Molding, Coremaking, and Casting Machine Setters, Operators, and Tenders, Metal and Plastic

51-4072.00

Median wage $44,350/yr150,470 employed (US)Rank #448 of 923 scored · top 49% by substitution

Set up, operate, or tend metal or plastic molding, casting, or coremaking machines to mold or cast metal or thermoplastic parts or products.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution28

Exposure22

Augmentation35

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

29 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

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%23

panel mean rating 1.9/5 → substitution pressure 23/100

Technical feasibility todayw 20%20

panel mean rating 1.8/5 → substitution pressure 20/100

Cost vs. human wagew 15%20

panel mean rating 1.8/5 → substitution pressure 20/100

Adoption barriersw 20%inverted — strong barriers lower the score58

panel mean rating 2.7/5 (barrier strength) → substitution pressure 58/100

Sector adoption velocityw 10%20

panel mean rating 1.8/5 → substitution pressure 20/100

Task breakdown (29 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.

Select and install blades, tools, or other attachments for each operation.

CI 35–84 · exposure 58 · augmentation 38 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	4/5	Metalworking and plastic casting are capital-intensive, digitized industries with strong incentives and demonstrated willingness to deploy robotic automation for repetitive machine setup and tool management.
Sector adoption velocity	claude-sonnet-5	2/5	Manufacturing floor tasks involving physical tooling changes are in a sector with slower, capital-intensive automation adoption compared to information-based work.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI vision systems and tool-management software can assist operators by suggesting optimal blade/tool selection and flagging mismatches, though the physical installation itself is either fully automated or purely manual.
Augmentation potential	claude-sonnet-5	2/5	AI can assist with predictive maintenance schedules or tool selection recommendations via data analytics, but does not materially transform this physical task today.
Task automatability	claude-haiku-4-5-20251001	5/5	AI-powered robotic systems can reliably select and install appropriate blades and tools by analyzing job specifications, consulting tool databases, and executing precise mechanical installation with vision guidance. This represents a well-defined, repeatable operation where end-to-end automation can exceed 50% time savings at equal quality.
Task automatability	claude-sonnet-5	2/5	This requires physical manipulation, judgment about tool selection, and fine motor skills that current general-purpose AI cannot perform end-to-end without specialized robotics integration.rating reflects that only narrow, highly engineered robotic cells could partially do this today.
Adoption barriers	claude-haiku-4-5-20251001	2/5	Few regulatory or licensing barriers prevent automation of tool installation; organizational adoption is primarily cost-benefit driven, though equipment-specific integration friction and maintenance overhead present modest adoption friction.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement, but physical safety protocols, machine-specific certification, and liability for improper tooling create moderate organizational friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	4/5	Robotic tool installation systems have high upfront capital cost but very low per-operation marginal cost; amortized across high-volume production runs, they are significantly cheaper than human labor, though not quite an order of magnitude for smaller operations.
Cost vs. human wage	claude-sonnet-5	2/5	Robotic tool-changing systems require significant capital investment, engineering, and maintenance, often exceeding the cost of a trained operator for lower-volume or varied setups.
Technical feasibility today	claude-haiku-4-5-20251001	4/5	Commercial robotic systems with tool changers and vision-guided installation are deployed in production metalworking and casting environments, though integration remains semi-automated rather than fully autonomous in most facilities. Mature solutions exist but often require human handoff or setup validation.
Technical feasibility today	claude-sonnet-5	2/5	Some automated tool-changers exist in CNC and injection molding contexts, but generalized AI-driven blade/tool selection and installation across varied machine setups is not a mature, widely deployed product.

Inventory and record quantities of materials and finished products, requisitioning additional supplies as necessary.

CI 46–72 · exposure 55 · augmentation 75 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	3/5	Automotive, aerospace, and larger contract manufacturers have adopted automated inventory systems; smaller foundries and job shops lag significantly. Overall adoption remains patchy—pilots and partial implementations common, but full end-to-end automation less universal in metalworking.
Sector adoption velocity	claude-sonnet-5	3/5	Manufacturing is a moderate-adoption sector for digitized inventory systems—common in larger plants but slower and more manual in smaller shops, reflecting a mid-tier digitization pattern.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI-driven inventory systems substantially assist operators by auto-flagging low stock, suggesting reorder quantities, and alerting to discrepancies, which significantly raises operator productivity in monitoring and requisitioning. The human remains in the loop for final approval and exception handling.
Augmentation potential	claude-sonnet-5	4/5	AI-enabled inventory systems substantially reduce manual counting and paperwork burden, letting operators focus on quality/production while automated systems flag low stock and suggest reorders.
Task automatability	claude-haiku-4-5-20251001	2/5	Current AI can handle data entry and record-keeping for inventory with high accuracy, but the full task requires physical counting or sensor integration, judgment about reorder thresholds, and supplier coordination that remains partially manual. Typical AI can automate perhaps 30–40% of the workflow without substantial custom setup.
Task automatability	claude-sonnet-5	4/5	Inventory counting and recording, and generating requisition requests based on stock levels, are well-suited to barcode/RFID scanning, ERP/inventory management systems, and automated reorder triggers that already exist and save significant time over manual tallying.
Adoption barriers	claude-haiku-4-5-20251001	2/5	Few legal barriers exist; inventory management is purely operational and non-licensed. Modest friction includes integration with legacy shop systems, supplier API compatibility, and worker familiarity, but no regulatory or authorization requirement protects the task.
Adoption barriers	claude-sonnet-5	2/5	No licensing or legal requirement mandates a human perform inventory counting or requisitioning, though some organizational oversight may exist to verify accuracy or authorize purchases.
Cost vs. human wage	claude-haiku-4-5-20251001	3/5	Barcode scanners, inventory software, and automated requisitioning systems cost hundreds to thousands annually, comparable to 1–2 FTE inventory technicians in a small shop. Larger facilities can achieve better economies of scale, placing overall cost parity around equivalence.
Cost vs. human wage	claude-sonnet-5	4/5	Automated inventory tracking and requisition software is inexpensive relative to manual labor hours spent counting and recording, especially at scale, though initial system setup and integration carry some cost.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	Inventory management software with barcode/RFID scanning and automated reorder systems exist in production at scale, but many metal/plastic foundries and small shops still rely on manual or semi-automated methods. Deployed solutions work reliably for well-structured inventory but often require human oversight and correction.
Technical feasibility today	claude-sonnet-5	4/5	Warehouse and manufacturing inventory management systems (e.g., ERP modules, barcode scanners, automated reorder point software) are mature and widely deployed in production settings today, though physical counting/reconciliation still sometimes requires human verification.

Maintain inventories of materials.

CI 32–77 · exposure 50 · augmentation 75 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	3/5	Manufacturing facilities have adopted inventory software and some RFID/barcode automation, but adoption of fully autonomous AI-driven inventory systems remains uneven. Larger facilities lead adoption; smaller job shops and contract manufacturers lag significantly.
Sector adoption velocity	claude-sonnet-5	3/5	Manufacturing has moderate digitization; many plants use ERP/inventory software but small-to-mid shops still rely on manual logs, so adoption is uneven.
Augmentation potential	claude-haiku-4-5-20251001	4/5	Inventory management systems and real-time dashboards substantially assist human operators by flagging low stock, suggesting reorder quantities, and reducing manual search time. AI augmentation here demonstrably raises human productivity without full replacement.
Augmentation potential	claude-sonnet-5	4/5	AI-driven inventory dashboards, demand forecasting, and automated alerts significantly boost an operator's ability to manage materials efficiently while still allowing human oversight.
Task automatability	claude-haiku-4-5-20251001	2/5	Material inventory maintenance involves physical tracking, stock counts, and location management in manufacturing environments. While inventory record-keeping can be partially automated with barcode scanning or RFID, the core physical tasks—locating, counting, and organizing materials in a shop—require on-site presence and manual verification that current AI cannot perform end-to-end.
Task automatability	claude-sonnet-5	4/5	Inventory maintenance is a data-tracking task well suited to existing inventory management software and IoT sensors, requiring minimal human judgment for routine tracking and reordering.
Adoption barriers	claude-haiku-4-5-20251001	3/5	There are no strict licensing barriers to automating inventory tracking, but organizational inertia, the need for periodic human audits, and shop-floor reality checks create moderate friction. Many plants require manual cycle counts for accountability and regulatory compliance.
Adoption barriers	claude-sonnet-5	1/5	No licensing or regulatory requirement mandates human inventory tracking; it's a purely operational task with low liability concerns.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Inventory management software and integration carry meaningful setup and maintenance costs, and AI cannot eliminate the human worker needed for physical stock verification and bin management on the shop floor. Cost parity or advantage depends heavily on implementation scale and labor rates.
Cost vs. human wage	claude-sonnet-5	4/5	Automated inventory systems are far cheaper per transaction than manual counting and logging by an operator, though initial integration costs exist.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Inventory management software exists and is widely deployed, but it typically requires human input (scanning, physical counts, manual updates) and struggles with real-time accuracy in dynamic manufacturing settings. AI-driven inventory optimization is nascent in production; most systems remain semi-manual with significant human oversight.
Technical feasibility today	claude-sonnet-5	4/5	Mature ERP and warehouse management systems (SAP, Oracle, barcode/RFID systems) already track material inventories reliably in manufacturing settings today.

Observe meters and gauges to verify and record temperatures, pressures, and press-cycle times.

CI 35–61 · exposure 47 · augmentation 63 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Metal and plastic casting remains a physically distributed, capital-constrained industry with many small and medium-sized shops. Adoption of continuous automated monitoring is limited; manual observation and recording remain the norm even in larger operations, reflecting slow digitization of shop floors.
Sector adoption velocity	claude-sonnet-5	2/5	Discrete metal/plastic manufacturing has moderate but uneven digitization; many smaller shops still rely on manual monitoring, and Industry 4.0 sensor adoption is a slow, capital-driven process.
Augmentation potential	claude-haiku-4-5-20251001	3/5	Real-time dashboards, automated data logging, and threshold alerts meaningfully assist operators by reducing manual note-taking and highlighting anomalies. However, the core task of observing meters is straightforward enough that AI assistance, while useful, does not substantially transform operator productivity.
Augmentation potential	claude-sonnet-5	4/5	Automated dashboards and alerts significantly help operators track multiple gauges simultaneously and catch anomalies faster, improving productivity while humans still oversee the process.
Task automatability	claude-haiku-4-5-20251001	2/5	While meter and gauge observation can be partially automated through sensor data logging and basic anomaly detection, the task requires judgment about acceptable variation, contextual interpretation of readings, and integration with broader equipment operation. Current AI systems can flag deviations but cannot fully replace the operator's situational awareness and decision-making.
Task automatability	claude-sonnet-5	3/5	Reading and logging meter/gauge values is easily automated via sensors, PLCs, and data historians, but 'observe' implies visual checks on physical equipment that may not all be digitally instrumented, requiring capital investment for full automation.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Operational safety and quality control standards often legally require human verification of critical parameters (temperature, pressure) before press activation. Equipment qualification and operator certification create regulatory friction, though these are not absolute prohibitions on sensor-based automation.
Adoption barriers	claude-sonnet-5	2/5	No licensing or legal requirement mandates human observation of these readings; the main friction is capital cost of retrofitting older machines and integration with existing workflows.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Automated sensor systems and data logging infrastructure require substantial capital investment and integration costs. For a single operator performing routine meter checks, the amortized cost of automation infrastructure often exceeds the labor cost saved, particularly in smaller or older facilities.
Cost vs. human wage	claude-sonnet-5	4/5	Once sensors and data acquisition systems are installed, continuous automated monitoring is far cheaper per data point than manual observation, though initial retrofit costs exist for older equipment.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	Industrial IoT systems and SCADA platforms already log sensor data automatically in many foundries, and monitoring software can alert operators to out-of-spec conditions. However, human verification and recording remain standard practice, and no production system fully eliminates the operator's role in this observation task.
Technical feasibility today	claude-sonnet-5	4/5	Industrial automation systems (SCADA, IoT sensors, data loggers) reliably capture and record process parameters like temperature, pressure, and cycle time in modern manufacturing plants today.

Observe continuous operation of automatic machines to ensure that products meet specifications and to detect jams or malfunctions, making adjustments as necessary.

CI 25–49 · exposure 38 · augmentation 63 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Molding and casting shops, especially smaller and mid-sized job shops, adopt automation slowly. The sector remains fragmented, with many legacy machines and manual workflows; while large automotive suppliers are deploying vision systems, the broader industry lags significantly behind tech and finance sectors.
Sector adoption velocity	claude-sonnet-5	2/5	Manufacturing sectors adopt automation steadily but slowly compared to information/professional services, with many small-to-mid manufacturers still relying on manual monitoring.
Augmentation potential	claude-haiku-4-5-20251001	4/5	Real-time visual dashboards, anomaly alerts, and predictive maintenance notifications can substantially boost operator productivity and decision-making. Operators using AI-assisted monitoring can respond faster to emerging defects and prevent downstream scrap, keeping them actively engaged while amplifying their impact.
Augmentation potential	claude-sonnet-5	3/5	AI-enabled sensors and predictive maintenance tools can alert operators to anomalies and jams earlier, improving efficiency while humans remain responsible for physical intervention.
Task automatability	claude-haiku-4-5-20251001	3/5	Computer vision systems can monitor machine output and detect visible defects or jams in real time, and condition-monitoring AI can flag anomalies in sensor data. However, the task requires physical adjustments and nuanced judgment about when deviations are tolerable versus requiring intervention, which still needs human oversight. Roughly half the observational and diagnostic work could be automated.
Task automatability	claude-sonnet-5	2/5	Vision-based monitoring systems can detect some jams/defects but the physical adjustment and diverse machine interactions still require human presence and dexterity, limiting full end-to-end automation today.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Product safety and quality liability create strong practical barriers: defects in molded parts can affect downstream manufacturing and end-product safety, making producers hesitant to remove human sign-off. Regulatory and insurance frameworks often require documented human inspection and sign-off on critical dimensions.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement, but safety protocols and liability for equipment damage or defective product batches create some organizational caution around fully unattended operation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	While computer vision hardware and monitoring software have dropped in cost, the integration, calibration, and ongoing oversight overhead are substantial. For small to mid-scale operations, the all-in cost of reliable AI inspection and anomaly detection still approaches or exceeds the loaded wage of an operator.
Cost vs. human wage	claude-sonnet-5	2/5	Sensor/vision monitoring systems require significant capital investment, integration, and maintenance, often comparable to or exceeding the cost of an operator for smaller production runs.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	Vision-based quality inspection systems and predictive maintenance platforms exist in production, but they typically operate alongside human operators rather than replacing them. Error rates and false positives remain material, and integration with legacy molding equipment varies widely, limiting full end-to-end automation.
Technical feasibility today	claude-sonnet-5	2/5	Some factories deploy machine vision and sensor-based anomaly detection, but reliable autonomous adjustment across varied machine types and materials is not yet standard production practice.

Cool products after processing to prevent distortion.

CI 34–39 · exposure 25 · augmentation 50 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Metalworking and casting are capital-intensive but traditional sectors with slower digital adoption; while large foundries may use advanced cooling systems, SMEs predominantly rely on manual or semi-automated setups, indicating slow overall adoption of autonomous cooling and distortion prevention.
Sector adoption velocity	claude-sonnet-5	2/5	Manufacturing sectors using casting/molding equipment have historically slower adoption of AI-driven process optimization compared to information-sector industries, relying more on traditional automation and control engineering.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI-powered sensors and real-time temperature monitoring can assist operators by alerting them to cooling anomalies and predicting distortion risk, improving operator productivity and quality without removing human oversight.
Augmentation potential	claude-sonnet-5	3/5	AI-based predictive analytics and sensor monitoring can help operators optimize cooling parameters and detect anomalies, offering moderate productivity assistance without replacing the physical task.
Task automatability	claude-haiku-4-5-20251001	2/5	Cooling requires physical environmental control (temperature, humidity, airflow) and monitoring for distortion, which are partially automatable via thermal systems and sensors, but detecting and preventing distortion in real-time requires judgment and manual intervention that current AI cannot reliably execute end-to-end without supervision.
Task automatability	claude-sonnet-5	2/5	This is a physical process control task involving timed cooling of molded parts, which requires monitoring equipment and physical handling rather than something purely informational AI can execute end-to-end.dinamik.It could be automated with dedicated industrial control systems, but this is not general-purpose AI, so under 50% time savings via off-the-shelf AI is unlikely.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Safety regulations around hot-material handling and quality control requirements create some friction, and many foundries and casting shops rely on operator judgment for distortion assessment, introducing organizational resistance to full automation.
Adoption barriers	claude-sonnet-5	2/5	No licensing or major regulatory barrier exists, though safety and quality control processes in manufacturing require validated equipment settings and some human oversight for defect prevention.
Cost vs. human wage	claude-haiku-4-5-20251001	3/5	Thermal control hardware and sensor systems are moderately expensive, while cooling operators earn modest wages; automation cost is comparable to or slightly higher than hiring operators for many smaller operations, making the ratio roughly even.
Cost vs. human wage	claude-sonnet-5	3/5	Automated cooling control (via PLCs/sensors) is already cost-effective in industrial settings, but this is machine automation rather than AI-driven cost savings per se, making the comparison to AI-labor cost ratios less applicable.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Automated cooling systems exist for metalworking (ovens, quench tanks, controlled environments), but comprehensive distortion prevention integrating visual inspection and process adjustment is not reliably deployed as a full autonomous system; most operations combine mechanical cooling with human oversight.
Technical feasibility today	claude-sonnet-5	2/5	Cooling cycles are typically managed by PLCs and industrial process controllers embedded in molding machines, not general AI products, and these are engineering control systems rather than AI systems performing the task autonomously.

Unload finished products from conveyor belts, pack them in containers, and place containers in warehouses.

CI 35–38 · exposure 25 · augmentation 25 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	3/5	Manufacturing and logistics are adopting robotic palletizers and conveyor systems at a moderate pace, especially in larger operations and automotive supply chains. However, adoption remains uneven; small and mid-tier metal/plastic casting shops lag significantly, and full end-to-end automation is still relatively uncommon.
Sector adoption velocity	claude-sonnet-5	2/5	Manufacturing floor tasks involving physical material handling see slower and more capital-intensive automation adoption compared to office/information work, though palletizing robots are gradually spreading in some large-scale plants.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI and robotics can assist with sorting and packing guidance (computer vision for defect detection, packing optimization), but the task is primarily physical handling where a human worker remains essential for real-time adaptation. The augmentation benefit is modest compared to the core labor requirement.
Augmentation potential	claude-sonnet-5	2/5	Some conveyor and warehouse automation assists workers by reducing manual handling, but this is mechanical/robotic automation rather than AI-driven augmentation of the human worker's cognitive task.
Task automatability	claude-haiku-4-5-20251001	2/5	While robotic arms and conveyor automation exist, the full pipeline—unloading varied product geometries, packing into containers with spatial reasoning, and warehouse placement—requires significant custom integration and still struggles with fragile/irregular parts. Current off-the-shelf systems cannot match the 50% time-saving threshold for the complete task end-to-end.
Task automatability	claude-sonnet-5	2/5	This requires physical manipulation of finished products, packing, and warehouse placement, which current general-purpose AI cannot do; robotic automation exists but is not 'AI' in the software sense and requires heavy capital investment specific to each line.'
Adoption barriers	claude-haiku-4-5-20251001	2/5	Few hard regulatory barriers exist for warehouse automation itself, but organizational friction is real: existing facility layouts, safety interlocks with human workers, and logistics integration create deployment friction. Product variability and occasional need for human judgment (damage detection, sorting) add friction but are not absolute blockers.
Adoption barriers	claude-sonnet-5	2/5	No licensing or legal requirement mandates a human for this task, though safety regulations around machinery and physical workspace shared with robots create some friction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Integrated robotic systems (hardware + integration + maintenance) typically cost $200k–$500k+ plus significant setup labor, while the task itself (unload, pack, place) may be performed by workers at $25–$35/hour loaded cost. The upfront and ongoing costs remain comparable to or exceed human labor for typical production volumes.
Cost vs. human wage	claude-sonnet-5	2/5	Industrial robotic systems for this work require significant capital investment, integration, and maintenance, often costing more than human labor unless run at very high volume with long amortization.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Partial automation is deployed in some facilities (robotic palletizers, conveyor systems), but reliable end-to-end performance across diverse product types, packaging constraints, and warehouse logistics remains rare in production. Most implementations require heavy engineering customization and still have material failure rates on non-standard items.
Technical feasibility today	claude-sonnet-5	2/5	Robotic pick-and-place and palletizing systems exist in some factories, but reliable unloading from conveyors, packing varied items, and warehouse placement across diverse product types remains narrow and product-specific, not broadly deployed for this exact combined task.

Trim excess material from parts, using knives, and grind scrap plastic into powder for reuse.

CI 31–35 · exposure 25 · augmentation 25 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Metal and plastic casting/molding shops are predominantly small to mid-sized, lower-digitization manufacturers. Adoption of advanced automation is laggard; most facilities still rely on manual or semi-automated equipment and prefer operator flexibility over rigid robotics.
Sector adoption velocity	claude-sonnet-5	2/5	Manufacturing, especially plastics/metal forming, is a physically-oriented, lower-digitization sector where automation adoption is slower and more capital-intensive compared to information-based industries.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI-driven vision systems could assist operators by identifying trim locations or alerting to excessive material, but core physical execution (knife trimming, grinder operation) requires human presence. Modest augmentation potential through computer vision feedback, but operator skill remains central.
Augmentation potential	claude-sonnet-5	2/5	AI-driven vision systems can help detect defects or guide trimming lines, offering some assistance, but the core manual cutting and grinding task itself sees little direct AI-based productivity enhancement for the human operator.
Task automatability	claude-haiku-4-5-20251001	2/5	Trimming excess material from parts and grinding scrap plastic require physical manipulation and dexterity in unstructured factory environments. Current robotics and AI can perform repetitive motions on standardized parts, but handling variable geometries, recognizing excess material edges, and safe grinding operations remain beyond reliable automation today.
Task automatability	claude-sonnet-5	2/5	This involves manual trimming with knives and physical grinding of scrap into powder, which requires dexterity and physical manipulation that current AI (software-based) cannot perform; robotic automation exists but is not 'AI' in the generative/agentic sense and requires significant hardware investment.deduction from full automation.
Adoption barriers	claude-haiku-4-5-20251001	2/5	Light regulatory oversight of the automation itself; however, machine safety requirements, worker proximity, and liability for tool-related injuries create modest friction. No licensed professional signature is required, but operator knowledge of material properties and grinder maintenance limits immediate substitution.
Adoption barriers	claude-sonnet-5	2/5	No licensing or legal barriers, but physical workspace constraints, machine safety requirements, and the need for physical retooling for different part geometries create moderate friction.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	Specialized robots for material trimming and grinding are capital-intensive ($100k–$500k+), require significant integration, and need frequent recalibration for part variety. Amortized cost per task cycle remains higher than a human operator's loaded wage for this routine work.
Cost vs. human wage	claude-sonnet-5	2/5	Industrial robotic trimming and grinding equipment requires substantial capital investment, tooling, and maintenance, making it costlier than human labor in many small-scale or variable-part operations despite long-term savings potential.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Some specialized industrial robots exist for deburring and material removal in controlled settings, but general-purpose systems capable of reliably trimming diverse part geometries and managing grinding equipment remain research-stage or extremely narrow in scope. No mainstream deployed product reliably performs both tasks across typical job shop conditions.
Technical feasibility today	claude-sonnet-5	2/5	Robotic deflashing/trimming systems and granulators exist in industrial settings, but these are hardware/mechanical automation solutions rather than AI products, and adoption is uneven across small-to-mid manufacturers.

Measure and visually inspect products for surface and dimension defects to ensure conformance to specifications, using precision measuring instruments.

CI 30–35 · exposure 25 · augmentation 50 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Adoption in metal and plastic casting is slow; most operations remain labor-intensive and fragmented across small to mid-size shops with limited digitization, though large manufacturers have begun pilot vision systems.
Sector adoption velocity	claude-sonnet-5	2/5	Manufacturing, especially metal/plastic molding, is a physical low-digitization sector where automated inspection adoption is growing but remains far behind information-sector AI adoption rates.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI-assisted measurement and flagging of suspicious areas can meaningfully support operators in speed and consistency, while the operator retains final judgment on specification conformance and defect severity.
Augmentation potential	claude-sonnet-5	3/5	AI-assisted vision systems and digital calipers with automated logging can help operators flag potential defects faster and reduce measurement errors, improving throughput while the operator remains responsible for final judgment.
Task automatability	claude-haiku-4-5-20251001	2/5	Visual inspection and dimensional measurement can be partially automated with computer vision systems, but current AI struggles with the full range of surface defects, material-specific interpretation, and real-time decision-making under variable lighting in industrial settings. Setup and oversight remain substantial.
Task automatability	claude-sonnet-5	2/5	Vision-based inspection systems exist for defect detection but full replacement of measurement plus visual inspection with precision instruments across varied part geometries still requires significant custom setup and human judgment for edge cases.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Quality assurance has some regulatory oversight (industry standards, customer contracts) and customer preference for human certification, but no strict licensing requirement prevents AI deployment. Organizational inertia and liability concerns for defects create moderate friction.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement, but quality control liability and the need for human sign-off on defect conformance in regulated industries (aerospace, automotive) create moderate friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Industrial vision systems with integration, calibration, and ongoing maintenance are capital-intensive and require dedicated infrastructure; human operators remain competitive on cost per unit inspected, especially at small-to-medium scales.
Cost vs. human wage	claude-sonnet-5	2/5	Vision inspection systems require significant capital investment, calibration, and integration with existing production lines, making the all-in cost often comparable to or higher than an operator's wage for lower-volume or varied production runs.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Vision-based inspection systems exist in research and limited production settings, but deployment remains narrow and error rates on complex surface defects remain material. Most facilities still rely on human operators for comprehensive conformance checking.
Technical feasibility today	claude-sonnet-5	2/5	Automated optical inspection and machine vision systems are deployed in some manufacturing lines, but many molding/casting operations still rely on manual gauge measurement and visual inspection due to variability in surface defects and part types.

Turn valves and dials of machines to regulate pressure, temperature, and speed and feed rates, and to set cycle times.

CI 30–35 · exposure 25 · augmentation 50 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Manufacturing sectors show slow adoption of full-task automation in molding and casting; most facilities rely on legacy equipment and incremental process control upgrades rather than AI-driven robotic actuation. Pilot projects are rare; production deployment is minimal.
Sector adoption velocity	claude-sonnet-5	2/5	Manufacturing, especially metal/plastic forming, is a physical, moderately digitized sector where automation adoption is real but slow and capital-intensive compared to information-sector adoption patterns.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI-powered monitoring dashboards and predictive parameter recommendations can assist operators in optimizing pressure, temperature, and cycle times, improving their decision-making while they retain manual control of valves and dials.
Augmentation potential	claude-sonnet-5	3/5	Sensor-based monitoring and predictive analytics can help operators fine-tune settings and detect drift, offering useful assistance even if full replacement is limited.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI could theoretically monitor and adjust machine parameters, the task requires direct physical manipulation of valves and dials plus real-time responsiveness to machine conditions. Current AI lacks deployed robotic systems that reliably perform this end-to-end in production foundries, though software monitoring could assist with parameter optimization.
Task automatability	claude-sonnet-5	2/5	While programmable logic controllers and sensors can automate parameter regulation in modern equipment, the task as stated (manual manipulation of valves/dials) implies legacy or semi-manual machines where physical adjustment and tacit judgment are still required, limiting full automation today.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Machine setup and operation may require operator certification or experience validation in some facilities, but no hard legal licensing requirement typically governs valve adjustment itself. Factory safety protocols and machine-specific documentation create moderate organizational friction.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement, but there are safety and quality-control considerations (equipment damage, defective parts) that create some organizational caution before removing human oversight.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	The hardware cost of robotic arms or automated valve systems, plus integration and maintenance, likely exceeds the loaded wage of a machine operator in most metal/plastic molding facilities, especially for smaller and mid-sized operations.
Cost vs. human wage	claude-sonnet-5	2/5	Retrofitting older machines with sensors and automated controls requires significant capital investment, so for many shops the human operator remains cheaper than a full automation retrofit in the short term.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No production systems today reliably set and adjust physical machine controls autonomously in molding/casting environments. Research prototypes exist, but deployed solutions are limited to software-only monitoring; the physical actuation and real-time feedback loop required are not yet mature at scale in this sector.
Technical feasibility today	claude-sonnet-5	2/5	Modern injection molding and casting machines increasingly have digital controls and some closed-loop automation, but many operations still rely on operator adjustment; fully autonomous parameter-setting products are not yet standard across the industry.

Remove finished or cured products from dies or molds, using hand tools, air hoses, and other equipment, stamping identifying information on products when necessary.

CI 30–35 · exposure 25 · augmentation 25 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Manufacturing automation in metalworking and plastics is advancing, but high-mix, lower-volume foundries and molding shops remain labor-intensive. Large-scale, high-volume automotive and consumer goods producers drive robotic adoption, but the broader sector lags behind information and finance.
Sector adoption velocity	claude-sonnet-5	2/5	Manufacturing/metal-plastic molding is a physically-oriented, moderate-digitization sector where robotic adoption is steady but slow and capital-intensive, not the fast AI-driven adoption seen in information/professional services.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI vision systems can assist by detecting defects or flagging out-of-spec parts for operator review, and robotic guidance aids handling of heavy molds. However, the core task of careful extraction and surface inspection remains primarily human-driven, limiting augmentation scope.
Augmentation potential	claude-sonnet-5	2/5	AI-driven vision systems can assist with defect detection or identifying stamping data placement, but they offer limited direct augmentation to the manual extraction and stamping steps themselves.
Task automatability	claude-haiku-4-5-20251001	2/5	Removing products from molds requires spatial reasoning, dexterity, and handling of varied geometries—tasks at which current AI vision + robotics systems struggle reliably. While stamping identifying information is automatable, the primary removal operation involves unpredictable product positions and fragile parts vulnerable to damage, limiting end-to-end automation to narrow, highly standardized scenarios.
Task automatability	claude-sonnet-5	2/5	This is a physical manipulation task requiring dexterity, force judgment, and handling of varied part geometries; current general-purpose AI (LLMs/vision models) cannot perform the physical extraction itself, though robotic automation exists but is not 'AI performing the task' in the generative-AI sense.'
Adoption barriers	claude-haiku-4-5-20251001	3/5	No strict licensing barrier exists, but operator experience and judgment—assessing product condition, detecting defects, handling delicate curing materials—create organizational friction. Safety and quality liability concerns slow automation adoption in molding plants.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement, but safety around hot dies, presses, and stamping equipment creates some liability and workplace-safety compliance friction that slows unstructured automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Robotic systems with vision and manipulation for this task require significant capital investment, integration, and maintenance, often exceeding the loaded wage of a single operator, especially where batch sizes or product variety are moderate.
Cost vs. human wage	claude-sonnet-5	2/5	Robotic de-molding cells require significant capital investment, engineering, and maintenance; for many small-to-medium batch operations this exceeds the cost of a human operator, though at very high volume it can pay off.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Deployed robotic systems for part removal exist in some high-volume settings, but error rates and need for human supervision remain material. Most facilities still rely on human operators because automation integration costs and part-specific tooling changes exceed benefits in moderate-volume production.
Technical feasibility today	claude-sonnet-5	2/5	Dedicated robotic de-molding/de-gating systems exist in high-volume manufacturing, but they are hard-automated/robotic cells engineered per part, not flexible AI systems performing this reliably across diverse molds and product types.

Position and secure workpieces on machines, and start feeding mechanisms.

CI 30–35 · exposure 25 · augmentation 25 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Casting and molding remain relatively traditional, lower-digitization sectors with many small and medium shops. Adoption of full automation for positioning and feeding is slow, with most facilities still relying on skilled human operators despite some robotic pilots in larger firms.
Sector adoption velocity	claude-sonnet-5	2/5	Manufacturing sectors using molding and casting machines have historically slower and more capital-intensive automation adoption compared to information/professional services, though robotic arms are used in some high-volume plants.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI-assisted vision systems could help operators align parts or predict feed rates, but the physical, dexterous nature of positioning and the immediate safety feedback required limit meaningful augmentation. Most augmentation remains at the level of numerical guides or alerts rather than substantial productivity gains.
Augmentation potential	claude-sonnet-5	2/5	AI-driven vision systems and sensors can assist in verifying correct part placement or detecting misalignment, offering some support, but the physical positioning task itself is not significantly transformed by current AI tools.
Task automatability	claude-haiku-4-5-20251001	2/5	While machine vision could identify and guide workpiece placement, this task requires physical manipulation in real-world factory conditions with variable part geometries, fixtures, and safety constraints. Current robotic systems can handle some positioning in controlled settings, but the variability and need for tactile feedback make end-to-end automation with 50% time savings rare without extensive custom engineering.
Task automatability	claude-sonnet-5	2/5	This requires physical manipulation and precise fixturing of workpieces in industrial machines, which current general-purpose AI cannot perform; it requires robotics with specialized end-effectors, not off-the-shelf AI systems.a
Adoption barriers	claude-haiku-4-5-20251001	3/5	While there are no hard legal restrictions on automating workpiece positioning, safety regulations around machinery guarding, lockout-tagout, and liability for dropped or misaligned parts create moderate friction. Worker preference, job security concerns, and the need for human oversight during setup also slow adoption.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement, but safety regulations around machine operation, physical workspace constraints, and the need for custom mechanical integration create moderate friction against quick substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Custom robotic positioning systems and integration are capital-intensive and require ongoing maintenance and programming. For the variable, low-volume casting jobs typical in this sector, per-unit automation costs often exceed the loaded wage of a skilled machine operator.
Cost vs. human wage	claude-sonnet-5	2/5	Industrial robotic automation for this task requires significant capital investment in custom tooling, fixtures, and integration, making it often costlier than human labor except in very high-volume, standardized production runs.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Deployed robots exist for picking and placing in manufacturing, but reliable production performance across diverse metal and plastic casting workpieces remains limited. Most existing systems operate in narrow, highly controlled scenarios; general-purpose positioning and feeding remain largely handled by human operators with occasional robotic assist.
Technical feasibility today	claude-sonnet-5	2/5	Robotic loading/unloading systems exist in some high-volume manufacturing lines, but generalized positioning and securing of varied workpieces across molding/casting setups is still narrow and engineering-intensive, not broadly deployed as a flexible product.

Mix and measure compounds, or weigh premixed compounds, and dump them into machine tubs, cavities, or molds.

CI 30–35 · exposure 20 · augmentation 25 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Manufacturing adoption of AI/robotics is slower than information-sector adoption; compound mixing and dumping is a lower-value per-unit task competing for investment with higher-ROI automation targets in the molding industry.
Sector adoption velocity	claude-sonnet-5	2/5	Metal/plastic molding manufacturing is a physical, moderately digitized sector where automation adoption is steady but slow compared to information-based industries, and this specific task lags behind more digitizable clerical work.
Augmentation potential	claude-haiku-4-5-20251001	2/5	Limited augmentation opportunity: automated measurement displays or scale integration could assist, but the core dumping action and material handling fundamentally require human physical presence and judgment in most current shop-floor setups.
Augmentation potential	claude-sonnet-5	2/5	Sensors and control systems can assist with measurement precision and process monitoring, but general AI tools offer limited direct augmentation for the physical act of mixing and dumping materials.
Task automatability	claude-haiku-4-5-20251001	2/5	Mixing and measuring compounds requires integration of scales, dumping equipment, and precise coordination in factory settings; current AI lacks reliable physical manipulation, computer vision for precise measurement verification, and safe autonomous handling of industrial machinery in unstructured environments.
Task automatability	claude-sonnet-5	2/5	This is a physical material-handling task requiring robotic manipulation and sensor-based measurement in a variable factory environment; current general AI systems cannot perform the physical dumping/mixing, though dedicated automation (not general AI) can handle portions.dosage.
Adoption barriers	claude-haiku-4-5-20251001	2/5	This is a hands-on industrial task with minimal licensing barriers, but operator expertise and safety compliance (handling chemicals, machinery guarding) create moderate organizational friction to full displacement.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement, but physical integration, safety regulations around chemical/material handling, and equipment compatibility create moderate organizational and capital barriers to substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Robotic solutions for compound handling exist but are expensive to deploy, require significant integration, and are typically cost-justified only in high-volume facilities; for most small-to-medium molding operations, human labor remains cheaper than full automation.
Cost vs. human wage	claude-sonnet-5	2/5	Retrofitting a manufacturing line with automated dispensing/mixing hardware and controls requires significant capital investment, so near-term cost may exceed a human operator's wage unless already integrated into a high-volume production line.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed production systems autonomously perform industrial compound mixing, measuring, and dumping at scale; this remains a hands-on manufacturing task without mature end-to-end automation products in general use.
Technical feasibility today	claude-sonnet-5	2/5	Automated dosing/dispensing systems exist in some plants but are hard-wired industrial automation rather than deployed AI products, and adoption for full compound mixing/measuring/dumping cycles remains narrow and equipment-specific.

Set up, operate, or tend metal or plastic molding, casting, or coremaking machines to mold or cast metal or thermoplastic parts or products.

CI 30–30 · exposure 25 · augmentation 50 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Manufacturing sectors that perform molding and casting are early-stage in AI/robotic adoption. While some large fabricators pilot vision-based inspection and basic robotics, the majority of molding shops remain manual-operator-dependent, with adoption concentrated in high-volume, high-margin production lines rather than broad across the sector.
Sector adoption velocity	claude-sonnet-5	2/5	Manufacturing is a physical, moderately digitized sector where robotic automation adoption is steady but slow compared to information-sector AI adoption, with many facilities still relying on manual tending.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI-assisted monitoring (computer vision for defect detection, predictive maintenance alerts) and automated parameter logging can meaningfully assist operators in quality control and equipment upkeep, raising productivity on inspection and record-keeping aspects while the operator remains responsible for setup and real-time troubleshooting.
Augmentation potential	claude-sonnet-5	3/5	AI-driven predictive maintenance, quality inspection, and process optimization tools can meaningfully assist operators in monitoring and adjusting machine performance, though the core physical operation still requires human presence.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI could assist with monitoring and basic parameter adjustments, metal and plastic molding/casting involves complex physical setup, real-time quality control, and troubleshooting of industrial equipment that current AI cannot perform end-to-end reliably. The task requires hands-on equipment operation and in-person intervention for jams, temperature fluctuations, and material variations—domains where automation falls well short of 50% time savings at equal quality.
Task automatability	claude-sonnet-5	2/5	This is a physical machine setup/operation task requiring manual handling of molds, materials, and machine adjustments that current AI systems cannot perform end-to-end without robotic embodiment.atable software agents alone do not suffice.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Physical safety requirements and equipment manufacturer liability create moderate friction; OSHA regulations and machine-guarding laws apply, but do not strictly forbid automation. Operator skill and judgment remain valued and difficult to fully replace, creating organizational reluctance alongside technical barriers.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement for the worker, but safety regulations, quality control standards, and capital costs of retrofitting machinery create real friction against rapid substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Integration of AI monitoring or robotic arms to assist molding/casting operations remains expensive relative to the wages of skilled operators. The capital cost of automation hardware, integration, and maintenance exceeds the operational savings on a per-task basis for most small-to-medium molding shops.
Cost vs. human wage	claude-sonnet-5	2/5	Industrial automation equipment for this task requires significant capital investment in robotics and sensors, and integration/maintenance costs often exceed savings versus a machine operator's wage, especially for small-batch or varied production.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No deployed AI products reliably perform the core setup and operation of metal/plastic molding machinery at production scale. Computer vision and sensors can assist with defect detection, but the full task of setting parameters, loading material, monitoring cycles, and responding to equipment faults remains manual. Research prototypes exist; production-grade autonomous systems do not.
Technical feasibility today	claude-sonnet-5	2/5	While CNC-style automation and PLC-controlled molding machines exist, fully autonomous setup, changeover, and tending by AI/robotics in production settings remains limited and typically still requires human operators for calibration and quality checks.

Read specifications, blueprints, and work orders to determine setups, temperatures, and time settings required to mold, form, or cast plastic materials, as well as to plan production sequences.

CI 25–30 · exposure 25 · augmentation 50 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Manufacturing, particularly small-to-medium molding shops, has historically lagged in digitization and AI adoption. Most facilities still rely on manual interpretation of blueprints by experienced operators, and digital workflow transformation is uneven across the sector.
Sector adoption velocity	claude-sonnet-5	2/5	Manufacturing, especially small-to-mid plastics/metal shops, has historically been slow to adopt AI-driven configuration tools compared to information-sector jobs, with adoption concentrated in pilots at larger firms.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI could usefully assist by automatically highlighting parameters from blueprints, cross-checking values against material databases, flagging anomalies, or suggesting setup sequences for operator review. Such augmentation could reduce cognitive load and error rate while the human operator remains responsible for final decisions and execution.
Augmentation potential	claude-sonnet-5	3/5	AI-assisted CAM software, parameter optimization tools, and digital work order systems can help operators interpret specs faster and suggest settings, meaningfully aiding but not replacing the task.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can parse specifications and blueprints to extract parameters, the task requires interpreting complex technical drawings, cross-referencing multiple documents, and determining optimal setup sequences that depend on material properties, equipment state, and production constraints. Current AI can assist with parsing but cannot reliably handle the full end-to-end decision-making at equal quality without substantial human oversight.
Task automatability	claude-sonnet-5	2/5	Interpreting varied blueprints, translating them into precise machine setups, and sequencing production requires physical-world knowledge and judgment that current AI cannot reliably perform end-to-end without heavy human verification.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Liability concerns are significant: incorrect temperature or timing settings can damage equipment, waste material, or create unsafe conditions, creating asymmetric error costs. Quality control requirements, equipment-specific calibration, and the need for a qualified human to sign off on production parameters all create friction against full automation.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement, but liability for defective parts, safety concerns around machine setup errors, and reliance on tacit shop-floor knowledge create moderate organizational friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Integration of document parsing AI with manufacturing systems, training, and necessary human oversight is substantial relative to the cost of a skilled operator reading and executing specifications. The wage loaded cost of a few minutes of human work is still low compared to the total cost of ownership for AI infrastructure and error correction.
Cost vs. human wage	claude-sonnet-5	2/5	Deploying AI systems capable of reading specs and configuring machine parameters requires significant integration with legacy manufacturing equipment, so costs remain comparable to or higher than experienced operators for many shops.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Vision models can read documents and blueprints, and LLMs can extract structured information from specifications, but no deployed product reliably translates these into complete, production-ready setup parameters and sequences without human validation. Existing systems lack the tight feedback loop needed to verify correctness against actual equipment behavior.
Technical feasibility today	claude-sonnet-5	2/5	Some CAM/CAD software and process-parameter recommendation tools exist, but they typically assist rather than autonomously determine full setups from raw blueprints and work orders in production settings.

Spray, smoke, or coat molds with compounds to lubricate or insulate molds, using acetylene torches or sprayers.

CI 19–35 · exposure 13 · augmentation 25 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Manufacturing has moderate AI/automation adoption overall, but spray coating and mold preparation remain labor-intensive in small-to-mid-tier foundries and plastic molding shops. Larger OEMs may use specialized robots, but penetration is slow and sector-specific.
Sector adoption velocity	claude-sonnet-5	1/5	Metal/plastic casting manufacturing is a physical, lower-digitization sector with slow automation uptake for this specific sub-task compared to information-sector AI adoption.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI vision systems could assist with defect detection or coating uniformity feedback post-application, but the physical act of spraying or torching—the core of the task—offers limited opportunities for meaningful human-AI collaboration in current tools.
Augmentation potential	claude-sonnet-5	2/5	AI could help optimize spray patterns or predict maintenance schedules via sensor data, but it offers minimal direct assistance to the hands-on spraying/coating action itself.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires precise physical manipulation of spraying/torching equipment in a 3D manufacturing environment with real-time sensory feedback and safety-critical decisions. Current AI lacks the dexterity, spatial reasoning in unstructured factory settings, and ability to respond to mold variability that this task demands.
Task automatability	claude-sonnet-5	2/5	This is a physical manual task requiring dexterity and judgment about coverage and mold condition; current AI systems (software/LLMs) cannot perform it, and robotic automation, while it exists, is not 'AI' in the generally-available sense and requires heavy hardware integration.utors.
Adoption barriers	claude-haiku-4-5-20251001	2/5	Safety regulations (fire codes, chemical handling, worker protection) create some friction, and custom mold geometries require operator oversight. However, no legal licensing requirement prevents automation attempts; the barriers are primarily technical and economic rather than regulatory.
Adoption barriers	claude-sonnet-5	2/5	No licensing or legal requirement for a human to do this, but physical workspace safety, equipment retrofitting, and variable mold geometries create moderate organizational friction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Specialized robotic systems for spray application exist but carry high capital costs ($200k–$500k+), integration overhead, and safety compliance expenses that exceed the loaded wage of a skilled operator in most metalworking shops. ROI is marginal for typical production runs.
Cost vs. human wage	claude-sonnet-5	2/5	Robotic spraying systems can be cheaper per unit at high volume, but the upfront capital, integration, and maintenance costs make it non-trivially more expensive than a human tender in most small-to-mid scale operations.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed AI products perform autonomous spray coating or torch-based mold treatment at production scale. The task requires manipulation of open flames/chemical sprays and judgment about coating uniformity and mold-specific conditions—beyond the scope of current commercial automation.
Technical feasibility today	claude-sonnet-5	2/5	Some robotic spray-coating systems exist in high-volume foundries, but they are engineered automation systems rather than deployed general AI products, and adoption is narrow and capital-intensive.

Pour or load metal or sand into melting pots, furnaces, molds, or hoppers, using shovels, ladles, or machines.

CI 23–30 · exposure 25 · augmentation 25 · importance 3.4/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Metal and plastic casting remains concentrated in traditional, physical manufacturing sectors with slower digitization and lower capital availability than software/finance. While large integrated foundries automate, small and mid-sized shops adopt piecemeal, and adoption remains pilot-heavy rather than deep.
Sector adoption velocity	claude-sonnet-5	1/5	Manufacturing/foundry work is a low-digitization, physical-labor-intensive sector with slow AI/robotics adoption compared to information-based industries.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI offers limited real-time assistance to this physical pouring task; basic material-flow monitoring and alerts exist, but the core task of load-and-pour lacks the document review, recommendation, or decision-support augmentation seen in knowledge work. Human control and physical intuition remain dominant.
Augmentation potential	claude-sonnet-5	2/5	AI-driven sensors and monitoring can assist operators with timing, temperature control, and quality checks, but does not fundamentally transform the physical pouring/loading task itself.
Task automatability	claude-haiku-4-5-20251001	2/5	While robotic loaders and automated material handling systems exist, pouring molten metal or managing sand with precision into varied mold configurations requires adaptive control, thermal sensing, and variable geometry handling that current AI-supervised automation struggles with end-to-end at 50% time savings. Human judgment on temperature, flow rate, and mold positioning remains critical.
Task automatability	claude-sonnet-5	2/5	This is a physical material-handling task requiring robotic manipulation of hot, heavy, or hazardous materials; current general-purpose AI cannot perform this end-to-end, though task-specific automation (e.g., automated pouring machines) exists as engineered hardware rather than 'AI' per se.'
Adoption barriers	claude-haiku-4-5-20251001	4/5	Foundry safety regulations mandate machine guarding, operator certification, and liability oversight; thermal and spill hazards create legal and insurance obligations that slow automation deployment. OSHA standards and worksite safety inspections create friction that favors human operators with formal training and sign-off responsibility.
Adoption barriers	claude-sonnet-5	3/5	No licensing barrier, but safety regulations around molten metal handling, workplace safety standards, and capital/organizational inertia in retrofitting foundries create moderate friction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Robotic pouring and material-handling systems are capital-intensive (installation, integration, maintenance) and typically justify only for high-volume, standardized casting runs. For variable small-batch work, human labor remains cost-competitive when amortized across diverse tasks and changeover requirements.
Cost vs. human wage	claude-sonnet-5	2/5	Automated pouring/casting equipment requires significant capital investment in specialized machinery, molds, and safety systems, often exceeding the cost of human labor unless at very high production volumes.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Heavy machinery automation (bucket loaders, conveyors) is deployed at scale, but intelligent pouring systems that adapt to different mold sizes, metal states, and safety conditions remain experimental. Narrow implementations exist but lack the flexibility and reliability needed for general production deployment across foundry operations.
Technical feasibility today	claude-sonnet-5	2/5	Some foundries use automated ladle/pouring systems, but these are specialized industrial automation, not adaptable AI products; most operations still rely on human operators or fixed robotics rather than flexible AI-driven systems.

Select coolants and lubricants, and start their flow.

CI 14–35 · exposure 20 · augmentation 25 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Manufacturing sectors, especially small to mid-size metal and plastic casting operations, adopt AI slowly. This specific task remains manually performed across most facilities with minimal digitization pressure.
Sector adoption velocity	claude-sonnet-5	2/5	Manufacturing and metal/plastic fabrication sectors are slower AI/automation adopters compared to information-based industries, though process automation and Industry 4.0 initiatives are growing gradually.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI could assist by recommending optimal coolant/lubricant selections based on material and temperature data, but the physical workflow and equipment-specific knowledge limits meaningful productivity gain over human operator judgment.
Augmentation potential	claude-sonnet-5	2/5	Sensors and monitoring systems can alert operators to optimal coolant/lubricant timing and flow rates, offering modest assistance, but the selection and manual start action remains largely human-driven.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI could theoretically identify appropriate coolants and lubricants from specifications, the physical task of starting their flow requires manual intervention on industrial equipment. Current AI systems cannot reliably end-to-end automate equipment startup and monitoring with 50% time savings.
Task automatability	claude-sonnet-5	2/5	This is a brief physical selection and setup step tied to specific machine configurations; while control logic can be automated in modern equipment, the generalized task across varied machines and materials requires physical presence and judgment not yet fully replaceable off-the-shelf.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Safety regulations, equipment-specific proprietary systems, liability for machinery malfunction, and the requirement for on-site human operators monitoring equipment create strong organizational and regulatory barriers to full automation.
Adoption barriers	claude-sonnet-5	2/5	No licensing barriers, but physical presence at machinery, safety protocols, and integration into a shop floor's existing equipment create moderate friction against automation.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	The equipment interaction cost (sensors, actuators, integration) combined with AI system costs significantly exceeds the hourly wage of a machine tender for this brief, routine task.
Cost vs. human wage	claude-sonnet-5	2/5	Retrofitting automated coolant/lubricant systems requires capital investment in sensors and dosing equipment, which may not be cheaper than a human operator performing this quick task within a broader tending role.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed AI product reliably performs physical coolant/lubricant selection and activation on molding/casting machines in production environments. This task requires direct equipment interaction and real-time adjustment that only exists in research or limited pilot settings.
Technical feasibility today	claude-sonnet-5	2/5	Some modern CNC-integrated molding machines have automated coolant/lubricant delivery systems, but many operations still rely on manual selection and startup, especially in older or smaller-scale plants.

Perform maintenance work such as cleaning and oiling machines.

CI 24–24 · exposure 16 · augmentation 25 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Manufacturing maintenance remains heavily manual; while large facilities may use condition-monitoring sensors, autonomous physical maintenance automation is not yet standard practice even in advanced plants.
Sector adoption velocity	claude-sonnet-5	1/5	Manufacturing floor physical maintenance tasks are in a low-digitization, low-robotics-adoption sector for this specific task type.
Augmentation potential	claude-haiku-4-5-20251001	2/5	Predictive maintenance software and IoT sensors can alert operators to maintenance needs, modestly improving scheduling efficiency, but AI provides limited real-time assistance during the actual hands-on cleaning and oiling work itself.
Augmentation potential	claude-sonnet-5	2/5	AI could assist with predictive maintenance scheduling or diagnostics, but offers minimal direct help with the physical act of cleaning and oiling.
Task automatability	claude-haiku-4-5-20251001	2/5	While some machine monitoring and cleaning schedules could be partially automated (e.g., sensors detecting need for maintenance), the physical manipulation of cleaning and oiling machines requires dexterous robotics that current general-purpose AI systems cannot reliably perform end-to-end without substantial custom engineering and integration.
Task automatability	claude-sonnet-5	2/5	Basic cleaning and oiling requires physical manipulation of machinery in varied conditions; current AI systems lack the embodied robotic capability to reliably perform this end-to-end.'
Adoption barriers	claude-haiku-4-5-20251001	2/5	Physical safety, equipment liability (damage from incorrect maintenance), and the need for operator judgment about machine condition create some friction, though no hard legal barriers prevent automation of maintenance tasks in manufacturing.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement, but physical access, safety protocols around machinery, and organizational maintenance schedules create some friction against quick substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	Specialized robotic systems capable of machine maintenance are expensive to purchase, integrate, and maintain, making them more costly than employing a human machine tender for routine cleaning and oiling tasks.
Cost vs. human wage	claude-sonnet-5	1/5	Without a viable automated solution, there is no AI cost basis to compare; human labor remains the only practical option.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed AI product reliably performs autonomous machine maintenance cleaning and oiling in production foundry or plastic molding environments today; this remains a domain-specific robotics challenge without mature commercial solutions.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product autonomously performs machine cleaning and oiling maintenance in production manufacturing settings; this remains a manual physical task.

Adjust equipment and workpiece holding fixtures, such as mold frames, tubs, and cutting tables, to ensure proper functioning.

CI 10–30 · exposure 8 · augmentation 25 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Manufacturing sectors adopting casting automation have focused on pouring and cooling operations, not fixture adjustment. Small and mid-tier foundries—where this task is most common—show low automation velocity and rely on skilled operator expertise.
Sector adoption velocity	claude-sonnet-5	2/5	Manufacturing/metalworking sectors are historically slower adopters of AI-driven physical automation compared to information-based industries, with adoption concentrated in large-scale automated plants rather than broad deployment.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI vision systems could assist by inspecting fixture alignment and alerting operators to misalignment, but the core manual adjustment requires human judgment and physical dexterity, limiting meaningful augmentation on this particular task.
Augmentation potential	claude-sonnet-5	2/5	AI-based sensors and predictive maintenance tools can flag issues or suggest adjustments, offering some assistance, but the core physical adjustment work still requires substantial human skill and judgment.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires physical manipulation of heavy, diverse equipment and real-time sensory feedback to verify proper alignment and functioning. Current AI systems lack the dexterous robotic capabilities and spatial reasoning needed to reliably adjust fixtures across varied metal and plastic casting setups.
Task automatability	claude-sonnet-5	2/5	This requires physical manipulation, sensory feedback, and dexterity to adjust fixtures and equipment, which current AI systems cannot perform end-to-end without robotic embodiment specialized for this exact task.'},
Adoption barriers	claude-haiku-4-5-20251001	2/5	While there are no strict licensing barriers, safety regulations governing machinery operation and the physical nature of the work create modest friction. The task is also tightly integrated with operator judgment about proper alignment, reducing pure substitution potential.
Adoption barriers	claude-sonnet-5	3/5	No licensing barrier, but safety, quality control, and machine-specific expertise create organizational friction and liability concerns around unsupervised equipment adjustment.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	Specialized robotic systems capable of fixture adjustment would require substantial capital investment, integration costs, and maintenance, far exceeding the loaded wage of a skilled machine operator who performs this task as part of their role.
Cost vs. human wage	claude-sonnet-5	2/5	Specialized automation/robotics could eventually be cheaper, but current AI-driven solutions for this specific fixture adjustment task require significant capital investment in sensors and robotics, making all-in cost comparable or higher than a skilled operator today.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed product reliably performs end-to-end fixture adjustment and verification in production casting environments. While industrial robots exist, they require task-specific programming and cannot handle the variability in mold frames, tubs, and cutting tables across different operations.
Technical feasibility today	claude-sonnet-5	1/5	No deployed general-purpose product performs this specific physical setup and adjustment task reliably in production; specialized robotics exist but are not general AI performing this diagnostic-adjustment task autonomously.

Remove parts, such as dies, from machines after production runs are finished.

CI 5–26 · exposure 8 · augmentation 25 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Manufacturing sectors adopting automation in metal and plastic casting are typically large operations with high-volume runs, but physical part removal remains labor-intensive and underautomated compared to information tasks. Adoption is slow, pilot-stage, and limited to high-value operations.
Sector adoption velocity	claude-sonnet-5	2/5	Manufacturing floor tasks involving physical part handling see slow AI adoption compared to office/information work, though some robotic automation exists in high-volume plants.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI offers minimal assistance for removing dies from machines; computer vision might help locate or identify parts, but the core task—physical extraction—remains purely manual. Augmentation potential is limited without significant robotics integration.
Augmentation potential	claude-sonnet-5	2/5	AI can support scheduling, predictive maintenance alerts, or digital work instructions, but offers minimal direct assistance in the physical act of removing dies.
Task automatability	claude-haiku-4-5-20251001	1/5	This task involves physical removal of dies and parts from machines, requiring manual dexterity, spatial navigation, and interaction with heavy industrial equipment. Current AI systems lack the embodied manipulation capabilities and real-time environmental adaptation needed to perform this safely and reliably.
Task automatability	claude-sonnet-5	2/5	This is a manual, physical task requiring dexterity and force to remove heavy dies/molds from machinery; current AI (software) cannot perform this physical action, and robotics for this specific unstructured task are not off-the-shelf solutions.ed
Adoption barriers	claude-haiku-4-5-20251001	4/5	Significant barriers exist: workplace safety regulations require human oversight of machinery, liability for equipment damage or injury, and the physical proximity needed to machines creates inherent human-contact requirements. Factory layouts are often customized, creating setup friction.
Adoption barriers	claude-sonnet-5	2/5	No licensing or legal barrier exists, but physical workplace safety protocols and machine-specific handling procedures create moderate practical friction against pure AI/robotic substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	Industrial robots capable of die removal and part handling are capital-intensive (six figures to millions), require custom integration, and demand ongoing maintenance. This cost far exceeds the loaded wage of a machine tender performing this task.
Cost vs. human wage	claude-sonnet-5	1/5	There is no viable AI-only solution to compare cost against; any automation would require custom robotic engineering far exceeding the cost of a human operator performing this routine task.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed product reliably performs industrial part removal from production machines at scale. While robotics research exists, production-grade systems capable of safely handling dies, avoiding thermal hazards, and working within existing factory layouts remain limited and not commercially standard.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product performs generalized die/mold removal from diverse molding machines; this remains a manual or specialized fixed-automation task, not an AI-driven one.

Connect water hoses to cooling systems of dies, using hand tools.

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

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Metalworking and plastics manufacturing remain capital-constrained sectors with low digital penetration and strong reliance on skilled labor; adoption of advanced robotic systems for such specialized physical tasks is rare and slower than in software-centric or logistics sectors.
Sector adoption velocity	claude-sonnet-5	1/5	Manufacturing floor tasks involving physical hose connections show minimal AI/robotic adoption; this is a laggard sector for this specific manual task despite some robotics in adjacent processes.
Augmentation potential	claude-haiku-4-5-20251001	1/5	AI offers minimal assistance for a task that is primarily physical assembly; computer vision or planning tools could theoretically help with hose routing documentation, but the task as stated is hands-on connection work where human tactile feedback and adaptation are essential.
Augmentation potential	claude-sonnet-5	1/5	AI offers essentially no assistance for the physical act of connecting hoses with hand tools; this is a manual mechanical task outside current AI's scope of support.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires precise physical manipulation of hoses and hand tools in a specific three-dimensional manufacturing environment. Current AI systems cannot perform end-to-end robotic assembly of cooling system connections with the dexterity and spatial reasoning needed, nor can they safely and reliably handle the varied physical configurations of dies and hoses in a shop floor context.
Task automatability	claude-sonnet-5	1/5	This is a physical manual task requiring dexterity to connect hoses with hand tools; no current AI system can perform this physical manipulation end-to-end.imb Robotics could theoretically do this but it's not an 'AI' automation in the software sense being assessed here.
Adoption barriers	claude-haiku-4-5-20251001	2/5	There are minimal regulatory or licensing barriers to automation itself, and no hard legal requirement for human sign-off. However, the task occurs in skilled manufacturing environments where integration friction and the specificity of die configurations to each production run create practical adoption friction.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement exists, but physical dexterity, variable die configurations, and safety around hydraulic/water connections create practical friction against automation.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	The capital and integration cost of a robotic system capable of hose connection work, plus ongoing maintenance and oversight, far exceeds the loaded wage of a skilled machine operator performing this task. The task is relatively quick and cost-effective for a human to perform.
Cost vs. human wage	claude-sonnet-5	1/5	There is no viable AI-driven solution performing this physical task at any meaningful scale, so cost comparison to human labor is inapplicable or favors the human by default.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed products reliably perform this physical assembly task autonomously in production metalworking or plastics environments. While some robotic arms exist in manufacturing, they are not general-purpose solutions for hose connection work and would require extensive custom engineering and programming for each die type.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product connects water hoses to cooling systems using hand tools; this requires physical robotic manipulation, which remains research-stage for such variable industrial tasks.

Smooth and clean inner surfaces of molds, using brushes, scrapers, air hoses, or grinding wheels, and fill imperfections with refractory material.

CI 15–15 · exposure 0 · augmentation 13 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Foundry and molding shops remain predominantly small to mid-sized, physical-process businesses with low digitization rates. Adoption of automation in this sector has historically lagged information industries, and no public data shows rapid AI adoption in mold finishing.
Sector adoption velocity	claude-sonnet-5	1/5	Metal/plastic manufacturing floor work involving physical mold maintenance is a low-digitization, physical-labor sector with minimal AI agent deployment for this type of task.
Augmentation potential	claude-haiku-4-5-20251001	2/5	Augmentation is minimal because the task is primarily manual skill and dexterity. While AI vision systems could theoretically help identify imperfections on mold surfaces, the core work—physical manipulation with tools—remains unaided by current assistive AI.
Augmentation potential	claude-sonnet-5	1/5	Current AI offers essentially no assistance for the tactile, hands-on process of smoothing mold surfaces or applying refractory material.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires dexterous physical manipulation in confined spaces (inner mold surfaces), tactile feedback to detect imperfections, and judgment about material application—capabilities far beyond current AI systems. No end-to-end automation of the smoothing, cleaning, and filling operations exists today.
Task automatability	claude-sonnet-5	1/5	This is a manual, tactile task requiring physical dexterity to inspect and manipulate tools inside mold cavities; no off-the-shelf AI system can perform this physical work end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	2/5	While not legally restricted, there are practical barriers: worker preference for manual craftsmanship, small shop sizes, and the need for customization and real-time quality judgment reduce substitution pressure, though these are organizational rather than regulatory barriers.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement exists, but the physical nature of the task and need for specialized robotic hardware (not AI software) creates practical barriers to any automation, AI-based or otherwise.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	Specialized equipment (grinding wheels, air systems, refractory materials) plus integration costs would exceed the loaded wage of a skilled mold operator, especially given the low volume of such niche automation relative to widespread manual labor.
Cost vs. human wage	claude-sonnet-5	1/5	There is no AI-based alternative to compare costs against; the task requires physical robotic manipulation which, if it existed, would likely be far more expensive than a human operator for this variable, low-volume task.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed robotic or AI systems reliably perform inner mold surface finishing and refractory material application in production environments. The task demands dynamic adaptation to variable mold geometry and quality assessment that current industrial automation cannot achieve.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product performs mold cleaning, smoothing, or refractory patching; this remains purely a manual/physical labor task performed by skilled workers or basic tools.

Obtain and move specified patterns to work stations, manually or using hoists, and secure patterns to machines, using wrenches.

CI 10–18 · exposure 5 · augmentation 25 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	2/5	Metal and plastic casting/molding is a traditional, slower-digitizing sector where automation adoption remains below professional services or finance; most shops rely on human operators for setup and material handling tasks.
Sector adoption velocity	claude-sonnet-5	1/5	Metal/plastic molding manufacturing is a physically intensive, low-digitization sector with slow uptake of AI/robotics for such manual material-handling tasks.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI could assist with pattern identification or location tracking via computer vision, but the core physical labor of moving and securing patterns offers limited augmentation opportunity without active human involvement in the mechanical work.
Augmentation potential	claude-sonnet-5	2/5	AI could assist with scheduling or optimizing which patterns to retrieve, but offers little direct help with the physical retrieval, hoisting, and wrench-based securing itself.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires physical manipulation in unstructured factory environments, moving patterns to multiple workstations and securing them with hand tools. Current AI/robotics cannot reliably perform the full end-to-end process of locating patterns, transporting them with hoists, and securely fastening them with wrenches across varying shop layouts.
Task automatability	claude-sonnet-5	1/5	This is a physical manipulation task requiring locating, transporting, and mechanically securing heavy patterns using hoists and wrenches; current AI systems cannot perform physical manual labor of this kind end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Factory environments have some automation-readiness, but the task involves safety-critical hoist operation and variability in part dimensions/locations that create moderate organizational and safety friction rather than absolute legal barriers.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement, but physical workplace safety protocols, machine-specific tooling, and the need for physical dexterity with hoists and wrenches create real operational friction against automation.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	A specialized robotic system capable of hoisting, positioning, and tool manipulation would require significant capital investment, integration, and maintenance costs that far exceed the loaded wage of a single machine operator performing this routine task.
Cost vs. human wage	claude-sonnet-5	1/5	Any automation would require expensive custom robotics/hoist integration far exceeding the cost of a human operator performing this routine physical task.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While robotic arms exist in manufacturing, deployed systems rarely handle the complete workflow of fetching patterns from variable storage, moving them via hoists, and manual fastening with wrenches. Most production automation focuses on narrower, fixed-sequence tasks rather than this multi-station, tool-requiring operation.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product performs this physical pattern-handling and fastening task in production; robotic solutions for such variable, heavy-object handling remain research/custom-integration stage, not general AI products.

Preheat tools, dies, plastic materials, or patterns, using blowtorches or other equipment.

CI 0–28 · exposure 8 · augmentation 25 · importance 3.6/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Metal and plastic casting manufacturing is a traditional sector with slower digitization; most shops still rely on human operators for preheating due to the safety-critical nature and low adoption of AI-capable robotics in this domain.
Sector adoption velocity	claude-sonnet-5	1/5	Metal/plastic manufacturing floor tasks are a laggard sector for AI adoption; automation here trends toward traditional industrial robotics/PLCs rather than AI agents, with slow uptake of new physical automation.
Augmentation potential	claude-haiku-4-5-20251001	2/5	While temperature monitoring sensors could provide real-time alerts or recommendations, the core task of physically operating preheating equipment offers limited augmentation potential since judgment and physical control remain essential and inseparable.
Augmentation potential	claude-sonnet-5	2/5	AI could assist with monitoring temperature sensors or optimizing preheat schedules via predictive analytics, but offers limited direct assistance to the physical act of preheating itself.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires physical manipulation of equipment like blowtorches, monitoring temperature conditions, and handling materials in a manufacturing environment. Current AI systems cannot physically operate blowtorches, adjust equipment, or safely manage heat-intensive processes in the real world.
Task automatability	claude-sonnet-5	2/5	This is a physical manipulation task requiring handling of blowtorches/heating equipment on varied materials, which current AI systems cannot perform end-to-end without robotic embodiment specifically engineered for this purpose.'
Adoption barriers	claude-haiku-4-5-20251001	5/5	This task involves direct operation of open flames and high-temperature equipment in a regulated manufacturing environment, requiring OSHA compliance, safety certifications, and legal liability for proper execution and worker safety.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement for this task itself, but physical workspace safety, equipment variability, and capital costs for robotic retrofitting create moderate practical barriers to AI/robotic substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	Robotic systems capable of autonomous preheating with blowtorches are expensive, specialized equipment with high capital and maintenance costs, making them far more expensive than the loaded wage of a skilled operator.
Cost vs. human wage	claude-sonnet-5	2/5	Specialized robotic/automated preheating systems exist but require significant capital investment in custom hardware, making them costlier than human labor for many small-to-medium operations.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed AI product can autonomously perform the physical preheating of tools and materials using blowtorches or equivalent equipment in an industrial setting today.
Technical feasibility today	claude-sonnet-5	1/5	No deployed general AI product performs manual preheating of tools/dies/patterns; existing automation in foundries uses fixed industrial controllers, not AI-driven robotic systems for this specific task.

Install dies onto machines or presses and coat dies with parting agents, according to work order specifications.

CI 5–21 · exposure 8 · augmentation 25 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Manufacturing—especially small- and mid-sized metal and plastic casting shops—lags in AI/automation adoption. Die setup is a specialized, low-volume task; only large facilities with high throughput and capital investment have deployed robotics, and human tenders remain dominant.
Sector adoption velocity	claude-sonnet-5	2/5	Manufacturing sectors, especially metal/plastic forming, show slow and uneven automation adoption for physical setup tasks compared to information-based industries.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI could assist by providing work-order interpretation, visual guidance, or pressure/temperature alerts during coating, but the core physical tasks of installation and application remain human-performed. Augmentation potential is limited because the operator must remain present for precision and safety compliance.
Augmentation potential	claude-sonnet-5	2/5	AI-driven monitoring or work-order digitization can support scheduling and quality checks, but offers limited direct assistance in the physical act of die installation and coating.
Task automatability	claude-haiku-4-5-20251001	1/5	Installing dies onto machines and coating them with parting agents requires precise physical manipulation, spatial positioning, and real-time adjustment in a factory setting—capabilities current AI systems cannot reliably perform. This task fundamentally depends on robotic embodiment and tactile feedback that deployed AI lacks.
Task automatability	claude-sonnet-5	2/5	This is a physical manipulation task requiring precise die installation and manual coating with parting agents; current AI (software/vision) cannot itself perform the physical labor, though robotics could theoretically assist in narrow, pre-engineered setups.dish
Adoption barriers	claude-haiku-4-5-20251001	4/5	Manufacturing safety regulations, machine-specific lockout/tagout (LOTO) procedures, and operator liability for equipment damage or defects create significant friction. Responsibility for correct die installation and parting-agent application typically rests with a licensed or certified operator, not an automated system.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement, but safety protocols, precision tolerances, and equipment-specific procedures create meaningful organizational and technical friction against quick automation.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	A trained operator costs a fraction of specialized industrial robotics with integration, plus the ongoing infrastructure and maintenance. The capital and operational expense of automating this task far exceeds the loaded wage of the human operator.
Cost vs. human wage	claude-sonnet-5	1/5	Robotic die-handling systems require expensive custom integration, tooling, and maintenance far exceeding the cost of a trained machine operator for this task.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No production AI system reliably performs this multi-step physical installation and coating task at manufacturing scale. While robotic arms exist, integrating them with spray coating, die alignment, and machine-specific setup remains a custom engineering problem, not a deployed product.
Technical feasibility today	claude-sonnet-5	1/5	No widely deployed product autonomously installs dies onto presses and applies parting agents in general manufacturing settings; this remains a manual or specialized robotic-cell task, not achieved via general AI systems.

Operate hoists to position dies or patterns on foundry floors.

CI 0–21 · exposure 8 · augmentation 25 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Foundry and metal-casting sectors are traditionally low-digitization, small-to-medium enterprise-dominated, with slow capital investment cycles and high risk aversion to autonomous equipment; adoption of AI-driven hoists remains negligible.
Sector adoption velocity	claude-sonnet-5	2/5	Foundry and metal/plastic manufacturing is a low-digitization, physically intensive sector with slow, capital-heavy automation adoption compared to information-sector AI adoption patterns.
Augmentation potential	claude-haiku-4-5-20251001	2/5	Limited augmentation potential exists: AI could assist with load-weight sensing or collision avoidance alerts, but the operator remains fully responsible for hoist control and positioning, so productivity gains are marginal.
Augmentation potential	claude-sonnet-5	2/5	Some automated hoist/crane controls and sensors can assist operators with precision and safety, but this offers modest rather than transformative productivity gains for this specific physical task.
Task automatability	claude-haiku-4-5-20251001	1/5	Operating hoists requires real-time spatial perception, load balancing, and precise physical positioning in a dynamic factory environment. Current AI systems lack the embodied control and sensorimotor feedback necessary to safely and reliably operate industrial hoists end-to-end.
Task automatability	claude-sonnet-5	2/5	Physical positioning of heavy dies/patterns via hoist requires perception, judgment, and manipulation in a variable physical environment that current general AI systems cannot perform end-to-end.'
Adoption barriers	claude-haiku-4-5-20251001	5/5	Heavy machinery operation in foundries is tightly regulated by OSHA and industry safety standards; liability for autonomous hoist failure (dropped dies, worker injury) creates substantial legal and organizational barriers to automation without licensed human oversight.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement specifically for this task, but safety regulations, liability for heavy equipment operation, and workplace safety oversight create meaningful friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	The hardware cost of an autonomous hoist system (sensors, controls, safety certification) far exceeds the loaded wage of a human hoist operator, and integration costs remain prohibitively high for this specialized task.
Cost vs. human wage	claude-sonnet-5	1/5	Robotic hoist automation for this task would require custom industrial engineering with high capital costs, making it more expensive than a human operator in most current foundry settings.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed AI products reliably operate physical hoists in production foundries today. This task requires integrated robotics, vision, and force control that exists only in narrow research settings, not commercial foundry operations.
Technical feasibility today	claude-sonnet-5	1/5	No deployed off-the-shelf product autonomously operates foundry hoists to position dies today; this remains a manual or semi-automated crane/hoist operation task requiring human control.

Skim or pour dross, slag, or impurities from molten metal, using ladles, rakes, hoes, spatulas, or spoons.

CI 0–19 · exposure 8 · augmentation 13 · importance 3.1/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Foundries and metal-casting operations are capital-intensive, legacy-heavy industries with slow digitization. Dross skimming remains a hands-on craft skill; adoption of even semi-automated dross removal systems is nascent, and AI/robotic adoption in foundries lags well behind information and finance sectors.
Sector adoption velocity	claude-sonnet-5	1/5	Metal casting and foundry work is a low-digitization, physically demanding sector with slow AI/robotics adoption; this specific task is not part of typical automation rollouts.
Augmentation potential	claude-haiku-4-5-20251001	2/5	Current AI tools offer minimal assistance for this task; computer vision could potentially flag molten-metal temperature anomalies or alert to unsafe conditions, but the core sensorimotor and judgment work remains unaugmented by available AI systems.
Augmentation potential	claude-sonnet-5	1/5	AI offers essentially no direct assistance to a human physically skimming molten metal; sensor-based monitoring might inform timing but doesn't materially change the manual task execution.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires real-time sensorimotor control in a hazardous environment with molten metal (typically 1000°C+), dynamic visual feedback to identify dross location and quality, and manual tool manipulation. Current AI systems lack embodied robotic capabilities to reliably and safely perform this operation at parity with human workers.
Task automatability	claude-sonnet-5	2/5	This is a physical manipulation task requiring dexterity and heat-tolerant handling of molten metal; current AI (software) cannot perform this, and robotics for this specific task are not off-the-shelf solutions.5% of time savings via AI assistance would be minimal without dedicated robotic hardware.
Adoption barriers	claude-haiku-4-5-20251001	5/5	This task is governed by OSHA regulations and industry safety standards that place strict liability on the employer and equipment operator for thermal and chemical hazards. The legal and insurance requirements for autonomous operation of molten-metal handling near workers create hard regulatory barriers to unattended automation.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement for this specific act, but safety regulations, extreme heat/hazard conditions, and quality-control liability create practical barriers to full automation without specialized engineered equipment.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	Specialized robotic systems capable of high-temperature molten metal handling (protective enclosures, precision thermal controls, safety interlocks) are extremely expensive; the amortized cost per dross-skimming task would exceed the loaded wage of a foundry worker by a significant margin.
Cost vs. human wage	claude-sonnet-5	1/5	There is no general AI system performing this physical task, so any hypothetical robotic solution would require expensive custom engineering, likely costing more than a human operator for this task.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed commercial product reliably performs manual dross skimming from molten metal at production scale. While foundry automation exists for other casting tasks, the heterogeneous nature of impurities and the need for tactile/visual judgment to avoid re-entrainment of cleaned metal remain manual operations in practice.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI/robotic product reliably skims dross or slag from molten metal in production settings; this remains a manual or specialized hard-automation task, not an AI product category.

Repair or replace damaged molds, pipes, belts, chains, or other equipment, using hand tools, hand-powered presses, or jib cranes.

CI 5–10 · exposure 0 · augmentation 25 · importance 3.7/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Manufacturing of metal and plastic components remains heavily physical and labor-intensive; digital transformation lags information-sector adoption, and robotics deployment in small-to-medium foundries is still minimal.
Sector adoption velocity	claude-sonnet-5	1/5	Manufacturing maintenance and machine repair is a low-digitization, physical-labor sector with minimal AI/robotic penetration into hands-on repair tasks.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI could marginally assist with diagnostic recommendations or equipment documentation, but the core task—hands-on repair with tools and cranes—offers limited scope for meaningful AI augmentation while the human remains primary.
Augmentation potential	claude-sonnet-5	2/5	AI could assist with diagnostics (e.g., predictive maintenance alerts, repair manuals/documentation lookup) but offers little direct help with the physical repair actions themselves.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires physical manipulation of heavy equipment in a factory setting—removing damaged parts, positioning replacements, and operating hand tools and jib cranes. Current AI systems cannot perform physical work in unstructured industrial environments, making end-to-end automation infeasible.
Task automatability	claude-sonnet-5	1/5	This is a physical maintenance/repair task requiring manual dexterity, diagnosis, and use of hand tools and cranes on heavy industrial equipment; no current AI system can perform this physically end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Safety regulations, equipment liability (incorrect repair could cause equipment failure or injury), and the requirement that a qualified technician validate repair work create strong legal and organizational barriers to full automation.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement typically applies, but physical plant safety protocols, equipment liability, and the need for skilled hands-on judgment create real organizational friction against any automation attempt.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	The equipment needed for autonomous physical repair—mobile manipulation, sophisticated sensors, and on-site integration—vastly exceeds the loaded cost of a skilled trade technician's wage, making AI substantially more expensive today.
Cost vs. human wage	claude-sonnet-5	1/5	There is no viable AI/robotic substitute performing this physical repair work, so the cost comparison favors the human by default since the AI alternative doesn't exist in deployable form.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed AI products can autonomously repair or replace molds and equipment using hand tools and cranes in a production facility. The task demands embodied dexterity and on-site troubleshooting that remains in the research or early prototype phase.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product performs physical mold/equipment repair; robotics for such unstructured maintenance work remains research-stage or highly bespoke, not commercially deployed at scale.

Related occupations — Production

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.