Separating, Filtering, Clarifying, Precipitating, and Still Machine Setters, Operators, and Tenders
51-9012.00Median wage $51,610/yr60,100 employed (US)Rank #333 of 923 scored · top 36% by substitution
Set up, operate, or tend continuous flow or vat-type equipment; filter presses; shaker screens; centrifuges; condenser tubes; precipitating, fermenting, or evaporating tanks; scrubbing towers; or batch stills. These machines extract, sort, or separate liquids, gases, or solids from other materials to recover a refined product. Includes dairy processing equipment operators.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution33
Exposure29
Augmentation43
Substitution — the headline: capability discounted by cost, barriers and adoption.
Exposure — technical capability alone, regardless of whether anyone deploys it.
Augmentation — how much AI assists without replacing. High here + moderate substitution = a changing job, not a disappearing one.
Tasks on the substitution scale
20 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
5%
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%30
panel mean rating 2.2/5 → substitution pressure 30/100
Technical feasibility todayw 20%27
panel mean rating 2.1/5 → substitution pressure 27/100
Cost vs. human wagew 15%29
panel mean rating 2.2/5 → substitution pressure 29/100
Adoption barriersw 20%inverted — strong barriers lower the score50
panel mean rating 3.0/5 (barrier strength) → substitution pressure 50/100
Sector adoption velocityw 10%24
panel mean rating 2.0/5 → substitution pressure 24/100
Task breakdown (20 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.
Maintain logs of instrument readings, test results, or shift production for entry in computer databases.
72CI 65–79 · exposure 70 · augmentation 63 · importance 3.8/5 · click for rater detail
Maintain logs of instrument readings, test results, or shift production for entry in computer databases.
72| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Manufacturing and chemical processing sectors with heavy instrumentation (the core audience for this role) are actively deploying sensor-to-database automation and SCADA upgrades; adoption is measurable and accelerating in digitized plants. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing and process industries adopt digitization more slowly than information/finance sectors; many smaller or older facilities still use manual logs and paper-based systems. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by auto-populating logs, flagging anomalies, and summarizing trends, reducing manual transcription burden; operators retain oversight of data quality and alarm interpretation. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI/automated systems can pull readings directly into databases, flag anomalies, and reduce transcription errors, significantly aiding operators who remain responsible for oversight and interpretation. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Log maintenance and data entry from instrument readings are highly structured, repetitive tasks. Current OCR, time-series parsing, and database APIs can automate ~80% of the workflow—reading gauges, timestamping, and inserting records—though occasional anomaly review or manual correction may remain. |
| Task automatability | claude-sonnet-5 | 4/5 | Recording instrument readings, test results, and production data into databases is a structured, repetitive data-entry task well within current automation capability, especially where sensors already generate digital signals. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory or licensing barriers apply to automated log maintenance itself; the main friction is legacy system integration and operator training. No requirement for a licensed human to physically perform the recording. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for logging data; some industries require certified human sign-off on certain quality/compliance records, creating minor friction but not a hard barrier. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated logging costs pennies per entry once infrastructure is in place (sensors, APIs, database), while manual data entry at loaded wage (including benefits, overhead) costs dollars per entry or higher, giving >10× cost advantage to automation. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated data logging via sensors and database integration is dramatically cheaper per data point than manual logging by an operator, once the sensor/integration infrastructure exists. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed products for automated industrial log capture (SCADA integration, ERP data ingestion, computer vision for gauge reading) exist and operate reliably in manufacturing settings; the gap is mainly in edge cases and anomaly handling rather than core capability. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Many plants already use SCADA/historian systems and IoT sensors that auto-log data, but many facilities still rely on manual logbooks or manual transcription of readings, especially for older equipment, so full deployment varies by plant. |
Measure or weigh materials to be refined, mixed, transferred, stored, or otherwise processed.
64CI 52–75 · exposure 62 · augmentation 50 · importance 4.0/5 · click for rater detail
Measure or weigh materials to be refined, mixed, transferred, stored, or otherwise processed.
64| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Pharmaceutical, chemical, food processing, and petrochemical industries—where this task is central—have rapidly adopted automated weighing and measurement systems as part of broader process automation and Industry 4.0 initiatives. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Process manufacturing and industrial sectors adopt automation more slowly than information/professional services, with capital-intensive retrofits and legacy equipment slowing deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | Automated measurement systems assist operators by providing real-time data visualization, alerts for out-of-spec batches, and historical trend analysis, improving decision-making on material quality and process adjustments without removing the human from oversight. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Digital scales, sensor readouts, and automated tracking systems assist operators by improving measurement accuracy and reducing manual calculation, though the physical handling aspect still requires human presence. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Modern vision systems and weight sensors can reliably measure and weigh materials with minimal human oversight, achieving significant time savings over manual measurement. Automation of this routine procedural task integrates readily into existing manufacturing workflows, though some setup and calibration oversight remains necessary. |
| Task automatability | claude-sonnet-5 | 3/5 | Measuring and weighing materials is highly routine and can be handled by automated sensors, scales, and PLC-controlled dosing systems, but full task automation requires physical integration with existing plant equipment which varies by facility. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory or legal barriers prevent automation; measurement systems are industry-standard in manufacturing. Some workplace safety protocols and quality oversight require human involvement, creating modest friction rather than hard prohibition. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for this specific measuring task, though safety protocols and quality control sign-offs in regulated industries (chemical, pharma) create some procedural friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Automated measurement and weighing systems cost thousands to tens of thousands annually but operate continuously across many cycles, whereas a single operator represents significant loaded wage costs per shift. The cost per measurement is substantially lower for automated systems. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Industrial-grade sensors and automation hardware require significant capital investment and integration costs, so while cheaper per-unit over time, upfront costs make the ratio only moderately favorable versus a machine operator's wage. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Weight sensors, automated measurement systems, and computer vision for material assessment are mature, deployed technologies in manufacturing and chemical processing environments. Production systems reliably perform this task, though some variability in material properties may require periodic human verification. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Automated weighing/metering systems (load cells, flow meters, batching controllers) are mature and widely deployed in process industries, though many facilities still rely on manual measurement steps for less standardized operations. |
Monitor material flow or instruments, such as temperature or pressure gauges, indicators, or meters, to ensure optimal processing conditions.
60CI 48–72 · exposure 62 · augmentation 75 · importance 4.3/5 · click for rater detail
Monitor material flow or instruments, such as temperature or pressure gauges, indicators, or meters, to ensure optimal processing conditions.
60| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Adoption is uneven: large petrochemical and pharmaceutical firms have deployed monitoring automation, but smaller manufacturers and batch facilities lag. Pilots are common; full autonomous operation (without humans in the loop) remains uncommon due to liability concerns. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Process manufacturing is a moderately digitized but physically-oriented sector; automation adoption is steady but slower than in information/professional services. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI dashboards and alert systems significantly amplify human operator productivity by surfacing anomalies faster than manual gauge-reading and enabling predictive action. The human operator remains central but can oversee more equipment with greater confidence. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-based predictive analytics and automated alerts substantially help operators monitor multiple parameters and catch deviations faster, even though a human typically remains responsible for response and judgment. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI-based monitoring systems with computer vision and sensor integration can continuously track gauges and instruments, detect deviations, and alert operators in real-time. This achieves significant time savings (>50%) on routine monitoring tasks, though human judgment on novel anomalies or complex failures still adds value. |
| Task automatability | claude-sonnet-5 | 3/5 | Monitoring gauges and instruments for process control is well-suited to sensor-based automation and SCADA/DCS systems with alarms and analytics, though physical sensor installation and edge-case judgment remain human-dependent in many plants.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Industrial facilities typically require human operators on-site for safety sign-off and intervention in emergencies. Regulatory frameworks (EPA, OSHA) often mandate human responsibility, creating friction but not an outright ban on AI monitoring as a support tool. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for automated monitoring, but safety-critical processes often require human oversight or sign-off for anomalies, creating moderate organizational and liability friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI monitoring systems have high upfront costs but very low per-task inference costs once deployed; amortized across many monitoring instances per day, they cost a fraction of continuous human operator oversight. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Sensor and control system infrastructure has significant upfront capital and integration cost, though once installed, ongoing monitoring is cheap per unit of output compared to continuous human observation. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed industrial IoT platforms and predictive maintenance systems (e.g., those from Siemens, GE, Honeywell) already perform automated monitoring of temperature, pressure, and flow in chemical plants and refineries. These are in production at scale, though integration complexity and site-specific calibration remain common. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Industrial control systems and IoT-based monitoring platforms are deployed in many process plants, but full automation with AI-driven anomaly detection is unevenly adopted across smaller or older facilities. |
Communicate processing instructions to other workers.
46CI 30–62 · exposure 45 · augmentation 63 · importance 3.9/5 · click for rater detail
Communicate processing instructions to other workers.
46| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Manufacturing and industrial processing are moderately digitizing; some plants use automated systems (displays, digital work orders) but human-to-human instruction remains standard and adoption of fully autonomous instruction agents is still in pilot phases. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing and process industries adopt AI more slowly than information-sector work, with digitization of floor-level communication still nascent.dummy |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can assist by drafting templates, translating instructions into multiple languages, or highlighting critical steps, enabling workers to communicate faster and more clearly while retaining final approval and real-time adaptation. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-enabled communication tools (translation, transcription, standardized instruction templates, alerts) can meaningfully support clarity and consistency of instructions even though a human remains central to conveying them.dummy |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI can generate and distribute processing instructions written from technical documentation or input parameters, achieving significant time savings. However, nuanced real-time adaptation to worker questions or unexpected conditions may require human oversight, preventing a full 5-rating. |
| Task automatability | claude-sonnet-5 | 2/5 | Communication of processing instructions is often verbal, context-dependent, and tied to real-time floor conditions, making full end-to-end automation impractical though message drafting or logging could be partially assisted.dummy |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Some friction exists due to worker preference for human-delivered instructions, safety liability concerns if instructions are misunderstood, and regulatory emphasis on clear chain-of-command communication in industrial settings. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement blocks AI here, but safety-critical plant environments create organizational friction and liability concerns around delegating instruction-giving to automated systems.dummy |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Automated instruction generation via AI is substantially cheaper than allocating skilled worker time to repetitive communication tasks, with minimal infrastructure cost once a system is in place. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Replacing this interpersonal, situational communication with AI would require sensors, integration, and oversight infrastructure that likely costs more than the marginal human communication effort involved.dummy |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Current AI systems can draft and send written instructions reliably, but real-world factory/plant communication often requires adaptive dialogue, accent/noise handling, or contextual judgment that deployed systems handle inconsistently at production scale. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | No deployed product autonomously communicates operational instructions between machine operators on a plant floor today; existing tools are limited to messaging or documentation support, not judgment-driven instruction relay.dummy |
Remove full containers from discharge outlets and replace them with empty containers.
44CI 15–72 · exposure 38 · augmentation 13 · importance 3.3/5 · click for rater detail
Remove full containers from discharge outlets and replace them with empty containers.
44| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Chemical, pharmaceutical, and beverage manufacturing have moderate to strong automation adoption, but many smaller facilities and batch operations still rely on manual container swaps. Adoption is growing but unevenly distributed across sectors and facility sizes. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Manufacturing/process industries with physical material handling are among the slowest sectors for AI-driven adoption, with automation here typically achieved via traditional industrial engineering rather than AI advances. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | This task offers minimal augmentation opportunity because it is almost entirely manual material handling with no judgment, decision-making, or cognitive elements that AI could assist. Full automation is the natural outcome rather than AI-assisted performance. |
| Augmentation potential | claude-sonnet-5 | 1/5 | AI offers essentially no direct assistance to a human physically removing and replacing containers; this is a manual task outside the scope of current AI augmentation tools. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | This task involves physically removing and replacing containers at fixed discharge points—a well-defined, repeatable mechanical operation. Current robotic systems can reliably perform container removal and replacement at industrial scales, with integration into existing machinery infrastructure. While some environmental variability (container positioning, pressure, thermal conditions) presents minor challenges, the core operation meets near-total time savings with deployed systems. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical material-handling task requiring perception, mobility, and manipulation in a plant environment; no general-purpose AI system can perform this end-to-end today. Robotic solutions exist only in narrow, highly engineered contexts, not as off-the-shelf AI. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory or licensing barriers exist for automating routine container handling; no licensed professional must sign off. Primary friction is capital investment and workplace safety integration (guarding, lockout-tagout procedures), which are surmountable organizational concerns rather than hard legal barriers. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates a human perform this specific task, but physical workplace safety standards and equipment certification requirements create some friction for automation retrofits. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Robotic systems for container handling typically cost $50k–$150k upfront with modest per-operation costs, amortized over thousands of cycles, making per-cycle cost 1–5× cheaper than loaded labor ($30–$50/hour including overhead). At sufficient throughput, the ratio strongly favors automation. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Deploying robotic arms, conveyors, or AGVs to replicate this task would require significant capital investment in custom automation, making it costlier than a human operator for most operations at current scale. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Robotic arms and automated bin/container management systems are deployed in production environments (breweries, chemical plants, pharmaceutical facilities) for precisely this task. Products like collaborative robots and conveyor-integrated systems reliably handle container swaps with minimal error, though integration complexity varies by installation. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed 'AI' product performs this generic physical swap task reliably across varied industrial settings; any solution would require custom robotics/automation engineering, not current AI systems. |
Examine samples to verify qualities such as clarity, cleanliness, consistency, dryness, or texture.
41CI 30–52 · exposure 38 · augmentation 50 · importance 4.2/5 · click for rater detail
Examine samples to verify qualities such as clarity, cleanliness, consistency, dryness, or texture.
41| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption is uneven and sector-dependent. Large pharmaceutical and chemical firms use automated inspection; smaller and mid-market operations rely heavily on manual inspection. Overall adoption remains moderate, with pilots common but full displacement rare outside capital-intensive settings. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing and process industries adopt automation more slowly than information sectors; sensor-based quality control is spreading but this occupation reflects a laggard, physically-oriented industrial sector. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-powered imaging and sensor dashboards assist operators by flagging anomalies and standardizing measurements, improving consistency and speed. The human remains essential for final judgment on borderline samples and for integrating cross-sensory cues (texture, appearance, context). |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-powered vision and sensor analytics can flag anomalies and support human judgment on sample quality, improving consistency and speed of inspection while the human remains responsible for final verification. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | Sampling and visual/tactile quality verification can be partially automated with machine vision systems and sensors, but real-world samples in chemical/pharmaceutical environments often require nuanced sensory judgment (texture, subtle clarity variations) that current AI struggles with reliably. Automation could handle roughly 40–60% of routine inspections. |
| Task automatability | claude-sonnet-5 | 2/5 | Visual/tactile sample inspection requires physical presence and sensor/robotic integration that off-the-shelf AI cannot yet fully replicate for diverse industrial process streams, though machine vision can assist with some visual checks like clarity or color consistency., dryness and texture largely require physical sensing. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | No legal requirement mandates human sign-off, but quality standards (GMP, ISO) and liability for defects create organizational friction. Sample examination directly affects product release, introducing error-cost asymmetry that encourages human oversight and validation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but quality control decisions affecting product safety or specifications may carry liability concerns, and physical sampling requires proximity to equipment, creating moderate friction against pure automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | Machine vision and sensor systems have capital and integration costs comparable to a technician's annual wage, but per-sample operational cost becomes competitive at scale. Overall cost parity is rough, depending on task volume and error tolerance. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Deploying specialized sensors, cameras, and integration for physical quality checks is capital-intensive relative to a machine operator visually/tactilely checking samples as part of routine duties. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Computer vision systems and spectroscopy tools exist in production labs, but they typically handle narrow, controlled applications (e.g., color or density matching). Broader quality verification across clarity, consistency, and texture requires human-in-the-loop validation in most industrial settings today. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Machine vision systems exist for quality inspection in some manufacturing lines, but generalized deployment for clarity, dryness, and texture verification across separating/filtering processes is narrow and not broadly production-proven. |
Test samples to determine viscosity, acidity, specific gravity, or degree of concentration, using test equipment such as viscometers, pH meters, or hydrometers.
33CI 30–35 · exposure 25 · augmentation 50 · importance 4.0/5 · click for rater detail
Test samples to determine viscosity, acidity, specific gravity, or degree of concentration, using test equipment such as viscometers, pH meters, or hydrometers.
33| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption is limited; most manufacturing and chemical firms still rely on human technicians for routine sample testing. A few large-scale operations have invested in automated lab systems, but widespread displacement remains slow due to the mixed technical and regulatory friction. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing and industrial processing sectors show slower, more capital-intensive adoption patterns for automation compared to information-based industries, with sensor-based monitoring being a mature but not universally deployed technology. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-driven data logging, automated report generation, and anomaly detection on historical test results can meaningfully assist operators in tracking samples and spotting trends. However, the core manipulation and initial measurement still requires human presence, limiting overall productivity lift. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Digital test equipment with automated data logging and AI-assisted anomaly detection can help operators interpret readings and flag out-of-spec conditions, improving efficiency in the testing and reporting process. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can interpret readings from digital test equipment (e.g., pH meters or viscometers that output numerical values), the task requires hands-on sample preparation, placement into equipment, and environmental monitoring. Current AI systems cannot reliably handle the physical manipulation and contextual judgment needed for consistent sample preparation and equipment operation end-to-end. |
| Task automatability | claude-sonnet-5 | 2/5 | While automated inline sensors can measure viscosity, pH, and specific gravity, the task as described involves manual sampling and handheld test equipment operation, which requires physical presence and dexterity that current AI cannot fully replace end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Chemical and manufacturing regulations often require documented operator oversight and chain-of-custody for test results, and some quality assurance standards mandate that a qualified human review and approve measurements. However, these are process-level controls rather than hard legal prohibitions on automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No formal licensing requirement typically applies, but quality control and safety protocols in industrial settings often require documented human verification of test results, creating moderate procedural friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Physical robotic systems capable of handling samples and operating lab equipment are expensive (significant capex), while a technician's hourly wage for this routine testing is modest. The integration, maintenance, and oversight costs of automation currently exceed the labor cost savings in most settings. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Retrofitting older separating/filtering equipment with automated sensors and inline testing systems requires significant capital investment, often exceeding the marginal cost of a machine operator performing periodic manual checks in smaller-scale operations. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some industrial analytics platforms can log and interpret sensor data from laboratory equipment, but no deployed system reliably performs the full cycle of sample handling, equipment operation, and quality judgment autonomously. Robotic arms with vision exist in research labs but are not mature, production-ready solutions for routine viscosity or pH testing. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Automated process analytical technology (PAT) systems exist in some modern plants for continuous inline monitoring, but widespread deployment of fully autonomous sampling-to-testing systems replacing manual operators is not yet standard in this occupation. |
Turn valves to pump sterilizing solutions or rinse water through pipes or equipment or to spray vats with atomizers.
33CI 30–35 · exposure 25 · augmentation 25 · importance 3.8/5 · click for rater detail
Turn valves to pump sterilizing solutions or rinse water through pipes or equipment or to spray vats with atomizers.
33| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of robotic valve control in industrial process plants is slow outside large pharma and food manufacturing; most chemical and sterilization operations remain labor-dependent with incremental automation only. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing and process industries adopt automation steadily but slowly compared to information sectors; CIP automation is common in large-scale food/pharma but adoption is uneven across smaller operators. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | Current AI offers minimal augmentation for real-time valve and spray control; sensors and alerts could assist monitoring, but the core manual dexterity and immediate sensory feedback tasks remain poorly supported by existing AI tools. |
| Augmentation potential | claude-sonnet-5 | 2/5 | Sensors and monitoring dashboards can alert operators to issues, offering some assistance, but the core physical valve operation itself is not meaningfully augmented by current AI tools. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Valve operation is nominally simple, but requires real-time physical manipulation in a piped system with sensory feedback—detecting proper flow, pressure, and spray patterns. Current robotic systems struggle with the coordination and real-world sensing needed for reliable, safe operation without on-site engineering. |
| Task automatability | claude-sonnet-5 | 2/5 | The physical valve-turning and equipment cleaning cycle requires actuation in a physical environment, which current general-purpose AI cannot perform; automation here means industrial control/robotics, not AI software alone, so the software-only time-saving is minimal., though PLC/SCADA systems already automate parts of this in some plants.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | No strict licensing barrier exists for this operator role, but workplace safety regulations, equipment-specific training, and the need for on-site human oversight of sterilization efficacy create moderate friction against full automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for this task, but safety regulations around sterilization processes and equipment certification create moderate procedural barriers to unsupervised automation changes. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | A deployed robotic system (hardware + integration + maintenance) to handle valve turning and spray atomization would exceed the cost of a shift operator performing these tasks, especially in smaller or medium-scale operations. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Retrofitting a plant with automated CIP/valve control systems requires significant capital investment in sensors and actuators; while cheaper long-term in high-volume plants, it is not universally cheaper than human labor especially in smaller operations. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Industrial automation exists for some sterilization cycles, but end-to-end valve manipulation and flow control in production environments remains largely manual or requires bespoke industrial PLC integration rather than off-the-shelf AI deployment. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Automated CIP (clean-in-place) systems exist and are deployed in some food/chemical plants, but many facilities still rely on manual valve operation and human tending, so this is not uniformly reliable AI-driven automation. |
Dump, pour, or load specified amounts of refined or unrefined materials into equipment or containers for further processing or storage.
30CI 30–30 · exposure 25 · augmentation 25 · importance 4.5/5 · click for rater detail
Dump, pour, or load specified amounts of refined or unrefined materials into equipment or containers for further processing or storage.
30| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Manufacturing and chemical processing sectors are adopting automated material handling, but adoption remains uneven and often confined to large, modernized facilities with standardized processes. Smaller plants and those handling diverse or hazardous materials lag significantly. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing and materials processing sectors adopt automation more slowly than information/professional services, with physical material handling tasks lagging due to capital costs and facility-specific engineering needs. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI and vision systems can assist in measuring or monitoring material quantities and flagging out-of-spec loads, but the core task is fundamentally physical manipulation. Augmentation value is limited compared to full automation potential, and the human remains the bottleneck for actual loading operations. |
| Augmentation potential | claude-sonnet-5 | 2/5 | Sensors and monitoring systems can assist operators with precise measurement feedback and alerts, but this offers only modest productivity gains for what is fundamentally a physical loading task. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | This task involves physical dumping, pouring, or loading of materials into equipment—operations that require robotic or automated systems, not just AI. While current robotics can handle some uniform dumping tasks, the requirement to measure 'specified amounts' and handle 'refined or unrefined materials' (which vary in properties) makes reliable end-to-end automation difficult without extensive task-specific engineering and vision systems. |
| Task automatability | claude-sonnet-5 | 2/5 | This involves physical material handling with specific quantities requiring perception and manipulation in industrial settings; while robotic dosing/dispensing systems exist, this is not a general AI (software) task and requires specialized hardware integration, not off-the-shelf automation.rialized industrial robotics setups. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Moderate barriers exist: equipment compatibility, safety standards (especially around hazardous materials), and process validation slow adoption. However, no hard licensing requirement forces human sign-off, and liability is manageable in many industrial contexts, reducing barriers below hard regulatory ones. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing barrier typically exists, but safety regulations, equipment certification requirements, and liability concerns around handling refined/unrefined materials (potentially hazardous) create moderate friction against wholesale automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Industrial robotics and vision systems for material handling are capital-intensive and require significant integration costs. For routine, low-skill manual dumping/loading at typical wages, upfront and maintenance costs often exceed the loaded wage of the human operator, especially in smaller or legacy facilities. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Industrial automation for material dumping/loading requires significant capital investment in specialized machinery, sensors, and integration, often exceeding the cost of a human operator for lower-volume or variable operations. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Deployed robotic systems exist for some material handling, but most production deployments remain task-specific and don't generalize to arbitrary refined or unrefined materials. Error rates in volume estimation and spillage remain material, and integration with existing equipment varies widely across industrial settings. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Automated dosing/filling systems exist in some processing plants but are narrow, purpose-built industrial equipment rather than flexible AI systems, and many facilities still rely on manual operators for this step. |
Turn valves or move controls to admit, drain, separate, filter, clarify, mix, or transfer materials.
28CI 25–30 · exposure 25 · augmentation 50 · importance 4.3/5 · click for rater detail
Turn valves or move controls to admit, drain, separate, filter, clarify, mix, or transfer materials.
28| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | While large chemical and petrochemical plants have invested in automation and SCADA systems, the broader manufacturing sector (food processing, pharmaceuticals, smaller operations) lags significantly; AI-driven autonomous agents are rarely in production for this task even in advanced industries. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing and process industries adopt automation steadily but slowly compared to information sectors, with capital-intensive upgrade cycles limiting the pace of new AI-driven control adoption. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | Real-time monitoring dashboards, predictive alerts for valve maintenance, and AI-assisted diagnostics of material flow anomalies can assist operators in decision-making, but the physical control action itself remains operator-driven; AI provides useful augmentation on the diagnostic and planning side. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-based predictive analytics and anomaly detection can assist operators in monitoring and optimizing valve/control settings, improving efficiency while the human remains responsible for physical adjustments and oversight. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Modern industrial systems can automate some valve operation via programmable logic controllers and sensors, but the task requires physical manipulation in industrial environments with real-time monitoring and exception-handling (leaks, pressure anomalies, material properties changing). Current AI agents cannot reliably perform end-to-end physical control and sensory feedback integration at 50% time savings across diverse industrial setups. |
| Task automatability | claude-sonnet-5 | 2/5 | This requires physical manipulation of valves and controls in an industrial plant setting, which current AI (software-based) cannot perform without robotic embodiment; only the decision-logic portion could be automated via existing SCADA/PLC systems, which are not 'AI' per se. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Industrial facilities face significant regulatory oversight (EPA, OSHA, ASME pressure vessel codes), operator certification requirements, and liability concerns around fluid handling, pressure control, and environmental compliance that legally mandate human supervision and sign-off, creating strong adoption barriers. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement for the operator role itself, but safety regulations, liability for equipment failure/spills, and the physical nature of intervention create meaningful organizational and safety barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Industrial automation and process control systems often cost millions and require extensive integration and maintenance; the operational cost of such systems for this task typically exceeds the loaded wage of a single equipment operator, especially for smaller or less standardized operations. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Retrofitting older plants with sensors, actuators, and control systems to replace human operators involves substantial capital investment that may exceed labor costs unless already automated, especially for smaller or older facilities. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Deployed industrial automation systems handle routine valve operations in narrow, controlled settings, but general-purpose automation of this task across variable material types, equipment configurations, and failure modes remains limited to specialized, custom installations rather than off-the-shelf AI systems that perform reliably at scale. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Distributed control systems and PLCs have automated valve control in process industries for decades, but these are traditional automation, not AI-driven perception/decision systems, and fully autonomous AI-based operation of legacy or variable plants is not widely deployed. |
Set up or adjust machine controls to regulate conditions such as material flow, temperature, or pressure.
28CI 25–30 · exposure 25 · augmentation 50 · importance 4.3/5 · click for rater detail
Set up or adjust machine controls to regulate conditions such as material flow, temperature, or pressure.
28| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Industrial manufacturing sectors are moving toward automation, but adoption of AI-driven machine setup is still in pilot and early adoption phases in most operations. Legacy equipment dominance and the need for human expertise in troubleshooting slow widespread deployment. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing and process industries adopt automation more slowly than white-collar sectors, with legacy equipment and capital cycles limiting AI-driven control adoption. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist operators by monitoring sensor data, alerting to parameter deviations, and suggesting control adjustments, improving decision-making speed. However, the human operator remains essential for final adjustment execution and real-time problem-solving. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-based predictive analytics and sensor monitoring can help operators anticipate needed adjustments and flag anomalies, improving decision quality even if humans still execute the physical control changes. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Setting up and adjusting machine controls requires real-time physical interaction with equipment, sensory assessment of conditions, and situational judgment. While AI can monitor and suggest adjustments, current systems cannot reliably perform end-to-end setup and hands-on control calibration without human oversight and physical intervention. |
| Task automatability | claude-sonnet-5 | 2/5 | This requires physical manipulation of machine controls and real-time sensory judgment on a shop floor, which current AI cannot perform end-to-end without robotic embodiment.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Process safety regulations, equipment liability, and manufacturing compliance standards require licensed operators or engineers to sign off on critical setup and calibration tasks. Union agreements and operator certification also create legal and contractual barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement typically applies, but safety, quality control, and liability concerns around process deviations create organizational friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Integration of AI monitoring and control systems is capital-intensive, and the cost of sensors, network infrastructure, and system configuration often exceeds the wage savings from one operator. Maintenance and oversight costs further erode the cost advantage. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Retrofitting sensors, control systems, and AI-driven regulation requires significant capital investment, and human operators remain cost-competitive for adjustment tasks in many plants. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some industrial control systems have partial automation (PID controllers, SCADA integration), but full setup and adjustment of separation/filtration machines by AI alone is not deployed at scale. Existing systems require human expertise to handle edge cases, equipment variability, and safety-critical configurations. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | While industrial control systems and PLC-based automation exist, fully autonomous AI-driven setup/adjustment of these machines in production is narrow and often still requires human tuning and oversight. |
Start agitators, shakers, conveyors, pumps, or centrifuge machines.
28CI 25–30 · exposure 25 · augmentation 38 · importance 4.1/5 · click for rater detail
Start agitators, shakers, conveyors, pumps, or centrifuge machines.
28| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Manufacturing sectors have slowly adopted industrial automation, but most chemical plants and processing facilities still employ human operators for hands-on equipment control. Autonomous machine startup remains a niche adoption in only the most advanced, greenfield facilities. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing and processing sectors involving physical equipment operation are slower adopters of AI/automation compared to information-based sectors, though some large-scale industrial automation exists. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could assist by monitoring equipment readiness and alerting operators when conditions are right to start machines, or by automating routine startup sequences through control system integration. This augmentation helps operators work more efficiently but does not eliminate the human role in verification and oversight. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI-based monitoring and predictive systems can inform operators when to start equipment or flag optimal timing, but this is a discrete physical action with limited scope for AI augmentation beyond basic scheduling assistance. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Starting machines is a discrete, physical action that current AI agents cannot perform without robotics integration. While the decision logic to start machines could be automated, the actual initiation requires physical manipulation or control system access that is not yet reliably deployed at scale in chemical/manufacturing settings. |
| Task automatability | claude-sonnet-5 | 2/5 | This is a simple physical machine-starting action requiring physical presence and interaction with equipment; current AI cannot physically operate these controls without robotic embodiment, though software could trigger start sequences in fully automated plants.atable only via existing automation, not general AI.=2 |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Heavy industrial equipment operation is often subject to safety regulations, operator licensing, and liability requirements. OSHA and industry standards typically mandate that a qualified human operator oversee startup procedures and be responsible for equipment state, creating legal and safety barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement for this specific action, but safety protocols, equipment liability, and the need for human oversight during hazardous machine starts create moderate organizational friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Installing robotic arms, sensors, and integrating AI control systems to replace a human pressing a button or toggle is substantially more expensive than the operator's wage for this simple task. The capital and integration cost far exceeds the labor savings from automating a single, quick action. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Retrofitting industrial control automation into existing plants requires significant capital investment in sensors, actuators, and integration, which may exceed the cost of a human operator especially in smaller facilities. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some industrial control systems allow remote or automated startup, but most chemical/manufacturing floors still rely on human operators to physically start equipment or use legacy control interfaces. No mainstream AI product reliably monitors conditions and autonomously initiates startup across diverse machine types in production environments. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some plants have automated control systems that start equipment on programmed schedules, but this reflects industrial automation/PLC systems rather than AI products, and generalized AI performing this physical task is not deployed at scale. |
Operate machines to process materials in compliance with applicable safety, energy, or environmental regulations.
25CI 25–25 · exposure 25 · augmentation 50 · importance 4.4/5 · click for rater detail
Operate machines to process materials in compliance with applicable safety, energy, or environmental regulations.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Heavy manufacturing and chemical processing sectors adopt automation cautiously due to safety and regulatory constraints; adoption has remained slow and focused on augmentation rather than replacement of human operators in production environments. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing and industrial processing sectors adopt automation more slowly than information-sector work, with physical plant upgrades requiring capital cycles and safety validation.' |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-powered monitoring dashboards, predictive alerts for equipment failures, and real-time regulatory checklist systems can meaningfully assist operators in maintaining compliance and optimizing processes while the human retains decision authority. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-based sensors, predictive maintenance, and monitoring dashboards can meaningfully assist operators in tracking compliance metrics and machine performance, though the human remains essential for physical operation. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While routine machine operation tasks can be partially automated through programmable logic controllers and monitoring systems, the requirement to ensure compliance with multiple, context-dependent regulatory frameworks and respond to real-time process variations requires human judgment and adaptation that current AI systems cannot reliably perform end-to-end. |
| Task automatability | claude-sonnet-5 | 2/5 | This is physical machine operation requiring on-site presence, manual controls, and regulatory compliance monitoring that current AI cannot perform end-to-end without robotic embodiment.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Safety regulations, environmental compliance requirements (EPA, OSHA), and liability frameworks typically mandate that a qualified human operator remains responsible for process outputs and regulatory adherence, creating a hard legal and liability barrier to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Safety, energy, and environmental regulations typically require human oversight, certified operators, and accountability for compliance failures, creating strong regulatory and liability barriers to full automation.' |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Integration of compliance-monitoring AI systems alongside existing machine control infrastructure is expensive relative to the wages of skilled operators; the cost of false-negative safety/environmental violations often exceeds the cost of human oversight. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Industrial automation retrofits require significant capital investment in sensors, control systems, and robotics, making near-term costs comparable to or higher than human operators for many facilities.' |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Deployed systems can monitor machines and flag deviations, but no current production-grade AI system independently operates these machines while guaranteeing regulatory compliance without human oversight; such systems remain primarily in proof-of-concept stages in specialized industrial contexts. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | While process control automation and sensor-based monitoring exist in industrial settings, fully autonomous machine operation with compliance judgment is not a mature deployed product across this occupation broadly.' |
Inspect machines or equipment for hazards, operating efficiency, malfunctions, wear, or leaks.
25CI 25–25 · exposure 25 · augmentation 63 · importance 4.0/5 · click for rater detail
Inspect machines or equipment for hazards, operating efficiency, malfunctions, wear, or leaks.
25| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Manufacturing and industrial process sectors show moderate digital adoption but remain conservative on autonomous inspection of safety-critical equipment; pilots are more common than full production deployment, particularly in smaller and mid-scale operations. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing and process industries are slower AI adopters relative to information-sector benchmarks; predictive maintenance sensors are spreading but full inspection automation remains uncommon. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | Computer vision and sensor-based alerts can assist operators by flagging anomalies or highlighting wear patterns for human review, but current systems rarely transform productivity enough to offset setup complexity; assistance is useful on specific, well-defined inspection subtasks. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-driven condition monitoring, vibration analysis, and thermal/visual anomaly detection significantly aid operators in prioritizing and identifying issues, even though a human typically still performs or verifies the physical inspection. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While visual inspection for obvious defects could be partially automated with computer vision, the task requires judgment about operating efficiency, subtle leaks, and contextual hazard assessment that demands understanding of process state and equipment-specific norms—difficult for current AI without extensive domain setup. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical inspection of industrial equipment for hazards, leaks, and wear requires sensory presence and manual checks that current AI systems cannot perform end-to-end without robotic hardware and sensor integration far beyond typical deployment.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Safety-critical inspection tasks often carry legal and regulatory requirements that a qualified operator must perform or sign off on; liability for missed defects, equipment damage, or workplace injury creates high error-cost asymmetry and organizational reluctance to remove human judgment from the loop. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Workplace safety regulations (e.g., OSHA) often require qualified personnel to inspect and certify equipment, and liability for missed hazards creates strong incentives to keep humans in the inspection loop. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current industrial computer vision systems, including hardware, integration, and ongoing calibration/oversight, remain expensive relative to straightforward machine operator wages, especially when factoring in false-positive costs and required human verification. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Sensor networks and monitoring software have upfront and maintenance costs that can approach or exceed the marginal cost of a technician's inspection round, especially for smaller or older installations lacking retrofit infrastructure. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Computer vision systems can detect some visual defects in controlled settings (research stage and early pilots exist), but reliable production deployment for complex industrial inspection across varied equipment, lighting, and contexts remains limited; most real-world use still requires human oversight. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | IoT sensor-based predictive maintenance and vision-based leak detection exist in some plants, but comprehensive machine-hazard inspection combining visual, auditory, and tactile cues is still narrow and supplementary rather than a replacement for human inspection. |
Collect samples of materials or products for laboratory analysis.
24CI 14–35 · exposure 20 · augmentation 38 · importance 4.0/5 · click for rater detail
Collect samples of materials or products for laboratory analysis.
24| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Chemical and pharmaceutical manufacturing remain moderately digitized; most sample collection is still performed by technicians on-site due to process complexity, safety protocols, and regulatory requirements. Adoption of full automation is slow and limited to high-volume, standardized operations. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Manufacturing and process industries with manual sampling tasks show slow AI adoption for physical operations, lagging far behind information-sector automation. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by automating sample-tracking workflows, recommending sampling schedules based on process data, and alerting technicians to anomalies that warrant sampling. Digital documentation and optimization of routes/timing increase efficiency, though the human remains essential for the physical act and certification. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can help schedule sampling intervals, analyze resulting lab data, or flag anomalies, but offers little direct assistance to the physical act of collecting the sample itself. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Sample collection requires physical manipulation of materials in varying contexts and environments, which current robotics and AI can only partially automate. While some routine sample-gathering steps could be digitally logged or sequenced, the task inherently demands in-situ physical presence and contextual judgment about where/how to collect representative samples. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical sample collection from industrial process equipment requires manipulation, access to machinery, and physical handling that current AI systems cannot perform without robotic embodiment. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Sample integrity and chain-of-custody for laboratory analysis often require documented human verification and sign-off, especially in regulated industries (pharmaceuticals, food, environmental). Legal and quality-assurance mandates typically require a human to certify sample collection procedures and authenticity. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement typically exists, but physical plant access, safety protocols, and equipment-specific procedures create organizational friction against replacing this with an AI system. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Custom automation or robotic sampling systems are capital-intensive and require ongoing integration and maintenance, making them more expensive than deploying a trained human technician for the same task across variable industrial environments. Cost per sample remains unfavorable for most applications. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI-only substitute for physical sampling, so the cost comparison favors the human worker who already performs this as part of routine duties. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No mature production systems today reliably collect physical samples from chemical/manufacturing processes end-to-end without human oversight. Robotic arms exist in controlled lab settings, but real-world industrial sampling involves diverse hazards, equipment layouts, and decision-making that deployed solutions do not handle consistently. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously collects physical material samples from separating/filtering equipment in production settings today; this remains a manual or at most semi-automated (fixed sampling ports, not AI-driven) task. |
Assemble fittings, valves, bowls, plates, disks, impeller shafts, or other parts to prepare equipment for operation.
20CI 10–30 · exposure 13 · augmentation 38 · importance 3.4/5 · click for rater detail
Assemble fittings, valves, bowls, plates, disks, impeller shafts, or other parts to prepare equipment for operation.
20| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Manufacturing automation in developed economies is longstanding but incremental; most separating and filtering equipment assembly still relies on skilled human technicians due to the customization and quality-control demands, with limited AI-agent or autonomous-system deployment. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Industrial machine operation and physical assembly in manufacturing plants is a low-digitization, physical-labor sector with minimal AI agent deployment for such tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could assist through guided assembly instructions, real-time quality verification via computer vision, and automated documentation, helping human technicians work faster and with fewer errors, though the core manual assembly remains human-performed. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could potentially assist with diagnostics, maintenance scheduling, or providing assembly instructions/manuals, but offers little direct assistance with the physical assembly action itself. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While some aspects like part identification could be automated, assembling physical components with precision tolerances and varying configurations requires dexterous manipulation, spatial reasoning, and contextual judgment that current general-purpose robots and AI systems struggle with reliably in unstructured factory settings. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical mechanical assembly task requiring manual dexterity, physical presence, and hands-on manipulation of industrial parts, none of which current AI systems can perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | While there are no strict legal barriers to automating assembly work, organizational friction (legacy equipment compatibility, workforce resistance, safety validation requirements, and need for human oversight during transition) creates moderate adoption friction. |
| Adoption barriers | claude-sonnet-5 | 3/5 | While not formally licensed in most cases, this task requires physical presence, mechanical skill, and often safety-critical judgment in industrial settings, creating natural barriers to remote AI substitution though not legal ones. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Custom industrial automation for assembly is capital-intensive (equipment, integration, maintenance), often approaching or exceeding the cost of sustained human labor, especially when accounting for reprogram costs as part variants change. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute for physical assembly, so any AI-based approach (e.g., robotics) would require far more capital investment than the human wage it replaces. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Specialized industrial robots exist for repetitive assembly in highly controlled environments, but they require extensive task-specific programming and fixturing. Deployed AI-powered general assembly systems that can handle the variety of parts and configurations described remain rare and typically limited to narrow scenarios. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs physical assembly of industrial machine parts like valves, bowls, and impeller shafts; this remains a manual labor task done by skilled technicians. |
Remove clogs, defects, or impurities from machines, tanks, conveyors, screens, or other processing equipment.
15CI 5–25 · exposure 8 · augmentation 38 · importance 3.8/5 · click for rater detail
Remove clogs, defects, or impurities from machines, tanks, conveyors, screens, or other processing equipment.
15| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Manufacturing and chemical processing sectors adopt AI/automation slowly for maintenance tasks like clog removal; most sites rely on human operators for episodic problem-solving. Adoption is driven by unscheduled failures rather than systematic digitization, limiting velocity. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Manufacturing plant maintenance and equipment operation is a low-digitization, physical-labor sector with minimal AI/robotics adoption for this specific unclogging task. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could assist by automating defect detection via camera feeds or predictive sensors to alert operators before clogs form, improving efficiency and safety planning. However, the actual removal still requires human skill, so augmentation is partial rather than transformative. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could potentially assist with predictive maintenance alerts or diagnostics to identify when clogs are forming, but it offers little help with the physical removal task itself. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | This task requires physical intervention in potentially hazardous environments with high variability in clog types, locations, and equipment configurations. While some diagnostic sensing could be automated, the actual removal of clogs demands dexterity, spatial reasoning, and safety judgment that current robotics cannot reliably perform end-to-end across diverse industrial settings. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical manipulation of industrial equipment to identify and remove clogs or debris, which current AI systems cannot perform as they lack embodied physical capability. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Safety regulations, equipment-specific authorization, and the need for human judgment during unpredictable failure modes create substantial barriers. Workers must often lock out/tag out equipment and certify safe conditions before intervention, limiting fully autonomous replacement. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement exists specifically, but safety protocols, lockout-tagout procedures, and physical access constraints create real organizational friction against automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Industrial robots capable of clog removal are expensive to deploy, integrate, and maintain relative to the episodic nature of the task. A worker's hourly wage is often cheaper than the capital and operational costs of specialized robotic systems for this unscheduled maintenance work. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute for the physical labor involved, so any AI-based approach would require robotic hardware far more costly than a human worker performing this task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No mature deployed AI system today can autonomously clear clogs from industrial equipment across varying tank, conveyor, or screen configurations. Vision systems can detect some impurities, but the physical manipulation and real-time hazard navigation required remain in the research phase for industrial automation. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs physical unclogging or debris removal from industrial processing equipment; this remains a manual maintenance task requiring physical presence and dexterity. |
Clean or sterilize tanks, screens, inflow pipes, production areas, or equipment, using hoses, brushes, scrapers, or chemical solutions.
13CI 5–21 · exposure 8 · augmentation 38 · importance 3.8/5 · click for rater detail
Clean or sterilize tanks, screens, inflow pipes, production areas, or equipment, using hoses, brushes, scrapers, or chemical solutions.
13| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption remains limited despite potential gains; most tank and equipment cleaning in pharma, food, and chemical manufacturing is still manual or semi-automated (hose + human guidance). Sectors are slow to deploy because of regulatory validation burdens, low technology maturity for the task, and organizational preference for human accountability in safety-critical cleaning. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Manufacturing and processing plant sanitation tasks are in a low-digitization, physical-labor sector with minimal AI/robotic adoption for this specific cleaning function. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI and automation can assist operators through real-time monitoring systems, scheduling optimization, chemical concentration guidance, or robotic arms handling repetitive motions (e.g., pump-spray on large surfaces), but the operator typically remains essential for inspection, adaptation to complex geometries, and regulatory sign-off. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could assist with scheduling, monitoring cleanliness via sensors, or documenting sanitation logs, but offers minimal direct assistance to the physical act of cleaning itself. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Cleaning and sterilizing equipment involves physical manipulation in unstructured environments with varied surface geometries, obstacles, and contamination types. Current AI systems lack the dexterous manipulation, environmental adaptation, and real-time decision-making needed to reliably perform this task end-to-end without significant human intervention, though some preparatory or localized cleaning steps could be partially automated. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical cleaning/sterilization task requiring manual manipulation of hoses, brushes, and scrapers in industrial settings; no off-the-shelf AI system can perform this physical labor end-to-end today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Regulatory barriers are substantial: FDA, EPA, and industry-specific sterilization standards often require validated cleaning procedures and human verification or sign-off; safety liability for inadequate sterilization (contamination risk) falls on the facility. These certification and accountability requirements create high legal and operational friction against pure automation. |
| Adoption barriers | claude-sonnet-5 | 3/5 | While not strictly licensed work, cleaning in food/chemical processing environments often falls under sanitation and safety regulations (e.g., food safety, hazardous chemical handling) requiring trained personnel and documentation, creating moderate friction against automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Specialized industrial robots capable of cleaning tasks remain expensive to purchase, integrate, and maintain, while human operators can be deployed flexibly across varied equipment and contamination scenarios. The all-in cost (hardware, integration, oversight, downtime) substantially exceeds the loaded wage of a cleaning operative. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no general-purpose AI system that substitutes for this manual task, so any specialized robotic solution would carry high capital and integration costs exceeding the wage cost of a human operator. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed commercial product reliably performs full tank/equipment cleaning and sterilization autonomously in production environments. While some robotic prototypes exist for specialized surfaces, they do not yet operate at production scale with the reliability, safety, and certification required in regulated manufacturing settings. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs physical tank/equipment cleaning; robotic cleaning systems for industrial vessels remain niche, research-stage, or highly customized rather than mature production products. |
Connect pipes between vats and processing equipment.
13CI 10–15 · exposure 0 · augmentation 13 · importance 3.5/5 · click for rater detail
Connect pipes between vats and processing equipment.
13| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Industrial chemical and processing facilities remain largely traditional in workforce deployment for such tasks; digitization is slow, and adoption of robotics for pipe connection is minimal relative to the population of such jobs. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | This task occurs in manufacturing/processing plants, a sector with low digitization of physical tasks and minimal AI-driven automation of manual piping work in production settings. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI offers minimal assistance for manual pipe connection; perhaps some augmentation through AR guidance or digital work-order systems, but the core task remains fundamentally human-hands-on with limited productivity gains from AI involvement. |
| Augmentation potential | claude-sonnet-5 | 1/5 | AI offers essentially no meaningful assistance for the physical act of connecting pipes; this is a hands-on mechanical task outside current AI capabilities. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Connecting pipes between vats and processing equipment requires physical manipulation, spatial reasoning in real environments, and precise alignment—capabilities that current AI systems lack. This is fundamentally a hands-on task unsuitable for today's technology without significant robotics development. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical manual task requiring dexterity to handle pipes, fittings, and tools in a physical plant environment; no current AI system can perform this end-to-end.rating |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Physical task execution in regulated industrial settings requires proper training and certification, but there are no strict legal prohibitions on automation; however, safety regulations and workplace standards create moderate friction to substitution. |
| Adoption barriers | claude-sonnet-5 | 2/5 | While not licensed work, physical plumbing/piping connections often require adherence to safety codes and may need trained personnel for safety-critical connections carrying hazardous or pressurized materials, but no formal licensing barrier exists. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Specialized robotics capable of this work remain expensive to acquire, integrate, and maintain, making the all-in cost substantially higher than paying a trained operator or technician to perform the task. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI/robotic solution deployed for this task, so any hypothetical system would require expensive custom robotics far exceeding the cost of a human worker performing this task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product performs this task reliably in production settings. While industrial robotics exist, general-purpose pipe-connection automation is not a solved, commercially available problem at scale. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product exists that autonomously connects industrial pipes between vats and processing equipment; this remains firmly in the domain of human manual labor and robotics research at best. |
Install, maintain, or repair hoses, pumps, filters, or screens to maintain processing equipment, using hand tools.
7CI 5–10 · exposure 0 · augmentation 25 · importance 3.8/5 · click for rater detail
Install, maintain, or repair hoses, pumps, filters, or screens to maintain processing equipment, using hand tools.
7| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Adoption of autonomous automation in maintenance-intensive manufacturing remains low and limited to highly controlled, repetitive operations. Most plants still rely on human technicians for hose, pump, filter, and screen maintenance due to task variability and safety constraints. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Manufacturing/process industries with physical equipment maintenance are slow AI adopters, especially for hands-on mechanical repair tasks which remain firmly human-performed. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI can minimally assist with diagnostic tools or monitoring alerts for equipment condition, but the core physical manipulation and decision-making about repair procedures remain human-driven. AI offers only marginal productivity gains for the core task itself. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could assist with diagnostics, maintenance scheduling, or repair manuals/guidance, but offers minimal help with the actual physical installation and repair work itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires physical manipulation of equipment components (installing, maintaining, repairing hoses, pumps, filters, screens) in industrial settings using hand tools. Current AI systems cannot perform end-to-end physical assembly, maintenance, or repair work in real-world industrial environments at scale or at parity with human speed and quality. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical maintenance/repair task requiring manual dexterity, hands-on hose/pump/filter replacement using hand tools; no current AI system can perform physical manipulation end-to-end.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong barriers exist: technicians must typically hold certifications or licenses for working on certain industrial equipment, there is legal liability for maintenance failures, and safety regulations often mandate human responsibility for equipment integrity and sign-off. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing typically required, but industrial safety protocols, equipment-specific knowledge, and physical workspace access create meaningful friction against remote or automated substitution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of implementing capable robotics (hardware, integration, maintenance, operator oversight) far exceeds the loaded wage of a skilled technician performing this maintenance work. Physical automation infrastructure costs millions of dollars compared to technician labor. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI/robotic substitute performing this hands-on repair work, so any hypothetical automation solution (specialized robotics) would be far more costly than a technician with hand tools. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI products reliably perform this task. While robotics exist for narrowly defined assembly tasks, there are no production systems that install, maintain, or repair industrial processing equipment with the flexibility and judgment this task requires. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product performs physical installation or repair of industrial hoses, pumps, filters, or screens; robotics for this specific unstructured maintenance work remains research-stage at best. |
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.