS apps, voice-to-text, or AI-assisted forms with high time savings and equal quality.
Adoption barriersclaude-haiku-4-5-202510012/5Few legal or regulatory barriers exist to automating this clerical task. The main friction is organizational (integration with legacy systems, staff familiarity) and potential need for human sign-off on accuracy, but nothing prevents substitution.
Adoption barriersclaude-sonnet-51/5No licensing or regulatory requirement mandates a human specifically record this data; it's routine administrative documentation with minimal liability concerns.
Cost vs. human wageclaude-haiku-4-5-202510014/5OCR, classification, and database entry via AI cost pennies per record compared to a technician's hourly wage for manual data transcription; once integrated into existing maintenance systems, overhead is minimal and scales.
Cost vs. human wageclaude-sonnet-54/5Digital logging tools are cheap relative to technician time spent on paperwork, and once integrated the marginal cost of software-assisted entry is very low compared to manual admin time.
Technical feasibility todayclaude-haiku-4-5-202510014/5Deployed products including document AI, RPA platforms, and ERP integrations already perform invoice parsing and work-order classification reliably in maintenance management software. Error rates are low for standard categories, though edge cases may need human verification.
Technical feasibility todayclaude-sonnet-54/5CMMS and field service management products (e.g., UpKeep, Fiix, ServiceTitan) already let technicians log work and costs digitally, often with voice or template-based entry, deployed widely in production.

Order parts, supplies, or equipment from catalogs or suppliers.

72

CI 65–79 · exposure 70 · augmentation 75 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510014/5Maintenance and facilities management sectors show steady adoption of integrated digital platforms and automated procurement, especially in manufacturing, hospitality, and large commercial operations. Cloud-based maintenance management systems with ordering automation are becoming standard.
Sector adoption velocityclaude-sonnet-52/5Maintenance and repair is a physically-oriented, often small-organization trade with lower overall digitization and slower enterprise software adoption compared to information/finance sectors.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can greatly assist by auto-suggesting correct part numbers based on equipment type and failure mode, flagging supply delays, comparing prices across suppliers, and alerting workers to availability issues—all while the maintenance worker retains final approval and control.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist workers by searching catalogs, comparing prices, identifying compatible parts, and drafting purchase orders, significantly speeding up the sourcing process even if a human confirms the final order.
Task automatabilityclaude-haiku-4-5-202510014/5Ordering parts from catalogs or suppliers is largely procedural: identifying part numbers, checking inventory, processing purchase orders, and managing delivery. Current AI systems can extract part specifications from work orders, search supplier catalogs, compare prices, and generate orders with minimal human intervention, achieving >50% time savings.
Task automatabilityclaude-sonnet-54/5Ordering parts from catalogs is a structured, transactional task involving identifying part numbers, comparing suppliers, and placing orders—well within reach of current AI/automation systems integrated with procurement software or agentic web tools.
Adoption barriersclaude-haiku-4-5-202510012/5Few legal barriers exist; most firms can substitute AI ordering without licensing requirements. Minor friction includes supplier relationship management and occasional need for human judgment on non-standard parts, but no hard regulatory or liability barriers prevent automation.
Adoption barriersclaude-sonnet-52/5No licensing requirement blocks automated ordering, though organizations may want human oversight to avoid costly ordering errors or mismatched parts, creating mild friction.
Cost vs. human wageclaude-haiku-4-5-202510015/5AI-driven ordering (API integration, automated form-filling, catalog searching) costs pennies per transaction, while a maintenance worker's time for manual ordering costs $15–30+ per order all-in. The ratio is highly favorable to automation.
Cost vs. human wageclaude-sonnet-54/5Automated ordering via integrated inventory/procurement systems is far cheaper per transaction than a technician's time spent researching and calling suppliers, though initial system setup and catalog integration carry some cost.
Technical feasibility todayclaude-haiku-4-5-202510014/5Mature procurement and supply-chain software with AI-assisted search and ordering exists in production (e.g., integrated maintenance management systems, supplier portals with ML-assisted part matching). Some friction remains around supplier authentication and approval workflows, but the core task is reliably automated in many organizations.
Technical feasibility todayclaude-sonnet-53/5Procurement software with automated reordering, e-commerce integrations, and AI agents exist and are used in some facilities, but many maintenance shops still rely on manual ordering via phone, in-person, or informal processes with humans verifying part compatibility.

Estimate costs to repair machinery, equipment, or building structures.

51

CI 35–67 · exposure 45 · augmentation 63 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Adoption is emerging in large facilities and service organizations with digitized maintenance records, but small-to-medium repair shops and field maintenance teams lag. Pilots are common among insurers and fleet managers, but production displacement remains spotty and modest.
Sector adoption velocityclaude-sonnet-52/5Maintenance and repair trades are physical, fragmented, small-business-dominated sectors with low AI adoption to date.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can draft cost estimates, flag unusual conditions requiring inspection, cross-reference historical data, and accelerate quote generation, allowing human estimators to focus on complex judgment calls and site-specific variables. This productivity lift is already being realized in managed service environments.
Augmentation potentialclaude-sonnet-53/5AI tools can help pull historical cost data, generate estimate templates, and speed up documentation, meaningfully assisting the estimator even though final judgment remains human.
Task automatabilityclaude-haiku-4-5-202510014/5AI systems can now process diagnostic data, images, and historical repair records to estimate costs with near-human accuracy. Computer vision can assess damage, and large language models can reason over parts catalogs and labor rates, achieving substantial time savings while maintaining comparable quality to human estimates in routine cases.
Task automatabilityclaude-sonnet-52/5Cost estimation for physical repairs requires on-site inspection, judgment about hidden damage, and knowledge of local material/labor pricing, which current AI cannot independently perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510012/5Few regulatory or legal barriers prevent automation of cost estimation itself, though liability for downstream repair decisions based on inaccurate estimates creates some friction. Most barriers are organizational (preference for human judgment, trust in estimators) rather than legal or licensing requirements.
Adoption barriersclaude-sonnet-52/5No licensing typically required for informal estimates, though liability for inaccurate estimates and customer trust in a human assessor create moderate friction.
Cost vs. human wageclaude-haiku-4-5-202510014/5AI inference and integration costs are low relative to the labor cost of a skilled estimator who must inspect sites, consult catalogs, and produce detailed quotes. Once trained systems are in place, marginal cost per estimate is an order of magnitude cheaper than a human visit and estimate.
Cost vs. human wageclaude-sonnet-52/5AI could cheaply generate cost ranges from historical data, but human inspection and validation are still required, so overall cost savings versus a skilled technician's estimate are modest.
Technical feasibility todayclaude-haiku-4-5-202510013/5Commercial products exist for cost estimation in maintenance contexts (e.g., damage assessment AI, automated quoting tools), but they require structured input, struggle with novel or complex scenarios, and typically need human review before deployment. Production use is growing but material error rates and scope limitations remain.
Technical feasibility todayclaude-sonnet-52/5Some estimating software and AI-assisted tools exist for construction cost estimation, but general maintenance/repair cost estimation across diverse machinery and structures is not reliably automated in production.

Plan and lay out repair work, using diagrams, drawings, blueprints, maintenance manuals, or schematic diagrams.

37

CI 30–44 · exposure 33 · augmentation 75 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Maintenance and repair is a traditional, physically distributed sector with lower digitization; adoption of AI-assisted planning is emerging in large facilities and fleet operations but remains uncommon in small to mid-sized repair shops.
Sector adoption velocityclaude-sonnet-52/5Maintenance and repair trades are a physically-oriented, lower-digitization sector where AI tool adoption for planning tasks is still nascent and mostly pilot-stage.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can substantially assist technicians by automatically extracting relevant diagrams from manuals, flagging potential hazards from schematics, and generating preliminary layouts that technicians refine, significantly reducing plan-preparation time while keeping human expertise in the loop.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully assist by quickly summarizing manuals, interpreting blueprints, and suggesting repair sequences, speeding up the planning phase even though a human executes and verifies the physical work.
Task automatabilityclaude-haiku-4-5-202510013/5AI can interpret technical diagrams and generate layout plans for routine repair scenarios, but real-world repair work often involves incomplete documentation, site-specific constraints, and unforeseen conditions that require human judgment and on-site assessment to achieve equal-quality outcomes.
Task automatabilityclaude-sonnet-52/5AI can help interpret manuals and generate plans from text/diagrams, but translating this into an actionable physical repair layout in a specific site context requires spatial reasoning and hands-on verification current AI cannot fully replace.
Adoption barriersclaude-haiku-4-5-202510013/5Repair work often occurs in regulated or safety-critical contexts (electrical, HVAC, equipment); liability and customer preference for human expertise create friction, though the task itself does not require formal licensing in most jurisdictions.
Adoption barriersclaude-sonnet-52/5No licensing generally required for planning repair work itself, though safety-critical systems (electrical, structural) may require certified oversight, creating some friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI infrastructure and oversight for repair planning is still relatively expensive compared to a skilled technician's hourly rate, especially when accounting for integration, validation, and the cost of errors in complex repairs.
Cost vs. human wageclaude-sonnet-52/5Using AI for planning assistance is cheap per query, but the overall task still requires a human technician's site visit and judgment, so all-in cost savings versus the human's total task time are modest.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI can digitize and analyze static diagrams and blueprints, existing products lack reliability in translating 2D/3D technical drawings into actionable repair sequences in production settings; most deployed solutions are narrow document-parsing tools rather than end-to-end repair planning systems.
Technical feasibility todayclaude-sonnet-52/5Some products (multimodal LLMs, AR overlays) can assist with reading schematics and suggesting steps, but no deployed system reliably plans full repair layouts across diverse equipment in production settings.

Inspect used parts to determine changes in dimensional requirements, using rules, calipers, micrometers, or other measuring instruments.

31

CI 28–35 · exposure 25 · augmentation 38 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of automated inspection in maintenance and repair is still in pilot stages, concentrated in high-volume manufacturing. General maintenance workers in smaller shops and field settings have shown limited uptake of AI-based measurement tools relative to established practices.
Sector adoption velocityclaude-sonnet-51/5General maintenance and repair is a low-digitization, physical-labor sector with minimal AI/automation penetration for ad-hoc inspection tasks; adoption is essentially absent outside large manufacturing lines.
Augmentation potentialclaude-haiku-4-5-202510013/5AI vision tools can assist workers by flagging potential out-of-spec measurements, flagging wear patterns, or recommending when deeper inspection is needed, improving their efficiency and consistency without replacing their judgment on parts disposition.
Augmentation potentialclaude-sonnet-52/5Digital calipers/micrometers with data logging and simple AI-assisted defect flagging can help record and compare measurements, but this offers only modest productivity gains over traditional manual measurement and judgment.
Task automatabilityclaude-haiku-4-5-202510012/5While AI vision systems can perform dimensional measurement analysis, the task requires physical inspection of used parts with hand-held precision instruments (calipers, micrometers) in variable environmental conditions. End-to-end automation would need robotics for consistent physical measurement and interpretation of part wear patterns, which current off-the-shelf systems cannot reliably achieve without significant custom integration.
Task automatabilityclaude-sonnet-52/5Physical measurement of used parts requires manual handling and instrument use; while some vision-based measurement systems exist, general-purpose AI cannot yet perform this hands-on inspection end-to-end.6, 60%+ time saving is not realistic across the diverse contexts general maintenance workers face.
Adoption barriersclaude-haiku-4-5-202510013/5Some regulatory and safety contexts require documented inspection by certified technicians, and many maintenance workflows depend on immediate physical feedback that human judgment provides. However, barriers are not absolute legal requirements in all sectors, creating moderate friction rather than hard prohibition.
Adoption barriersclaude-sonnet-52/5No licensing requirement dictates a human must take these measurements, but liability for missed defects and the physical nature of part handling create moderate friction against pure automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5The equipment cost (specialized cameras, lighting, robotics, custom software) plus integration and human oversight overhead exceeds the typical wage cost of a maintenance worker performing manual measurement with standard instruments. AI vision solutions for precision metrology remain expensive relative to skilled labor.
Cost vs. human wageclaude-sonnet-52/5Deploying calibrated automated measurement systems (robotic arms, vision systems, fixtures) for the broad, variable part types in general maintenance work costs more than the marginal cost of a technician using hand tools, unless massively scaled in a single repetitive context.
Technical feasibility todayclaude-haiku-4-5-202510012/5Computer vision systems exist for parts inspection, but deployed products struggle with the variability of used parts, wear assessment, and the need for sub-millimeter precision measurement in real-world conditions. Most production systems require human-in-the-loop validation and are limited to controlled environments or high-contrast parts.
Technical feasibility todayclaude-sonnet-52/5Automated optical/dimensional inspection systems exist in manufacturing quality control, but these are narrow, fixed-setup solutions, not generalizable products deployed for the varied ad-hoc parts inspection general maintenance workers handle.

Inspect, operate, or test machinery or equipment to diagnose machine malfunctions.

30

CI 25–35 · exposure 25 · augmentation 50 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption is slow outside large-scale manufacturing and utilities; small and mid-sized maintenance shops rely on technician expertise and resist automation due to capital costs, regulatory uncertainty, and the craft nature of diagnostics. Pilot programs exist but production displacement remains minimal.
Sector adoption velocityclaude-sonnet-52/5Manufacturing and facilities maintenance sectors are physical, moderately digitized industries with slower AI adoption compared to information/professional services, though IoT-based predictive maintenance is growing in some segments.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist by analyzing sensor data, suggesting probable faults, and flagging anomalies—raising technician productivity in diagnosis—but the human expert must still operate equipment, validate findings, and make final decisions in variable conditions.
Augmentation potentialclaude-sonnet-53/5AI-powered sensors, diagnostic apps, and predictive analytics dashboards can meaningfully assist technicians in narrowing down likely fault causes and prioritizing checks, even though the human still performs physical testing.
Task automatabilityclaude-haiku-4-5-202510012/5Current AI can assist with some diagnostics through image recognition and predictive maintenance models, but requires human presence to operate machinery safely and interpret ambiguous sensor data in diverse, real-world industrial settings. The task involves physical operation and contextual judgment that remains largely human-dependent.
Task automatabilityclaude-sonnet-52/5Diagnosis often requires physical inspection, hands-on operation, and sensory judgment (sounds, smells, vibration, touch) that current AI cannot perform without embodiment; sensor-based diagnostics help but don't cover the full task.
Adoption barriersclaude-haiku-4-5-202510014/5Significant regulatory and safety barriers exist: machinery operation often requires licensed technicians or certification in many jurisdictions, liability for incorrect diagnosis can be substantial, and human sign-off is typically mandated before repairs proceed. Organizational risk aversion is high given potential downtime costs.
Adoption barriersclaude-sonnet-52/5No licensing typically required for general maintenance diagnostics, though liability for missed diagnoses on critical equipment and organizational reliance on human judgment create moderate friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-based diagnostic systems require substantial integration, sensor infrastructure, and ongoing human oversight to validate results. For general maintenance workers across multiple equipment types, all-in AI costs remain comparable to or exceed the loaded wage of experienced technicians.
Cost vs. human wageclaude-sonnet-52/5Sensor-based monitoring systems can be cost-effective for large fleets of specific equipment, but retrofitting sensors, integration, and covering the physical inspection portion for varied general equipment is often costlier than a human doing rounds.
Technical feasibility todayclaude-haiku-4-5-202510012/5Narrow deployments exist for specific equipment types (e.g., predictive maintenance in large manufacturing), but general-purpose AI systems struggle with the diverse machinery, varying failure modes, and safety-critical nature of malfunction diagnosis. Production reliability remains low across heterogeneous environments.
Technical feasibility todayclaude-sonnet-52/5Predictive maintenance and vibration/thermal analytics products exist in some industrial settings, but general-purpose physical inspection and hands-on testing by AI is not deployed at scale for generalist maintenance workers.

Perform general cleaning of buildings or properties.

25

CI 10–40 · exposure 13 · augmentation 13 · importance 3.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of cleaning robots is slow and concentrated in large, high-budget facilities (hospitals, airports, offices). Most small-to-medium buildings and residential cleaning remain labor-intensive; digital integration is low, and cost barriers delay broader rollout.
Sector adoption velocityclaude-sonnet-51/5Building maintenance and janitorial sectors show minimal AI/robotic adoption; this remains a physical, low-digitization task category.
Augmentation potentialclaude-haiku-4-5-202510012/5AI augmentation for cleaning is minimal; cleaning is fundamentally a physical, sensorimotor task where machine assistance (e.g., autonomous floor buffers, leak detection alerts) offers modest productivity gains but does not significantly transform human productivity in typical roles.
Augmentation potentialclaude-sonnet-51/5Current AI offers essentially no meaningful assistance to a human performing general physical cleaning tasks.
Task automatabilityclaude-haiku-4-5-202510012/5Current AI robotics can handle narrow, repetitive cleaning subtasks (vacuuming, floor scrubbing) in controlled environments, but general building cleaning requires adaptive navigation, handling irregular obstacles, responsive decision-making about what needs cleaning, and safe operation around people—capabilities that fall far short of 50% time savings at equal quality today.
Task automatabilityclaude-sonnet-51/5Physical cleaning of buildings requires mobile manipulation, navigation of varied environments, and dexterity that current general-purpose AI systems and robots cannot reliably perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510013/5Cleaning does not require professional licensing, but facilities management practices, worker safety regulations, insurance liability for autonomous equipment failure, and customer preference for visible human presence in occupied buildings create moderate friction against full automation.
Adoption barriersclaude-sonnet-51/5No licensing, liability, or regulatory barriers specifically restrict automation of general cleaning tasks.
Cost vs. human wageclaude-haiku-4-5-202510012/5Specialized cleaning robots cost $15k–$100k+ upfront with ongoing maintenance, training, and integration overhead, while general cleaners earn roughly $25k–$35k annually. The capital and operational costs typically exceed the equivalent labor cost for general cleaning over medium timescales.
Cost vs. human wageclaude-sonnet-51/5Specialized cleaning robots have high capital and maintenance costs relative to low-wage human cleaning labor, and cannot match human versatility across tasks, making AI more expensive per task-equivalent.
Technical feasibility todayclaude-haiku-4-5-202510012/5Cleaning robots exist in narrow, structured domains (airport terminals, parking garages) but deployed systems have significant limitations: they require pre-mapped environments, struggle with stairs and uneven surfaces, and cannot reliably handle the full scope of general building cleaning. Production deployments are rare and spotty.
Technical feasibility todayclaude-sonnet-51/5No mature deployed AI product performs general building cleaning autonomously; robotic vacuums/floor scrubbers exist for narrow subtasks but not general cleaning across varied surfaces and property types.

Design new equipment to aid in the repair or maintenance of machines, mechanical equipment, or building structures.

25

CI 20–30 · exposure 20 · augmentation 50 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of AI-assisted design tools is emerging in specialized sectors (aerospace, automotive) but remains limited and typically augmentative rather than replacing core design roles. Most small maintenance shops and general repair operations have minimal AI tool integration.
Sector adoption velocityclaude-sonnet-52/5Maintenance and repair is a physically-oriented, lower-digitization occupation where AI adoption for custom equipment design is still nascent and largely confined to pilot use of CAD/generative tools.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist by generating initial design concepts, automating routine CAD tasks, or suggesting standard solutions based on past examples, which speeds up the human designer's work. However, the human expert must verify feasibility, safety, and manufacturability, limiting the productivity gain.
Augmentation potentialclaude-sonnet-53/5Generative design and CAD-assist tools can help brainstorm equipment concepts, generate technical drawings, or suggest modifications, meaningfully speeding up parts of the design process while humans retain control.
Task automatabilityclaude-haiku-4-5-202510012/5Designing new equipment requires creative problem-solving, domain expertise, and understanding of specific mechanical constraints that vary case-by-case. Current AI can assist with drafting or suggesting standard solutions but cannot reliably handle the iterative design process, testing, and validation needed for novel equipment without substantial human oversight.
Task automatabilityclaude-sonnet-52/5Designing new equipment requires physical understanding, iterative prototyping, and hands-on validation that current AI cannot fully replace; AI can assist with ideation and CAD drafting but not execute the full design-build-test cycle.
Adoption barriersclaude-haiku-4-5-202510014/5Equipment design often requires professional engineering judgment and accountability; liability for faulty designs typically falls on the responsible engineer. Many jurisdictions require designs to be certified or signed by licensed engineers, creating a strong legal barrier to full automation.
Adoption barriersclaude-sonnet-53/5No licensing requirement specifically bars AI from generating designs, but liability for faulty equipment and organizational reliance on experienced technicians creates moderate friction against pure automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI assistance for design (CAD generation, idea suggestion) has moderate costs, but the total cost including human expert review, iteration, testing, and validation remains high relative to the skilled labor required. Human engineers still dominate the cost structure.
Cost vs. human wageclaude-sonnet-52/5AI tools can cheaply generate design concepts or sketches, but the overall task still requires skilled human engineering judgment, fabrication oversight, and testing, keeping all-in costs comparable to or above human-only work.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI can generate design sketches or CAD-like outputs, no deployed product reliably designs equipment ready for manufacture or use without significant human expertise and revision. Design tools exist but require expert-in-the-loop validation and domain knowledge to produce functional outputs.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously designs custom maintenance equipment for facilities; this remains a human engineering task with occasional CAD/generative-design tool assistance.

Diagnose mechanical problems and determine how to correct them, checking blueprints, repair manuals, or parts catalogs, as necessary.

23

CI 16–30 · exposure 20 · augmentation 50 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption is slow and fragmented; while large industrial facilities may pilot AI-assisted diagnostics, small and medium repair shops and field maintenance remain human-centric with limited digital infrastructure for AI integration.
Sector adoption velocityclaude-sonnet-52/5General maintenance and repair is a low-digitization, physically-embedded trade sector with slow AI adoption limited mostly to reference lookup tools and basic diagnostics apps.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist by cross-referencing manuals, flagging known fault patterns, and accelerating documentation lookup, but the core diagnostic reasoning and physical validation remain human-dependent, making this a partial augmentation rather than transformative.
Augmentation potentialclaude-sonnet-53/5AI chatbots and manual-search tools can help workers quickly find blueprint specs, repair manual sections, or parts information, meaningfully speeding up part of the diagnostic research process.
Task automatabilityclaude-haiku-4-5-202510012/5AI can assist with documentation lookup and pattern matching against known fault codes, but diagnosing mechanical problems typically requires physical inspection, hands-on testing, and contextual judgment about system interactions that current AI cannot reliably perform end-to-end without human verification.
Task automatabilityclaude-sonnet-52/5Diagnosis of physical mechanical faults requires hands-on inspection, sensory feedback, and physical manipulation that current AI cannot perform; AI can assist with information lookup but cannot execute the core diagnostic task.
Adoption barriersclaude-haiku-4-5-202510014/5Safety and liability are high barriers: a misdiagnosis can cause equipment failure, workplace injury, or damage; many jurisdictions require licensed technicians to sign off on repairs, and customers typically expect human accountability for critical diagnostics.
Adoption barriersclaude-sonnet-53/5No licensing barrier for most general maintenance tasks, but physical presence and liability for misdiagnosis leading to equipment failure or safety issues create meaningful friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI diagnostic tools incur significant integration and customization costs per machine type or system, and human oversight remains mandatory; the per-task cost is not clearly below a skilled technician's loaded wage for complex diagnostics.
Cost vs. human wageclaude-sonnet-51/5AI cannot substitute for physical diagnosis, so the relevant cost comparison favors the human worker who must be physically present regardless of any AI assistance costs.
Technical feasibility todayclaude-haiku-4-5-202510012/5Some diagnostic decision-support tools exist (e.g., fault-tree systems for industrial equipment), but they narrow in scope and still require human technicians to gather sensor data, interpret ambiguous symptoms, and validate recommendations in real conditions.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously diagnoses general mechanical equipment problems in the field; existing tools are limited to narrow, sensor-instrumented predictive maintenance systems, not general hands-on troubleshooting.

Provide groundskeeping services, such as landscaping or snow removal.

23

CI 10–35 · exposure 13 · augmentation 38 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of autonomous landscaping and snow removal systems remains slow and concentrated in large institutional grounds (golf courses, universities) or niche applications. Most small to mid-size landscaping businesses and municipalities continue with traditional labor-intensive approaches.
Sector adoption velocityclaude-sonnet-51/5Groundskeeping and building maintenance are low-digitization, physically-oriented sectors with minimal AI/robotic adoption in production settings.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-powered tools can assist with route optimization for snow removal, landscape design visualization, and equipment maintenance scheduling, raising worker productivity on planning aspects. However, the physical execution remains largely human-dependent, limiting overall augmentation impact.
Augmentation potentialclaude-sonnet-52/5Basic tools like scheduling apps, weather forecasting, or route optimization software can somewhat help planning, but there is little AI assistance for the physical execution of the task itself.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can plan landscaping designs and optimize snow removal routes, the actual execution requires physical manipulation of terrain, plants, and equipment in unstructured outdoor environments. Current robotic systems for these tasks are limited in adaptability and reliability, far short of 50% time savings at equal quality for the full task.
Task automatabilityclaude-sonnet-51/5This is a physical outdoor task requiring mobility, dexterity, and equipment operation (mowers, snow blowers, hand tools) in variable terrain and weather; no current AI system can perform this end-to-end.
Adoption barriersclaude-haiku-4-5-202510013/5Groundskeeping is typically not regulated at the professional licensing level, but organizational preference for human workers, customer expectations for human judgment (e.g., plant selection), and safety liability create moderate adoption friction.
Adoption barriersclaude-sonnet-52/5No licensing requirement blocks automation, but physical property access, liability for damage/injury on client property, and unpredictable terrain/obstacles create real friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current robotic landscaping and snow removal equipment is capital-intensive and still requires human supervision and maintenance. The all-in cost (hardware, integration, oversight) remains comparable to or exceeds the loaded wage of a groundskeeping worker for equivalent output.
Cost vs. human wageclaude-sonnet-51/5Robotic groundskeeping equipment (autonomous mowers, snowplows) has high capital and maintenance costs relative to low-wage manual labor, making AI more expensive per task-equivalent today.
Technical feasibility todayclaude-haiku-4-5-202510012/5Robotic lawn mowers and snow removal systems exist but have narrow scope (flat terrain, structured layouts) and require significant human oversight and intervention. No deployed product reliably handles the full range of groundskeeping variability and decision-making in production at scale.
Technical feasibility todayclaude-sonnet-51/5No deployed autonomous product performs general landscaping or snow removal reliably at scale; existing robotic mowers and limited snow-clearing machines are narrow, semi-autonomous, and require human setup and supervision.

Perform routine maintenance, such as inspecting drives, motors, or belts, checking fluid levels, replacing filters, or doing other preventive maintenance actions.

21

CI 10–31 · exposure 13 · augmentation 50 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5While large industrial facilities are piloting condition-monitoring AI and predictive maintenance systems, widespread production adoption remains limited. Most maintenance remains manual and decentralized across small to medium facilities with lower digitization.
Sector adoption velocityclaude-sonnet-51/5Facilities and industrial maintenance is a low-digitization, physical-labor sector with minimal AI/robotic adoption for hands-on preventive maintenance tasks.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-powered diagnostic and monitoring tools (thermal imaging, vibration analysis, fluid analysis recommendations) can meaningfully assist technicians in identifying problems and prioritizing maintenance tasks, though the human still performs the physical work and final judgment.
Augmentation potentialclaude-sonnet-53/5AI-driven predictive maintenance software and IoT sensors can help schedule and prioritize inspections, but the physical execution still relies entirely on the human worker.
Task automatabilityclaude-haiku-4-5-202510012/5While AI-driven robotic systems could theoretically inspect and diagnose some equipment issues, the task requires physical manipulation (replacing filters, checking fluid levels) and contextual judgment about equipment condition that current general-purpose systems cannot reliably perform end-to-end. Most maintenance still requires human hands and site-specific adaptation.
Task automatabilityclaude-sonnet-51/5This is physical inspection and hands-on maintenance requiring manipulation, sensory judgment, and mobility in varied environments—far outside current AI capability without robotics that don't exist at scale.
Adoption barriersclaude-haiku-4-5-202510013/5Routine maintenance often occurs on-site in industrial settings where safety compliance and liability concerns require human sign-off; however, there are no strict legal bars to robotic inspection or condition monitoring in many contexts, creating moderate friction rather than hard prohibition.
Adoption barriersclaude-sonnet-52/5No licensing typically required for general maintenance, but the physical nature of the task itself is the primary barrier rather than regulation.
Cost vs. human wageclaude-haiku-4-5-202510011/5Specialized inspection robots and condition-monitoring systems exist but remain expensive to deploy, configure, and oversee for routine maintenance. The loaded cost of such AI systems currently exceeds the wage of a maintenance technician performing these tasks.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute for the physical labor involved, so any AI cost is irrelevant relative to a human still needing to perform the hands-on work.
Technical feasibility todayclaude-haiku-4-5-202510012/5Some narrow computer vision systems can detect obvious equipment faults (cracked belts, leaks), but deployed products do not reliably perform the full range of routine maintenance tasks—inspection, decision-making, and physical replacement—at production scale in varied real-world environments.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs physical drive/motor/belt inspection or fluid checks; sensor-based condition monitoring exists but doesn't replace the physical task itself.

Clean or lubricate shafts, bearings, gears, or other parts of machinery.

20

CI 5–35 · exposure 13 · augmentation 38 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of autonomous maintenance automation remains slow and concentrated in large manufacturing plants with standardized, high-value equipment. Most small-to-medium enterprises and job shops still rely entirely on human technicians, reflecting laggard adoption outside core automotive/semiconductor sectors.
Sector adoption velocityclaude-sonnet-51/5General maintenance and repair work is a low-digitization, physical-labor-heavy sector with minimal AI/robotic adoption in routine tasks like lubrication.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-assisted tools like condition-monitoring sensors and predictive maintenance software help technicians identify when and where to lubricate, improving scheduling and reducing emergency repairs. However, the physical work of cleaning and lubrication itself sees limited augmentation from current AI systems.
Augmentation potentialclaude-sonnet-52/5AI can assist with predictive maintenance scheduling or diagnostics indicating when lubrication is needed, but does not directly aid the physical act of cleaning or lubricating.
Task automatabilityclaude-haiku-4-5-202510012/5While AI-powered robots could theoretically perform lubrication and cleaning at scale, this task requires navigation of complex machinery, precise identification of components, and physical dexterity in confined spaces that current systems handle inconsistently. End-to-end automation with ≥50% time savings at equal quality is not demonstrated in production today.
Task automatabilityclaude-sonnet-51/5This is a physical, hands-on task requiring manipulation of tools, lubricants, and machinery in varied physical environments; no AI system can perform the physical manipulation involved.'
Adoption barriersclaude-haiku-4-5-202510014/5Factory and facility managers face substantial organizational friction: equipment downtime risk, safety liability if automation fails, regulatory workplace safety requirements around hazardous machinery, and resistance from maintenance unions and incumbent workforces. The physical risk and asset criticality create high barriers to substitution.
Adoption barriersclaude-sonnet-52/5No licensing requirement typically applies, but physical access, safety protocols, and equipment-specific knowledge create moderate organizational friction against automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Robotic maintenance systems require significant capital investment, specialized integration for each machinery type, and ongoing programming, making them more expensive than hiring trained maintenance workers except in very high-volume, standardized environments. Most general maintenance scenarios remain costlier to automate.
Cost vs. human wageclaude-sonnet-51/5There is no AI-only cost structure for this physical task; any automation would require expensive robotics and sensors far exceeding the cost of a human worker doing routine lubrication.
Technical feasibility todayclaude-haiku-4-5-202510012/5Robotic systems for equipment maintenance exist in research and limited pilot phases, but deployed products rarely handle the variability of industrial machinery layouts, component identification under dust/oil, or the judgment needed to detect wear before lubrication. Practical deployment remains narrow and error-prone.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product performs physical cleaning or lubrication of machinery parts; this remains a manual maintenance task performed by humans with occasional robotic assistance only in narrow, controlled industrial settings.

Set up and operate machine tools to repair or fabricate machine parts, jigs, fixtures, or tools.

20

CI 5–35 · exposure 13 · augmentation 38 · importance 3.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of AI-driven machine tool automation in small and mid-sized maintenance shops remains slow; most deployments are in large, high-volume manufacturing where ROI is clearer. General maintenance workers in diverse sectors show low velocity of AI tool adoption today.
Sector adoption velocityclaude-sonnet-51/5General maintenance and repair work is a low-digitization, physical-labor sector with minimal AI/robotic adoption in production environments today.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist with design optimization, CNC code generation, and measurement verification, moderately improving a technician's productivity. However, the hands-on nature of setup and the need for real-time adaptive judgment limit how much AI can transform the core task while the human remains in the loop.
Augmentation potentialclaude-sonnet-52/5AI can assist with diagnostics, CAD/CAM programming guidance, or documentation lookup, but offers little direct help with the physical setup and operation of machine tools.
Task automatabilityclaude-haiku-4-5-202510012/5Machine tool setup and operation involve significant physical manipulation, spatial reasoning, and real-time adjustment to material feedback. While AI can assist in planning and design, current systems cannot reliably handle the full sequence of physical setup, measurement, and adaptive control needed for precision fabrication end-to-end.
Task automatabilityclaude-sonnet-51/5This is a hands-on physical task requiring machine setup, tool operation, and precision manipulation that current AI systems cannot perform without robotic embodiment, which is not generally available for this work.
Adoption barriersclaude-haiku-4-5-202510014/5Physical safety regulations, operator licensing for certain equipment, liability for precision and part quality, and the requirement for on-site troubleshooting and human judgment create substantial barriers to full automation. Many jurisdictions require a licensed operator or responsible human oversight.
Adoption barriersclaude-sonnet-52/5No licensing requirement specifically blocks automation, but physical embodiment, safety requirements around machinery, and lack of technology readiness create substantial practical barriers.
Cost vs. human wageclaude-haiku-4-5-202510012/5Deploying robotic systems capable of flexible tool setup and operation remains capital-intensive (hardware, integration, maintenance) compared to a technician's loaded wage, especially for the variety and low-volume work typical in general maintenance and repair shops.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI-only substitute for this physical task, so any comparison favors the human worker; automation would require expensive specialized robotics far exceeding labor costs for this generalist role.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed products reliably perform complete machine tool setup and operation independently. Robotic systems exist for narrowly scoped tasks (e.g., CNC machining of known designs), but they lack the adaptability and judgment humans bring to tool setup, fixture design, and handling variability in real shop conditions.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product operates machine tools autonomously to fabricate or repair parts in general maintenance settings; this remains far beyond current commercial robotics/AI capability.

Align and balance new equipment after installation.

18

CI 5–30 · exposure 13 · augmentation 38 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Maintenance and repair work remains largely in small, dispersed operations with limited digitization. Adoption of AI for hands-on physical tasks is slow compared to information-intensive sectors, with most facilities still relying on trained human technicians.
Sector adoption velocityclaude-sonnet-51/5Maintenance and repair trades are physical, low-digitization occupations where AI/robotic adoption for hands-on mechanical tasks remains minimal and largely experimental.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-powered measurement and diagnostic tools can assist technicians by suggesting adjustments or flagging misalignment detected by sensors, moderately improving productivity. However, the human remains essential for execution and final sign-off, limiting the depth of augmentation.
Augmentation potentialclaude-sonnet-52/5AI-powered diagnostic software and sensor analytics can help interpret vibration or alignment data to guide technicians, but the actual physical adjustment work sees limited AI-driven productivity enhancement.
Task automatabilityclaude-haiku-4-5-202510012/5Alignment and balancing of new equipment involves physical manipulation and precise measurements in highly variable installation contexts. While some diagnostic measurements can be partially automated with sensors, the hands-on adjustment and calibration steps require human dexterity and real-time problem-solving that current AI systems cannot reliably perform end-to-end.
Task automatabilityclaude-sonnet-51/5Aligning and balancing equipment requires physical manipulation, precision tools (dial indicators, laser alignment tools, vibration analyzers), and hands-on adjustment that current AI cannot perform without embodiment in advanced robotics, which is not generally available.
Adoption barriersclaude-haiku-4-5-202510014/5Equipment manufacturers often require certified technicians to perform alignment and balancing to maintain warranties and safety certifications. Liability for improper alignment (which can cause equipment failure or safety hazards) creates strong regulatory and contractual barriers to full automation.
Adoption barriersclaude-sonnet-53/5While not licensed in the way medical or legal work is, this task often involves safety-critical machinery where improper alignment can cause equipment failure or injury, creating strong organizational and liability-driven preference for skilled human technicians.
Cost vs. human wageclaude-haiku-4-5-202510012/5The specialized equipment, sensors, and calibration tools required would need to be paired with AI systems, and the total integration cost would likely exceed the labor cost of a skilled maintenance worker performing the task directly in most scenarios.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this physical task, so the cost comparison favors the human worker by default since AI cannot deliver the output at all.
Technical feasibility todayclaude-haiku-4-5-202510012/5No mature deployed products perform full alignment and balancing autonomously. Some diagnostic tools and measurement aids exist, but actual execution requires specialized robotic systems (not general AI) or human technicians, limiting the applicability of current general AI to production environments.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously performs physical equipment alignment and balancing; this remains a manual mechanical task performed by skilled technicians using specialized instruments.

Operate cutting torches or welding equipment to cut or join metal parts.

18

CI 5–30 · exposure 13 · augmentation 38 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption is mostly limited to large manufacturers with high-volume, standardized work. Small and mid-market maintenance shops, field operations, and repair-on-demand work remain predominantly manual, with robotic systems too inflexible and capital-intensive for these segments.
Sector adoption velocityclaude-sonnet-51/5Maintenance and repair is a physically intensive, low-digitization sector where AI/robotic adoption for hands-on tasks like welding is minimal and not scaling in production outside narrow manufacturing niches.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-assisted positioning guides, real-time quality monitoring, and decision-support for parameter selection (heat, speed, rod type) can improve worker productivity and reduce defects. However, the human must remain in the loop for safety and final sign-off, so augmentation is useful but not transformative of the core task.
Augmentation potentialclaude-sonnet-52/5AI can assist with related planning tasks like reading schematics, calculating material specs, or training simulations, but it offers little direct real-time assistance during the physical act of cutting or welding itself.
Task automatabilityclaude-haiku-4-5-202510012/5Operating cutting torches or welding equipment requires precise spatial coordination, real-time feedback, and context-specific judgment about metal properties and joint quality. While AI can control robotic arms in structured factory settings, the variety of metal types, joint geometries, and real-world field conditions means current general-purpose systems cannot reliably perform this end-to-end at 50% time savings across typical maintenance scenarios.
Task automatabilityclaude-sonnet-51/5This is a physical manual skill requiring precise hand-eye coordination, torch/welder manipulation, and real-time sensory feedback (heat, molten metal flow) that current AI systems cannot perform end-to-end; robotic welding exists but is confined to fixed industrial setups, not general maintenance work.
Adoption barriersclaude-haiku-4-5-202510014/5Safety regulations, quality certification standards, and liability create meaningful barriers. Welded and cut joints often require inspection and sign-off by certified personnel; many codes and customer contracts explicitly require licensed or certified human approval. Operator error in welding can cause structural failure, creating error-cost asymmetry that deters full automation.
Adoption barriersclaude-sonnet-53/5While not licensed in the same way as medical or legal work, welding often requires certification for safety and quality assurance, and physical hazards (fire, fumes, structural integrity) create liability concerns that favor trained human oversight.
Cost vs. human wageclaude-haiku-4-5-202510012/5Robotic welding systems require significant capital investment ($50K–$500K+), integration, programming, and ongoing maintenance. For one-off or small-batch maintenance and repair jobs, the amortized cost per task typically exceeds the loaded wage of a skilled welder or torch operator.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI-driven substitute for this physical task in general maintenance settings, so any capable automation (specialized robotic welding cells) would be far more expensive to deploy than a human worker for varied, low-volume repair jobs.
Technical feasibility todayclaude-haiku-4-5-202510012/5Robotic welding and cutting exist in production for high-volume, repetitive manufacturing, but these are narrowly scoped to fixed geometries and materials. Deployed systems struggle with variable positioning, inspection of work quality, and adaptation to non-standard parts. General maintenance workers face unpredictable conditions that exceed what current deployed products handle reliably.
Technical feasibility todayclaude-sonnet-51/5Deployed products for autonomous, general-purpose cutting/welding in variable maintenance contexts do not exist; fixed robotic welding arms in manufacturing are a different, highly structured task and not applicable to general repair work.

Maintain or repair specialized equipment or machinery located in cafeterias, laundries, hospitals, stores, offices, or factories.

18

CI 5–30 · exposure 13 · augmentation 50 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Maintenance is performed across fragmented, often small operations with older equipment and low digitization levels. Adoption of AI-assisted diagnostics is slow and concentrated in large facilities; most maintenance remains labor-intensive.
Sector adoption velocityclaude-sonnet-51/5Facilities maintenance is a low-digitization, physically dependent sector with minimal AI-driven displacement observed to date, though diagnostic tools are slowly emerging.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist with equipment diagnostics, parts identification, and maintenance history analysis, moderately improving technician efficiency. However, augmentation is limited by the physical and judgment-intensive nature of actual repair work.
Augmentation potentialclaude-sonnet-53/5AI can assist with diagnostics, manuals lookup, troubleshooting guidance, and predictive maintenance alerts, improving efficiency even though the physical repair remains human-performed.
Task automatabilityclaude-haiku-4-5-202510012/5Specialized equipment repair requires physical intervention, diagnostic judgment, and adaptive troubleshooting across diverse machinery types. While AI can assist with diagnostics and documentation, current systems cannot reliably perform hands-on repair tasks that require manipulation, calibration, and contextual assessment of unique equipment failures.
Task automatabilityclaude-sonnet-51/5This requires physical diagnosis, disassembly, part replacement, and hands-on repair of diverse equipment types; current AI has no general physical manipulation capability to perform this.
Adoption barriersclaude-haiku-4-5-202510014/5Safety regulations, equipment manufacturer liability, and on-site facility authorization requirements create substantial barriers. Legal responsibility for equipment failure often rests with licensed technicians, preventing full substitution by automated systems.
Adoption barriersclaude-sonnet-53/5No licensing typically required for general maintenance work, but physical presence, safety liability, and hands-on troubleshooting create strong practical barriers to any automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI diagnostic tools cost significant capital and integration overhead but eliminate only small portions of the repair workflow. The loaded wage of a skilled technician is often lower than the amortized cost of AI infrastructure for marginal diagnostic acceleration.
Cost vs. human wageclaude-sonnet-51/5AI cannot perform the physical repair work at all, so there is no viable substitute cost comparison; a human technician remains necessary for all physical labor.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed AI system performs end-to-end equipment repair today. Diagnostic support tools exist but are narrow in scope; actual repair work remains entirely human-dependent in production environments.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs physical equipment repair across varied settings; AI is at most used for diagnostic support or scheduling, not the hands-on repair itself.

Assemble, install, or repair wiring, electrical or electronic components, pipe systems, plumbing, machinery, or equipment.

16

CI 5–26 · exposure 13 · augmentation 50 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of robotics and automation in maintenance and repair remains limited. Most work occurs in small, dispersed service contexts (field visits, emergency repairs) where standardization is poor and human judgment is expected. Sectors are slow to digitize, and customer expectations still favor licensed human technicians.
Sector adoption velocityclaude-sonnet-51/5Building maintenance and skilled trades are physical, low-digitization sectors with minimal AI/robotics deployment for hands-on repair work.
Augmentation potentialclaude-haiku-4-5-202510013/5AI tools moderately assist technicians through diagnostic support (fault detection from sensor data), repair procedure documentation, parts identification, and predictive maintenance alerts. These features improve productivity and reduce downtime but do not transform the core task; the human remains essential for executing the physical repair.
Augmentation potentialclaude-sonnet-53/5AI can assist via diagnostic apps, augmented reality repair guides, troubleshooting chatbots, and predictive maintenance alerts, improving efficiency without performing the physical work itself.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with diagnostics and planning (e.g., identifying what needs repair from images or descriptions), the physical assembly, installation, and repair of hardware requires dexterous manipulation in unstructured environments that current robotics and automation systems struggle with at scale. The task involves real-time problem-solving, spatial reasoning, and fine motor control that remain largely beyond today's AI capabilities.
Task automatabilityclaude-sonnet-51/5This is hands-on physical work requiring manipulation of wiring, pipes, and machinery in varied real-world environments; current AI cannot perform physical manipulation tasks end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5High regulatory and safety barriers protect this task: electrical work typically requires licensing/certification, building codes govern plumbing and equipment installation, and liability exposure is substantial if automated systems fail. Insurance and compliance requirements create strong friction against full substitution.
Adoption barriersclaude-sonnet-53/5Electrical and plumbing work often requires licensed trades, code compliance, and safety inspections in many jurisdictions, creating regulatory and liability barriers, though not for every task variant.
Cost vs. human wageclaude-haiku-4-5-202510011/5The capital cost of specialized robots, combined with integration, maintenance, and the low error tolerance for safety-critical repairs, exceeds the loaded wage of skilled trades workers. For most contexts, human technicians remain far more cost-effective than current automation.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI-driven substitute performing the physical labor, so any comparison favors the human worker entirely; robotic solutions for generalized repair would be far costlier than a technician's wage.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed products reliably perform this task end-to-end in production. Specialized robotics exist for narrow, highly controlled scenarios (e.g., pipe assembly in factories), but general electrical/plumbing/machinery repair in varied field conditions lacks mature automation. AI vision systems can support diagnosis, but the hands-on execution remains human-dependent.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously assembles, installs, or repairs electrical/plumbing/mechanical systems in general facility maintenance settings; robotics for such unstructured physical tasks remain research-stage.

Adjust functional parts of devices or control instruments, using hand tools, levels, plumb bobs, or straightedges.

14

CI 10–19 · exposure 8 · augmentation 25 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Maintenance and repair is dominated by small firms and field work with low digitization. Robot adoption in this sector is minimal; most adoption remains in controlled factory settings, not customer sites or diverse field conditions.
Sector adoption velocityclaude-sonnet-51/5General maintenance and repair work is a low-digitization, physical-labor sector with minimal AI/robotic adoption in production environments.
Augmentation potentialclaude-haiku-4-5-202510012/5AI can assist with diagnostic guidance, safety checklists, and procedure documentation, but the core task—physically adjusting mechanisms with hand tools—offers limited augmentation scope since the human must do the physical manipulation regardless.
Augmentation potentialclaude-sonnet-52/5AI can assist with diagnostics, manuals, or troubleshooting guidance, but offers little support for the physical adjustment process itself.
Task automatabilityclaude-haiku-4-5-202510012/5Adjusting physical parts with hand tools requires real-world spatial reasoning, tactile feedback, and precise mechanical manipulation. Current AI systems excel at planning and instruction but cannot reliably perform physical assembly/adjustment tasks end-to-end without human intervention, and robotics for this task remain limited to narrow, controlled environments.
Task automatabilityclaude-sonnet-51/5This requires physical manipulation of hardware using hand tools and precision measuring instruments in varied real-world settings, which is beyond current AI capability without a robotic embodiment that doesn't exist at scale.
Adoption barriersclaude-haiku-4-5-202510013/5While there are no strict licensing requirements for general maintenance work, organizational and safety friction exists: responsibility for equipment failure after adjustment, customer preference for human technicians, and liability questions for AI-performed work create moderate adoption barriers.
Adoption barriersclaude-sonnet-53/5No licensing requirement typically applies, but physical presence, tool dexterity, and liability for improperly adjusted equipment create practical barriers to any automated substitution.
Cost vs. human wageclaude-haiku-4-5-202510011/5The hardware, integration, and training costs of a robot capable of general hand-tool manipulation remain far higher than the loaded wage of a general maintenance worker, especially given the low complexity of individual adjustment tasks.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI/robotic substitute performing this physical task, so the human remains the only cost-effective option.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system reliably performs general physical device adjustment with hand tools in field conditions. Specialized robotics exist in labs, but general maintenance robots that can adapt to diverse device types and conditions are not in production at scale.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs physical adjustment of mechanical/control devices using hand tools; this remains a human manual task performed on-site.

Paint or repair roofs, windows, doors, floors, woodwork, plaster, drywall, or other parts of building structures.

14

CI 5–24 · exposure 8 · augmentation 25 · importance 3.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Adoption in building maintenance remains minimal; the sector is fragmented into small and medium firms with low digital integration, high physical dependence, and conservative attitudes toward automation. No measurable displacement of maintenance workers by autonomous systems exists in production today.
Sector adoption velocityclaude-sonnet-51/5Building maintenance and construction trades are among the least digitized, lowest AI-adoption sectors, with virtually no production robotic deployment for general repair tasks.
Augmentation potentialclaude-haiku-4-5-202510012/5AI offers limited on-the-job assistance—tools like damage-detection algorithms or automated diagnostics could support planning, but most of the task requires human judgment, physical skill, and real-time problem-solving that current AI systems do not materially augment during execution.
Augmentation potentialclaude-sonnet-52/5AI can assist with diagnosing problems via image recognition, sourcing repair guides, or estimating materials, but offers minimal help with the actual physical execution of painting or repairing structures.
Task automatabilityclaude-haiku-4-5-202510012/5While some preparatory steps (inspection, estimation) could be partially automated, the core task—physically painting, repairing, or replacing building components—requires embodied manipulation in varied, unstructured environments. Current AI cannot reliably handle the dexterity, environmental adaptation, and quality judgment at 50% time savings compared to skilled humans.
Task automatabilityclaude-sonnet-51/5This is physical manual labor requiring dexterity, mobility, and adaptation to unpredictable building conditions—current AI systems cannot perform painting, roofing, or drywall repair without embodied robotics far beyond deployed capability.
Adoption barriersclaude-haiku-4-5-202510014/5Strong adoption barriers exist: building codes often require licensed contractors to perform structural repairs and sign off on work; liability and safety regulations mandate human accountability; and customer preference strongly favors human expertise for non-trivial repairs. Worker certification and warranty requirements further protect human workers.
Adoption barriersclaude-sonnet-52/5No licensing typically required for general handyman repair work, but physical access to buildings, liability for damage, and customer trust in a human presence create moderate friction against automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5The specialized hardware (mobile robots, drones, manipulators) required for autonomous building repair is expensive and often requires extensive setup and oversight per job site. The all-in cost per task (hardware amortization, integration, human supervision, quality verification) far exceeds the loaded wage of a journeyman maintenance worker.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI/robotic substitute performing this full task, so cost comparison favors human labor entirely; any robotic solution would require expensive specialized hardware exceeding human wages for general repair work.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed commercial products reliably perform full roof, window, door, or structural repair autonomously in production. Robotic painting exists in controlled factory settings, but field-deployed systems for general building repair remain research-stage and cannot handle the variability of real construction sites.
Technical feasibility todayclaude-sonnet-51/5No deployed products autonomously paint or repair building structures in general maintenance settings; any relevant robotics remain research-stage or narrow demos (e.g., painting drones in controlled settings) not used for general repair.

Install equipment to improve the energy or operational efficiency of residential or commercial buildings.

13

CI 5–21 · exposure 8 · augmentation 50 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Physical trades show slow AI adoption; most building maintenance remains localized, small-firm work with limited digitization. Pilot projects for automated installation exist but production deployment remains minimal.
Sector adoption velocityclaude-sonnet-51/5Building maintenance and repair is a physically-oriented, low-digitization sector with minimal AI/robotic adoption for hands-on installation tasks.
Augmentation potentialclaude-haiku-4-5-202510013/5AI could assist with energy efficiency audits, equipment selection recommendations, and installation planning documentation, but the core on-site installation work requires sustained human direction and problem-solving.
Augmentation potentialclaude-sonnet-53/5AI can assist with diagnostics, product selection, energy audits, and generating installation plans or checklists, but doesn't help with the physical installation itself.
Task automatabilityclaude-haiku-4-5-202510012/5Installation of physical equipment requires on-site manual dexterity, spatial reasoning, and adaptation to variable building conditions. While AI could assist with planning and documentation, the hands-on installation work cannot be meaningfully automated by current systems, making end-to-end automation infeasible.
Task automatabilityclaude-sonnet-51/5This task requires physical installation of equipment (insulation, smart thermostats, HVAC upgrades, sensors, etc.) in real buildings, involving manual dexterity, spatial reasoning, and physical manipulation that current AI systems cannot perform.'
Adoption barriersclaude-haiku-4-5-202510014/5Building code compliance, electrical/mechanical licensing requirements, liability for improper installation, customer preference for human oversight, and on-site safety regulations create substantial barriers to full automation of this task.
Adoption barriersclaude-sonnet-53/5While no formal licensing is typically required for general maintenance installation, electrical or gas-related efficiency equipment may require certified technicians, and physical access/liability concerns create some friction.
Cost vs. human wageclaude-haiku-4-5-202510011/5Robotic installation systems, where they exist, remain expensive research prototypes far exceeding the loaded wage of a maintenance worker. Integration and oversight costs would compound hardware and software expenses.
Cost vs. human wageclaude-sonnet-51/5There is no AI system capable of substituting for the physical labor involved, so AI cost is not comparable—human labor remains the only viable option.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product reliably performs physical equipment installation in real buildings. Current systems lack the embodied robotics, environmental sensing, and safety certification required for autonomous installation work at scale.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs physical equipment installation; robotics for unstructured building environments remains research-stage and far from production-ready for this kind of general maintenance work.

Repair machines, equipment, or structures, using tools such as hammers, hoists, saws, drills, wrenches, or equipment such as precision measuring instruments or electrical or electronic testing devices.

13

CI 10–15 · exposure 0 · augmentation 38 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Adoption of AI/robotic repair automation in maintenance sectors remains negligible; most repairs still rely on human technicians, with limited pilot programs and no meaningful production deployment of autonomous repair systems in general manufacturing or facility maintenance.
Sector adoption velocityclaude-sonnet-51/5Maintenance and repair is a physical, low-digitization trade with minimal AI/robotic deployment in production; adoption of automation in this field is very slow.
Augmentation potentialclaude-haiku-4-5-202510012/5AI can assist with diagnostic recommendations (fault detection, parts identification) via image analysis or sensor data, and some planning tools help technicians, but these are incremental aids; the hands-on repair work itself remains human-driven with limited augmentation from current systems.
Augmentation potentialclaude-sonnet-53/5AI can assist with diagnostics (e.g., interpreting sensor data, suggesting repair procedures, or providing troubleshooting guides) but the physical execution of repairs remains entirely human-driven.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires physical manipulation of tools and equipment in real environments with spatial reasoning, dexterity, and tactile feedback. Current AI systems cannot perform end-to-end repair work autonomously; mobile manipulation remains in early stages and cannot reliably diagnose and fix diverse equipment failures at scale.
Task automatabilityclaude-sonnet-51/5This task requires physical manipulation of tools, diagnosis of mechanical/electrical faults, and hands-on repair work that current AI systems cannot perform end-to-end; robotics for general-purpose repair remains research-stage.
Adoption barriersclaude-haiku-4-5-202510012/5While there is no strict licensing requirement for general maintenance repair, liability for faulty repairs (safety hazards, equipment damage) and the physical presence requirement create moderate adoption barriers; organizations often prefer human accountability for critical infrastructure.
Adoption barriersclaude-sonnet-53/5While no formal licensing is typically required for general maintenance work, safety liability, physical access to equipment, and the need for hands-on dexterity create substantial practical barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5Current robotic systems capable of any repair work are capital-intensive ($100k+) with high integration costs, maintenance, and oversight requirements, far exceeding the loaded wage of a general maintenance worker ($40–60k annually).
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing the physical repair work itself, so cost comparison favors the human by default; any AI-assisted diagnostics only marginally reduce total cost.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed commercial product reliably performs general machine repair end-to-end today. Specialized robotic systems exist for narrow, controlled tasks, but general repair across diverse equipment types, problem diagnosis, and tool selection remains infeasible in production environments.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously repairs diverse machines/equipment using hand tools and precision instruments; robotic manipulation for unstructured repair tasks is not commercially mature.

Perform routine maintenance on boilers, such as replacing burners or hoses, installing replacement parts, or reinforcing structural weaknesses to ensure optimal boiler efficiency.

13

CI 5–20 · exposure 8 · augmentation 50 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of AI-assisted diagnostics in industrial maintenance is growing but slow; the physical, capital-intensive nature of boiler work and reliance on licensed technicians limits velocity. Most organizations still rely on traditional preventive and reactive maintenance models.
Sector adoption velocityclaude-sonnet-51/5Building maintenance and repair trades are a low-digitization, physically dominated sector with minimal AI/robotics adoption in the field today.
Augmentation potentialclaude-haiku-4-5-202510013/5Predictive maintenance dashboards and diagnostic AI can guide technicians toward problem areas and optimal repair strategies, modestly improving planning efficiency and reducing downtime. However, the augmentation is limited to the decision-support phase rather than execution.
Augmentation potentialclaude-sonnet-53/5AI can assist via diagnostic apps, manuals, parts lookup, or troubleshooting guidance, but does not materially change the physical execution of the repair.
Task automatabilityclaude-haiku-4-5-202510012/5While AI-driven diagnostics could identify maintenance needs, the hands-on work of replacing burners, installing parts, and reinforcing structures requires dexterity, situational judgment, and physical presence that current robots and AI lack reliably. Only diagnostic and planning phases are automatable; execution remains dependent on human workers.
Task automatabilityclaude-sonnet-51/5This is a physical hands-on task requiring manipulation of hardware, disassembly, and fitting of parts in confined mechanical spaces—current AI systems cannot perform physical labor of this kind.
Adoption barriersclaude-haiku-4-5-202510014/5Boiler maintenance typically falls under licensed trades (jurisdictional licensing), carries significant liability (safety, regulatory compliance, insurance), and often requires certified technicians to perform and sign off on work. Legal and safety requirements create hard adoption barriers.
Adoption barriersclaude-sonnet-54/5Boiler work often requires certified technicians due to safety codes (pressure vessels, gas lines), and liability for improper repair is high, creating strong regulatory and safety barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Diagnostic AI is cheap, but the heavy capital and infrastructure costs of deploying robotic systems for boiler maintenance, combined with human oversight and safety requirements, make the all-in cost comparable to or higher than human technicians for most operations.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing the physical repair, so the human remains the only cost-effective option for this labor.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system reliably performs physical maintenance tasks on boilers end-to-end. Diagnostic tools and predictive maintenance exist, but replacement and installation work remains entirely human-dependent in production environments.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs physical boiler maintenance; robotics for this specific unstructured mechanical work is not commercially available.

Position, attach, or blow insulating materials to prevent energy losses from buildings, pipes, or other structures or objects.

13

CI 10–15 · exposure 0 · augmentation 25 · importance 3.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Insulation installation is primarily performed by small contractors and maintenance workers in fragmented, low-digitization sectors. Adoption of automation is minimal; the industry remains labor-intensive and location-dependent.
Sector adoption velocityclaude-sonnet-51/5Building maintenance and construction trades are among the least digitized sectors with minimal AI/robotics adoption for physical installation tasks.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could assist with task planning (e.g., estimating material needs, identifying problem areas via thermal imaging), but the core physical work of positioning and securing materials requires human judgment and dexterity that AI tools have not meaningfully augmented to date.
Augmentation potentialclaude-sonnet-52/5AI could help with planning material quantities, identifying insulation gaps via thermal imaging analysis, or generating instructions, but offers little direct assistance during the physical placement task itself.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires physical manipulation in varied, site-specific environments—positioning insulating materials, securing them, and applying blown-in insulation to irregular surfaces. Current AI and robotics cannot reliably perform the end-to-end physical execution at scale across diverse building geometries and structures.
Task automatabilityclaude-sonnet-51/5This is a physical installation task requiring manipulation of materials, ladders/scaffolding access, and fitting insulation into irregular spaces; no current AI system can perform the physical placement or blowing of insulation.
Adoption barriersclaude-haiku-4-5-202510012/5While there are few licensing barriers for the task itself, occupational safety regulations (OSHA compliance, handling hazardous materials), building code verification, and customer preference for skilled, accountable human workers create modest friction to full automation.
Adoption barriersclaude-sonnet-53/5No licensing typically required for basic insulation work, but physical access to structures, safety requirements, and building codes create moderate friction against any automated approach.
Cost vs. human wageclaude-haiku-4-5-202510011/5Current robotic systems capable of any insulation work (e.g., spray-foam rigs) are expensive, require skilled operators, and demand significant setup and integration per site. Human labor remains cheaper for most residential and commercial applications.
Cost vs. human wageclaude-sonnet-51/5No viable AI/robotic substitute exists, so any AI-based approach would require expensive custom robotics far exceeding the cost of a human worker with basic tools.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed commercial product reliably performs general insulation installation autonomously. Prototype robotic systems exist in narrow, controlled settings (flat, uniform surfaces), but production-grade systems that handle arbitrary building configurations do not exist.
Technical feasibility todayclaude-sonnet-51/5There are no deployed robotic or AI products performing insulation installation in buildings at any meaningful scale; this remains manual skilled trade work.

Dismantle machines, equipment, or devices to access and remove defective parts, using hoists, cranes, hand tools, or power tools.

12

CI 5–19 · exposure 8 · augmentation 25 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Adoption of automation in general maintenance work remains minimal; the sector is still predominantly manual and relies on skilled technician judgment. Only large industrial facilities have invested in specialized robotic systems for specific, high-volume disassembly tasks, and these remain exceptions rather than the norm.
Sector adoption velocityclaude-sonnet-51/5Maintenance and repair trades show minimal AI/robotic adoption for physical disassembly tasks, remaining a low-digitization, hands-on sector.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could marginally assist by providing disassembly documentation, identifying defective parts from images, or guiding technicians through procedures, but current systems offer limited augmentation since technicians rely heavily on experience, physical inspection, and real-time problem-solving that AI cannot yet enhance meaningfully.
Augmentation potentialclaude-sonnet-52/5AI can assist with diagnostics, repair manuals, or identifying defective parts beforehand, but offers little direct help during the physical dismantling process itself.
Task automatabilityclaude-haiku-4-5-202510012/5While disassembly sequences could be partially documented and some mechanical guidance generated by AI, the task requires real-world spatial reasoning, physical manipulation with hand and power tools, and safe handling of heavy equipment with hoists/cranes. Current AI cannot reliably perform the physical work itself or navigate the unpredictable variations in equipment design and defect locations that require adaptive tactile feedback.
Task automatabilityclaude-sonnet-51/5This is a hands-on physical disassembly task requiring manipulation of hoists, cranes, and tools in unstructured environments; current AI systems cannot perform physical manipulation tasks like this at all.
Adoption barriersclaude-haiku-4-5-202510014/5Significant barriers exist: safety liability for automated machinery operation, OSHA regulations governing equipment dismantling and tool use, insurance requirements, and the need for human judgment in identifying defects and safe disassembly sequences. Organizations typically require licensed technicians to perform or directly oversee this work.
Adoption barriersclaude-sonnet-53/5No licensing barrier specifically prevents automation of disassembly, but safety requirements around heavy equipment, liability for improper dismantling, and physical workspace constraints create real friction.
Cost vs. human wageclaude-haiku-4-5-202510011/5Physical robots capable of disassembly are extremely expensive to purchase, program, and maintain, far exceeding the loaded wage of a maintenance technician. Integration costs and limited reusability across different equipment designs make AI-based automation economically unviable today.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this physical labor, so any hypothetical robotic system would be far more costly than a human technician given current hardware costs and limitations.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed products today can autonomously dismantle machines and remove defective parts end-to-end. Robotic systems for disassembly exist in highly controlled research environments but lack the dexterity, adaptability, and safety compliance needed for general maintenance work across varied equipment types.
Technical feasibility todayclaude-sonnet-51/5No deployed product performs physical dismantling of machinery; robotics for such unstructured, variable mechanical disassembly remains research-stage at best.

Fabricate or repair counters, benches, partitions, or other wooden structures, such as sheds or outbuildings.

10

CI 5–15 · exposure 0 · augmentation 25 · importance 3.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Maintenance and repair workers in small trades operate in fragmented, low-digitization sectors with physical on-site constraints. Adoption of automation in this skilled trades area remains minimal, with limited capital investment in robotics at small scales.
Sector adoption velocityclaude-sonnet-51/5Skilled trades and physical maintenance work show minimal AI/robotic adoption due to the unstructured, variable nature of on-site fabrication tasks.
Augmentation potentialclaude-haiku-4-5-202510012/5AI can assist with design suggestions, material estimation, or reference planning via CAD or vision analysis, but such tools offer limited productivity uplift for a task that is fundamentally craft-driven and dependent on physical execution and on-site judgment.
Augmentation potentialclaude-sonnet-52/5AI can assist with design plans, measurements, material estimates, or generating cut lists, but offers little help with the actual physical fabrication or repair process.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires physical manipulation of materials (wood cutting, joining, fastening) and site-specific measurement and customization that current AI systems cannot perform end-to-end. Robotic woodworking exists in controlled factory settings but cannot handle the variable, on-site repair and fabrication work described.
Task automatabilityclaude-sonnet-51/5This requires physical fabrication, cutting, joining, and installation of wooden structures—manual dexterity and physical world manipulation that current AI systems cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Strong adoption barriers exist: structural work often requires licensed tradespeople in jurisdictions with building codes, customer expectations for human craftsmanship, liability for structural integrity, and the physical, site-dependent nature of the work creates high organizational friction.
Adoption barriersclaude-sonnet-52/5No licensing generally required for general carpentry repair work, though building codes may apply to structures like outbuildings; the main barrier is physical infeasibility rather than regulation.
Cost vs. human wageclaude-haiku-4-5-202510011/5The capital cost of specialized woodworking robotics, combined with integration and setup for each job site, far exceeds the loaded wage of a skilled maintenance worker for the variable, small-to-medium batch work typical in this domain.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI-driven automation solution for this physical task, so any hypothetical robotic system would be vastly more expensive than a human carpenter/repair worker.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product performs full fabrication or repair of wooden structures autonomously. While AI can assist in design or planning, the actual physical construction requires skilled human hands and real-time problem-solving in unstructured environments.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously fabricates or repairs wooden structures like sheds or benches; this remains far outside current robotics/AI product capability for unstructured carpentry work.

Assemble boilers at installation sites, using tools such as levels, plumb bobs, hammers, torches, or other hand tools.

5

CI 5–5 · exposure 0 · augmentation 25 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5HVAC and building maintenance sectors show laggard AI adoption patterns; on-site assembly work is physical, site-specific, and remains heavily dependent on skilled human labor with minimal deployed automation.
Sector adoption velocityclaude-sonnet-51/5Skilled trades and field maintenance work are among the least digitized, slowest-adopting sectors for AI and robotics due to physical variability and lack of standardization.
Augmentation potentialclaude-haiku-4-5-202510012/5While AI could assist with documentation, code lookup, or sequencing guidance, it offers limited assistance with the core physical assembly task itself, which remains primarily manual.
Augmentation potentialclaude-sonnet-52/5AI can assist with instructions, diagrams, or troubleshooting guidance via mobile devices, but offers minimal help with the core physical assembly task itself.
Task automatabilityclaude-haiku-4-5-202510011/5On-site boiler assembly requires precise physical manipulation in varied real-world environments, coordination of multiple hand tools, and real-time spatial judgment. Current AI cannot perform this end-to-end mechanical assembly task with adequate speed and reliability.
Task automatabilityclaude-sonnet-51/5This requires physical manipulation of heavy components, hand-eye coordination, and adaptive fitting on-site, none of which current AI systems can perform end-to-end; robotics for this remains research-stage.'
Adoption barriersclaude-haiku-4-5-202510014/5Boiler installation involves building codes, safety certifications, and often requires a licensed/credentialed technician to sign off on the work, creating legal and regulatory barriers to full automation.
Adoption barriersclaude-sonnet-54/5Boiler installation is often subject to codes, permits, and licensed trade requirements (e.g., pipefitter/plumber certification), creating strong regulatory and liability barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5Specialized robotic systems capable of assembly work would require substantial capital investment, integration, and maintenance—far exceeding the loaded wage of a skilled maintenance worker for this task.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI/robotic substitute performing this task, so the human remains the only cost-effective option; any attempted automation would require expensive custom robotics far exceeding a technician's wage.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed products reliably perform on-site boiler assembly autonomously. Robotics systems capable of this work exist only in research or highly controlled lab settings, not in production at installation sites.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously assembles boilers at installation sites; industrial robots exist for factory work, not mobile field assembly with hand tools.

Train or manage maintenance personnel or subcontractors.

4

CI 0–7 · exposure 0 · augmentation 38 · importance 3.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Maintenance sectors are traditionally less digitized, and the people-management aspect has no algorithmic substitute being deployed in production today. Adoption of AI management tools remains minimal and experimental.
Sector adoption velocityclaude-sonnet-52/5Building trades and facilities maintenance sectors show slow, uneven AI adoption, especially for supervisory/management functions rather than technical diagnostics.
Augmentation potentialclaude-haiku-4-5-202510012/5AI can marginally assist with scheduling suggestions or training material drafting, but provides limited productivity transformation given that relationship-building and judgment remain central to the task.
Augmentation potentialclaude-sonnet-53/5AI can assist with generating training materials, tracking certifications, scheduling, and documenting procedures, improving efficiency of the human manager's supporting tasks.
Task automatabilityclaude-haiku-4-5-202510011/5Training and managing personnel requires human judgment, relationship-building, conflict resolution, and contextual decision-making that current AI systems cannot perform end-to-end. Supervisory duties inherently depend on understanding individual worker capabilities and interpersonal dynamics.
Task automatabilityclaude-sonnet-51/5Training and managing people requires interpersonal leadership, coaching, performance evaluation, and hands-on skill demonstration that current AI cannot execute end-to-end.'
Adoption barriersclaude-haiku-4-5-202510015/5Labor law, collective bargaining agreements, and organizational norms require a human manager with accountability for hiring, discipline, safety oversight, and performance evaluation. Legal liability for worker safety and employment decisions creates a hard barrier to automation.
Adoption barriersclaude-sonnet-54/5Managing personnel involves employment law, safety accountability, and organizational authority that generally requires a human in a supervisory role, creating strong structural barriers.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI tools lack the capability to replace a maintenance supervisor's core functions, making direct cost comparison moot. Any AI assistance would require significant human oversight, negating cost savings.
Cost vs. human wageclaude-sonnet-51/5Since AI cannot perform the core management/supervisory function, there is no viable substitute cost comparison; human managers remain necessary.
Technical feasibility todayclaude-haiku-4-5-202510011/5No production AI system today reliably manages people, assigns work, evaluates performance, or provides effective training at scale. While chatbots can assist with documentation or scheduling, they cannot substitute for the human oversight required to manage and develop maintenance teams.
Technical feasibility todayclaude-sonnet-51/5No deployed product manages or trains maintenance staff autonomously in production; at best AI provides supplementary training content or scheduling support.

Related occupations — Installation, Maintenance & Repair

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.