Farmworkers and Laborers, Crop, Nursery, and Greenhouse

45-2092.00

Median wage $35,660/yr265,500 employed (US)Rank #414 of 923 scored · top 45% by substitution

Manually plant, cultivate, and harvest vegetables, fruits, nuts, horticultural specialties, and field crops. Use hand tools, such as shovels, trowels, hoes, tampers, pruning hooks, shears, and knives. Duties may include tilling soil and applying fertilizers; transplanting, weeding, thinning, or pruning crops; applying pesticides; or cleaning, grading, sorting, packing, and loading harvested products. May construct trellises, repair fences and farm buildings, or participate in irrigation activities.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution29

Exposure22

Augmentation39

Substitution — the headline: capability discounted by cost, barriers and adoption.

Exposure — technical capability alone, regardless of whether anyone deploys it.

Augmentation — how much AI assists without replacing. High here + moderate substitution = a changing job, not a disappearing one.

Tasks on the substitution scale

27 rated tasks, binned by substitution score.

Position among all scored occupations

Distribution of 923 occupation scores; the marker is this occupation.

Tasks with substitution ≥ 70

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%22

panel mean rating 1.9/5 → substitution pressure 22/100

Technical feasibility todayw 20%21

panel mean rating 1.8/5 → substitution pressure 21/100

Cost vs. human wagew 15%21

panel mean rating 1.8/5 → substitution pressure 21/100

Adoption barriersw 20%inverted — strong barriers lower the score63

panel mean rating 2.5/5 (barrier strength) → substitution pressure 63/100

Sector adoption velocityw 10%14

panel mean rating 1.6/5 → substitution pressure 14/100

Task breakdown (27 tasks)

Substitution pressure per task, weighted by O*NET importance in the composite. Expand a task for the full rater audit trail — every rating, every model, every rationale.

Record information about crops, such as pesticide use, yields, or costs.

CI 65–76 · exposure 75 · augmentation 75 · 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	3/5	Farm digitization is accelerating but remains patchier than in finance or tech; early adopters and larger operations use automated logging, but many small and mid-size farms still rely on manual record-keeping, yielding middling overall velocity.
Sector adoption velocity	claude-sonnet-5	2/5	Agriculture, especially crop and nursery labor, remains a low-digitization sector with slower and more uneven adoption of digital record-keeping tools compared to information or finance sectors.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI significantly assists farmworkers by automating tedious data entry, flagging anomalies (e.g., unusual pesticide usage or low yields), and summarizing trends; the human remains in the loop to validate and decide, boosting their productivity substantially.
Augmentation potential	claude-sonnet-5	4/5	AI-enabled farm management apps significantly streamline and reduce the burden of record-keeping, letting workers input data quickly via mobile devices with automatic organization, calculations, and reporting.
Task automatability	claude-haiku-4-5-20251001	4/5	Recording crop information (pesticide use, yields, costs) is highly structured data entry and logging that AI systems can automate end-to-end via sensor data integration, image recognition for yield estimation, and database population. The task requires minimal judgment and is easily 50% time-saving with off-the-shelf OCR, vision models, and form-filling automation.
Task automatability	claude-sonnet-5	4/5	Recording structured data like pesticide use, yields, and costs is a straightforward data-entry/documentation task that can largely be handled by mobile apps, voice-to-text, and farm management software with human input of raw observations.
Adoption barriers	claude-haiku-4-5-20251001	2/5	Few regulatory or legal barriers exist to automating crop record-keeping; no licensing requirement mandates human entry. Primary barriers are organizational adoption friction and farmer preference for manual oversight, which are modest and declining.
Adoption barriers	claude-sonnet-5	3/5	Pesticide use records often have regulatory compliance requirements (EPA, state agencies) that may require certified applicator sign-off or verification, adding moderate friction beyond simple automation.
Cost vs. human wage	claude-haiku-4-5-20251001	5/5	AI-driven data logging and sensor-to-database pipelines cost far less than hiring labor to manually record and transcribe crop data; the operational cost per record is orders of magnitude lower than the loaded wage of a farmworker or data clerk.
Cost vs. human wage	claude-sonnet-5	4/5	Digital record-keeping software is inexpensive per record compared to manual paperwork time, though some human observation and initial data capture is still needed.
Technical feasibility today	claude-haiku-4-5-20251001	4/5	Multiple deployed products (farm management software with automated logging, computer vision for yield estimation, IoT sensor integrations) demonstrably perform parts of this task reliably in production. Some components like precise cost reconciliation remain semi-manual, but core recording is mature and widely used in digital agriculture.
Technical feasibility today	claude-sonnet-5	4/5	Deployed farm management software (e.g., Climate FieldView, Granular, John Deere Operations Center) reliably logs pesticide applications, yields, and costs in production agricultural operations today.

Maintain inventory, ordering materials as required.

CI 40–57 · exposure 42 · augmentation 63 · 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	Agricultural sectors, especially smaller operations where farmworkers predominate, lag in digital adoption. Inventory automation is more common in large agribusiness than in the small-to-medium farms where most crop laborers work, limiting overall adoption velocity.
Sector adoption velocity	claude-sonnet-5	2/5	Agriculture, especially crop/nursery/greenhouse labor, is a low-digitization sector where such tools are adopted slowly compared to office-based industries.
Augmentation potential	claude-haiku-4-5-20251001	3/5	Inventory systems can assist farmworkers by forecasting material needs, tracking stock levels, and flagging reorder points, improving planning efficiency. However, the augmentation is modest because human judgment on crop-specific timing and supplier selection remains central and is not dramatically amplified by current AI tools.
Augmentation potential	claude-sonnet-5	4/5	AI-enabled inventory software can significantly reduce time spent tracking stock and predicting reorder needs, meaningfully boosting efficiency for workers who still handle physical inventory.
Task automatability	claude-haiku-4-5-20251001	2/5	While inventory tracking systems can automate record-keeping and generate reordering alerts, the task requires judgment about crop cycles, seasonal demand, supplier reliability, and field conditions that vary by site. Current systems can support ordering but cannot fully replace human decisions on material quantities and timing at equal quality without significant manual oversight.
Task automatability	claude-sonnet-5	3/5	Inventory tracking and reorder triggering can be handled by software (ERP/inventory management with AI-assisted forecasting), but requires physical counting and integration with farm-specific systems that may not be automated end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	2/5	No legal licensing requirement exists to order farm materials, and no strict regulatory mandate prevents automation. However, organizational friction is real: many smaller agricultural operations use informal or manual processes, and supplier relationships often require human negotiation.
Adoption barriers	claude-sonnet-5	1/5	No licensing or regulatory requirement mandates a human perform inventory ordering; it's a low-stakes administrative task with minimal liability concerns.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Agricultural inventory systems have moderate setup and maintenance costs, and oversight by farmworkers remains necessary to validate automated suggestions. For small-to-medium farms typical of this occupation, system costs may approach or exceed the hourly wage savings from partial automation.
Cost vs. human wage	claude-sonnet-5	3/5	Off-the-shelf inventory software is affordable and cheaper than dedicated labor time, but setup, integration with suppliers, and physical counting still require human involvement, moderating savings.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	Deployed inventory management and ERP systems exist and handle basic reordering in farm operations, but they often require manual input of usage rates and frequent adjustments. Integration with farm-specific conditions (weather, pest outbreaks, growth stage) remains semi-manual; no mature end-to-end solution reliably handles the full complexity of crop-specific ordering without human intervention.
Technical feasibility today	claude-sonnet-5	3/5	Inventory management software with automated reorder points is commonly deployed in agriculture, though many small farm operations still rely on manual tracking and physical stock checks.

Operate tractors, tractor-drawn machinery, and self-propelled machinery to plow, harrow and fertilize soil, or to plant, cultivate, spray and harvest crops.

CI 35–55 · exposure 38 · augmentation 50 · importance 3.5/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	3/5	Adoption is accelerating in large-scale commodity farming (corn, soybeans in the US Midwest) and among well-capitalized farms, but remains slower in smaller farms, specialty crops, and developing regions. Pilot and early-production deployment is visible, but sector-wide displacement is still years ahead of current levels.
Sector adoption velocity	claude-sonnet-5	2/5	Agriculture is a historically slow-adopting, low-digitization sector; autonomous equipment adoption is growing but remains concentrated among large-scale row-crop operations, not typical for crop/nursery/greenhouse labor.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI-assisted guidance systems (GPS, variable-rate application) meaningfully improve farmworker productivity and precision in fertilizer/herbicide use, but the human operator typically remains at the controls for now. As autonomy increases, the augmentation role diminishes, so current systems offer moderate assistance short of full replacement.
Augmentation potential	claude-sonnet-5	3/5	GPS-guided steering, variable-rate application, and sensor-based monitoring meaningfully assist human operators in precision and efficiency while they remain in control of the machinery.
Task automatability	claude-haiku-4-5-20251001	3/5	Autonomous tractors and machinery can handle repetitive field operations like plowing, harrowing, and fertilizing with GPS guidance; however, crop harvesting remains variable due to crop heterogeneity, and selective spraying requires environmental sensing and judgment. Current systems handle 40–60% of the work but require setup, oversight, and human intervention for edge cases.
Task automatability	claude-sonnet-5	2/5	Autonomous tractors and precision ag equipment exist but require specific field conditions, setup, and human oversight; most farms still rely on human operators for the variety of tasks and terrains involved.
Adoption barriers	claude-haiku-4-5-20251001	2/5	Agricultural automation faces minimal legal barriers—no license requirement to operate autonomous machinery—but practical barriers include farmer capital constraints, equipment compatibility with diverse farm sizes, and organizational inertia toward traditional methods. Liability for crop damage and machine failure introduces some friction but not hard legal bars.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement for a human to operate farm equipment, but liability for equipment damage, crop loss, and safety around autonomous machinery in variable field conditions creates some adoption friction.
Cost vs. human wage	claude-haiku-4-5-20251001	3/5	Initial capital cost for autonomous machinery is high, but per-hour operational cost (fuel, maintenance, no operator wage) is competitive with or slightly cheaper than hired farmworker labor in high-income countries; the comparison is less favorable in low-wage regions, making the cost ratio context-dependent and roughly comparable overall.
Cost vs. human wage	claude-sonnet-5	2/5	Autonomous equipment requires substantial capital investment, GPS infrastructure, and maintenance, making it costly relative to seasonal farm labor except at very large scale operations.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	Autonomous farm equipment exists in production (e.g., John Deere autonomous tractors, spray drones) and is deployed in large-scale operations, but reliability and scope remain limited to structured fields and simple crops. Error rates and weather/soil variability constraints mean systems are not yet universally reliable across farm contexts.
Technical feasibility today	claude-sonnet-5	2/5	Autonomous tractor products (e.g., John Deere's autonomous tractor line) are commercially available but deployment is limited to specific operations like tillage on large, well-mapped fields, not the full range of plowing, planting, spraying, and harvesting.

Regulate greenhouse conditions, and indoor and outdoor irrigation systems.

CI 30–52 · exposure 38 · augmentation 75 · 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	Large-scale commercial greenhouses and nurseries are piloting automated climate control and drip irrigation, but adoption remains concentrated in high-value crops and capital-rich operations. Most smaller farms and nurseries operate with manual or simple timer-based systems due to upfront cost and integration complexity.
Sector adoption velocity	claude-sonnet-5	2/5	Agriculture, especially crop and nursery farming, is a historically slow-adopting, capital-constrained sector with uneven digitization, though large-scale greenhouse operations are ahead of the curve.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI-driven monitoring dashboards, sensor alerts, and automated logging substantially increase a farmworker's ability to detect problems early and make informed irrigation decisions across multiple zones. Real-time data augmentation and predictive alerts clearly enhance productivity and decision quality while the human remains in direct control.
Augmentation potential	claude-sonnet-5	4/5	Sensor-based monitoring, automated alerts, and AI-driven irrigation scheduling significantly help workers manage greenhouse conditions more efficiently even when humans remain responsible for final decisions.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI systems can monitor and log greenhouse sensor data, and simple rule-based irrigation scheduling is feasible, current systems lack the real-time adaptability and failure recovery needed for end-to-end autonomous operation of complex physical systems. Significant human oversight remains required for equipment troubleshooting, unexpected conditions, and system calibration.
Task automatability	claude-sonnet-5	3/5	Automated climate and irrigation control systems can regulate greenhouse conditions and irrigation with sensors and controllers, but full task includes physical adjustments, monitoring outdoor systems, and responding to anomalies not fully covered by off-the-shelf AI.'
Adoption barriers	claude-haiku-4-5-20251001	3/5	No strict licensing requirement exists for greenhouse operation, but there are practical barriers: equipment liability if autonomous systems fail (crop loss, plant damage), grower preference for experienced manual management, and organizational reluctance to trust fully automated systems without redundancy and backup. Oversight and human sign-off remain the norm.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement blocks automation, but physical infrastructure investment, equipment maintenance, and variable outdoor conditions create moderate practical friction to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Initial hardware and software costs for integrated environmental control systems are substantial, and the labor savings are modest compared to a farmworker's loaded wage since much calibration and maintenance still requires human expertise. Cost parity has not yet been achieved for small to mid-sized operations.
Cost vs. human wage	claude-sonnet-5	3/5	Automated control systems can reduce labor costs over time but require significant upfront capital investment in sensors, actuators, and software, making cost parity rather than dramatic savings typical for smaller operations.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Sensor monitoring and automated irrigation controllers exist in commercial deployments, but full autonomous regulation of greenhouse microclimates and irrigation requires integration across multiple systems with reliable fail-safes that current agricultural automation products handle only partially. Most production systems still require substantial human decision-making and manual intervention.
Technical feasibility today	claude-sonnet-5	3/5	Commercial greenhouse automation products (climate computers, smart irrigation controllers) exist and are deployed, but many operations still rely on manual oversight and adjustment, especially smaller farms and outdoor irrigation.

Record information about plants and plant growth.

CI 30–44 · exposure 33 · augmentation 50 · 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	Agricultural technology adoption varies widely by farm size and region; precision agriculture tools are growing in larger operations but remain sparse in small to mid-sized farms where most farmworkers operate. Overall adoption remains slow relative to information-intensive sectors.
Sector adoption velocity	claude-sonnet-5	2/5	Agriculture, especially crop/nursery/greenhouse labor, is a sector with historically low digitization and slow AI adoption compared to knowledge-work industries, though some greenhouse operations are adopting IoT sensors gradually.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI-powered plant monitoring dashboards and image analysis tools can assist farmworkers by flagging anomalies and automating data entry from photos, moderately raising productivity without removing human inspection and judgment from the recording process.
Augmentation potential	claude-sonnet-5	3/5	Mobile apps, barcode/QR scanning, and sensor dashboards can meaningfully speed up and standardize how workers record and log plant data, even though full automation of decision-making observation remains limited.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can analyze images of plants and interpret some growth metrics, the task requires physical inspection, nuanced observation of multiple plant parameters, and integration with contextual farm-specific details. Current systems cannot reliably achieve 50% time savings end-to-end without substantial manual verification and setup.
Task automatability	claude-sonnet-5	3/5	Recording plant/growth data is often structured data entry (measurements, counts, observations) that AI-enabled apps or sensors combined with simple logging software can partially automate, but manual observation and data collection in the field still requires human presence and judgment for many crops.
Adoption barriers	claude-haiku-4-5-20251001	3/5	No legal licensing requirement exists, but organizational friction is moderate: farmers rely on embodied knowledge, prefer direct observation, and integrate records with existing informal practices. Some farms have regulatory record-keeping obligations that create friction against full automation.
Adoption barriers	claude-sonnet-5	2/5	No licensing or legal requirement mandates a human record this information, though quality control and accuracy needs create some organizational friction against fully trusting automated systems.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Deploying vision systems, integrating with farm management software, and maintaining oversight adds non-trivial costs. A farmworker's wage remains competitive with the all-in expense of reliable automated monitoring infrastructure.
Cost vs. human wage	claude-sonnet-5	2/5	Sensor networks and digital tracking systems require upfront capital, integration, and maintenance costs that often exceed the low wages of manual laborers for this specific task, making cost parity uncertain in many small-to-mid operations.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Some image recognition products can identify plants and growth stages, but reliable production deployment for consistent real-world recording across diverse crop types and conditions remains limited. Most solutions are research-stage or require heavy manual annotation.
Technical feasibility today	claude-sonnet-5	2/5	Some ag-tech products offer digital logging, IoT sensors, and image-based growth tracking, but widespread reliable deployment specifically for manual field recording tasks by laborers is limited and often supplements rather than replaces human recording.

Feel plants' leaves and note their coloring to detect the presence of insects or disease.

CI 33–35 · exposure 25 · augmentation 50 · 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	2/5	Crop monitoring AI adoption is growing in large-scale commodity agriculture, but nurseries and small farms (where this task is prevalent) lag significantly. Adoption remains pilot-heavy rather than production-standard in the sectors where this work is concentrated.
Sector adoption velocity	claude-sonnet-5	1/5	Crop, nursery, and greenhouse farm labor is a low-digitization, physical sector with slow, uneven AI adoption compared to information or finance industries.
Augmentation potential	claude-haiku-4-5-20251001	3/5	Image recognition tools can flag suspicious leaf areas for human inspection and help document disease progression, moderately improving a farmworker's efficiency. However, the tactile and contextual judgment required limits augmentation value compared to visual-only tasks.
Augmentation potential	claude-sonnet-5	3/5	AI-powered diagnostic apps and imaging tools can assist workers by confirming or flagging suspected disease/pest issues, improving accuracy without replacing the physical inspection itself.
Task automatability	claude-haiku-4-5-20251001	2/5	Visual inspection via cameras and machine learning can detect some leaf discoloration and pest damage, but tactile feedback (leaf texture, wilting firmness) is critical to diagnosis and currently cannot be reliably assessed by AI. Humans must still perform substantial manual tactile inspection.
Task automatability	claude-sonnet-5	2/5	While computer vision can detect some plant diseases from images, the tactile 'feel' component and integrated sensory judgment across diverse plants in situ is not automatable end-to-end with off-the-shelf systems today.'
Adoption barriers	claude-haiku-4-5-20251001	2/5	Minimal regulatory barriers exist, but significant practical friction remains: farmers rely on embodied knowledge, trust humans to make nuanced pest/disease judgments, and require rapid in-field decision-making that current AI systems cannot yet support reliably enough to fully displace inspection.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement, but organizational friction exists since this task is bundled with physical plant handling and requires equipment investment not easily justified for the task alone.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Current camera systems and agronomic AI require infrastructure investment, integration, and ongoing model maintenance. For small-scale and nursery operations, the per-task cost remains comparable to or higher than a farmworker's hourly inspection labor.
Cost vs. human wage	claude-sonnet-5	2/5	Deploying sensors, drones, or imaging systems plus agronomist oversight for small-scale nurseries/crops is often costlier than low-wage manual labor already performing this task as part of broader duties.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Computer vision systems exist for crop disease detection but struggle with fine-grained pest identification and miss tactile cues essential to diagnosis. No deployed agricultural system reliably replaces hands-on leaf inspection across the diversity of crops and conditions encountered.
Technical feasibility today	claude-sonnet-5	2/5	Agricultural imaging/scouting drones and phone-based diagnostic apps exist but are narrow, error-prone on early-stage or subtle symptoms, and rarely handle tactile inspection or unstructured greenhouse/nursery settings at production scale.

Set up and operate irrigation equipment.

CI 30–38 · exposure 25 · augmentation 63 · 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	Irrigation automation is spreading in larger commercial operations and intensive greenhouse settings, but adoption remains uneven across small farms and seasonal labor-intensive crops; adoption is neither laggard nor leading-edge.
Sector adoption velocity	claude-sonnet-5	2/5	Agriculture is a historically low-digitization sector; while precision ag and IoT irrigation tools are growing, adoption remains slow and concentrated in large-scale operations rather than widespread among crop/nursery laborers.
Augmentation potential	claude-haiku-4-5-20251001	4/5	AI-driven irrigation monitoring, predictive maintenance alerts, and soil-moisture-based scheduling substantially assist farmworkers in setting up and maintaining systems more efficiently and responsively than manual-only approaches.
Augmentation potential	claude-sonnet-5	3/5	AI-driven sensors and scheduling software can meaningfully assist workers by optimizing water use timing and flagging equipment issues, improving efficiency even though physical setup remains manual.
Task automatability	claude-haiku-4-5-20251001	2/5	While some irrigation components (timers, sensors, flow monitoring) can be automated, the task requires field assessment, equipment troubleshooting, and adaptive response to weather and soil conditions that demand human judgment and physical presence today.
Task automatability	claude-sonnet-5	2/5	While automated/smart irrigation controllers can manage scheduling, physically setting up pipes, sprinklers, and equipment in varied field conditions still requires substantial human manual labor and adaptability that current AI/robotics cannot fully replace.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Water rights regulations, environmental compliance, and liability for crop loss create moderate barriers; however, no legal requirement mandates human sign-off, and many growers already adopt automated systems, lowering adoption friction.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirements block automation, but physical infrastructure costs, variable terrain, and the need for on-site labor to connect/move equipment create practical friction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Irrigation automation hardware (sensors, controllers, water management software) has significant upfront capital costs; when amortized over small farms or seasonal operations, the cost-per-task remains comparable to or higher than farmworker wages.
Cost vs. human wage	claude-sonnet-5	2/5	Precision irrigation tech requires significant capital investment (sensors, controllers, sometimes robotics) that is often costlier than low-wage farm labor for equipment setup, though software-only scheduling optimization can be cheap.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Deployed systems exist for automated irrigation scheduling and remote monitoring, but they typically require human setup, maintenance, and manual intervention for equipment failures or unusual conditions—no production system fully operates irrigation systems end-to-end without human oversight.
Technical feasibility today	claude-sonnet-5	2/5	Smart irrigation controllers and sensor-based systems are deployed in some commercial agriculture, but full setup and operation of physical irrigation equipment by autonomous systems remains largely research-stage or limited to high-value crops.

Provide information and advice to the public regarding the selection, purchase, and care of products.

CI 28–39 · exposure 25 · augmentation 50 · 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	2/5	Agricultural retail, especially smaller nurseries and farm cooperatives, lags in digital adoption; while large garden centers may pilot chatbots, deep production deployment across the sector is minimal and slow.
Sector adoption velocity	claude-sonnet-5	1/5	Agriculture, nursery, and retail garden centers are low-digitization, physically-oriented sectors with minimal AI agent deployment in customer-facing advisory roles.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI tools can assist workers by providing quick reference information, care guides, and product specs they can relay or refine for customers, moderately raising their efficiency without replacing the human advisory role.
Augmentation potential	claude-sonnet-5	3/5	AI can help workers quickly look up plant care information, pest identification, and product specs, moderately boosting their ability to advise customers.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can generate product information and care advice, this task requires live interaction with customers, understanding their specific needs, and adaptive conversation—capabilities that current AI systems handle poorly at scale in production agricultural retail environments. Some chatbot support exists, but customer satisfaction and error rates remain problematic for full automation.
Task automatability	claude-sonnet-5	2/5	Basic plant-care Q&A can be handled by chatbots, but nuanced in-person advice tailored to specific products, local conditions, and customer needs still requires human judgment and physical demonstration.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Customers in agricultural retail often prefer human interaction for complex plant/product decisions; liability concerns around incorrect care advice create moderate friction; no hard legal barrier exists, but organizational preference for human touch provides meaningful resistance.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement, but customers often prefer face-to-face advice and physical product handling, creating moderate friction against full substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Deploying and maintaining a reliable AI advisory system (chatbot, integration, oversight, failure handling) often exceeds the wage cost of a part-time farmworker in rural or small-business agricultural settings, particularly when factoring in liability for bad advice.
Cost vs. human wage	claude-sonnet-5	3/5	AI chat support is cheap per query, but integrating it into physical retail settings with product-specific knowledge and human backup keeps costs roughly comparable to low-wage retail labor once implementation is factored in.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Basic product information systems and limited chatbots exist, but reliable deployment in actual nurseries and farm supply contexts is rare; most real operations still rely on human staff for customer advice due to the need for contextual judgment and personalized recommendations.
Technical feasibility today	claude-sonnet-5	2/5	Chatbots and search tools exist for generic gardening advice, but no deployed product reliably replaces in-person retail advisory interactions at nurseries/greenhouses today.

Participate in the inspection, grading, sorting, storage, and post-harvest treatment of crops.

CI 30–35 · exposure 25 · augmentation 50 · 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	Adoption is slow and concentrated in large commodity operations; most farmwork remains labor-intensive and geographically dispersed across small and medium operations with limited capital for automation and minimal digital infrastructure.
Sector adoption velocity	claude-sonnet-5	2/5	Agriculture is traditionally a slow-adopting sector for AI and robotics due to low digitization, seasonal variability, and fragmented small-to-mid-size operations, though some large agribusinesses are piloting automated sorting technology.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI-powered optical grading and sorting systems can meaningfully assist workers by flagging defects and organizing crops by category, but workers typically retain final quality judgment and must manage unpredictable crop variation and storage decisions.
Augmentation potential	claude-sonnet-5	3/5	AI-assisted vision systems and sensors can help workers flag defects or quality issues faster and more consistently, improving throughput and reducing error rates in grading and sorting, even if humans remain central to inspection and handling.
Task automatability	claude-haiku-4-5-20251001	2/5	While some post-harvest operations like sorting and grading can be partially automated with optical systems, the task requires substantial human judgment for quality assessment, physical handling of delicate crops, and dynamic decision-making about storage conditions that current AI cannot reliably perform end-to-end at 50% time savings.
Task automatability	claude-sonnet-5	2/5	Some sorting/grading tasks can be done by machine vision systems, but the broader task including inspection, storage decisions, and post-harvest treatment across diverse crops still requires significant human dexterity, judgment, and physical handling that current AI cannot fully replicate end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Moderate barriers exist: food safety regulations and traceability requirements create oversight needs, customer quality preferences favor human expertise, and the physical environment (seasonal labor availability, crop variability) creates organizational friction against full substitution.
Adoption barriers	claude-sonnet-5	2/5	No licensing or legal requirement mandates human performance of this task, though food safety regulations may require certain inspection standards; the main barriers are practical (variable produce, outdoor/field conditions, capital cost) rather than legal or professional in nature.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Automated sorting equipment requires high capital investment, integration costs, and ongoing maintenance; for small to medium operations and specialty crops, the total cost per unit processed remains comparable to or exceeds hiring farmworkers.
Cost vs. human wage	claude-sonnet-5	2/5	Automated sorting/grading machinery requires substantial capital investment, making it cost-effective mainly for large operations with high volume; for many farms the loaded cost of low-wage manual labor remains cheaper or comparable to specialized equipment plus integration and maintenance.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Deployed sorting and grading systems exist (e.g., optical sorters, packing lines) but handle only specific crops and standardized parameters; they require significant human oversight, cannot match complex human judgment on quality, and remain narrow in scope compared to the full task.
Technical feasibility today	claude-sonnet-5	2/5	Automated sorting/grading equipment exists in some large-scale packing operations (e.g., optical sorters for fruit), but these are narrow, crop-specific deployments rather than general solutions covering the full range of inspection, storage, and treatment tasks described.

Maintain and repair irrigation and climate control systems.

CI 30–35 · exposure 25 · augmentation 50 · 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	Adoption of precision agriculture and automated monitoring is growing but concentrated in large, capital-intensive operations. Small and mid-sized farms—which employ the majority of farmworkers—lag significantly, making sector-wide adoption slow and uneven.
Sector adoption velocity	claude-sonnet-5	2/5	Agriculture is a historically slow-adopting, low-digitization sector, with IoT sensors gaining some traction but actual maintenance automation lagging.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI-powered monitoring systems and predictive maintenance alerts can assist technicians by flagging problems early and recommending diagnostics, reducing troubleshooting time. However, the physical repair work remains human-directed, limiting the overall productivity lift.
Augmentation potential	claude-sonnet-5	3/5	AI-based monitoring systems can predict failures and guide maintenance scheduling, improving efficiency of the human technician even though hands-on repair remains manual.
Task automatability	claude-haiku-4-5-20251001	2/5	While monitoring sensors and scheduling adjustments could be partially automated, the physical repair work (replacing pipes, fixing equipment) requires on-site manipulation that current mobile robotics cannot reliably perform in diverse agricultural environments. Diagnostics could be aided by AI, but hands-on maintenance remains predominantly manual.
Task automatability	claude-sonnet-5	2/5	Diagnosing and physically repairing pumps, valves, wiring, and HVAC/greenhouse climate hardware requires manual dexterity and situational troubleshooting that current AI cannot perform end-to-end.'
Adoption barriers	claude-haiku-4-5-20251001	3/5	Agricultural operations vary widely in technical sophistication and capital availability; no licensing barrier exists, but organizational friction is moderate—small farms may lack digitized systems to automate, and equipment heterogeneity complicates standardized solutions. Liability for system failure adds some friction.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement typically, but physical access to equipment, safety concerns, and lack of robotic actuation create practical barriers to full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI monitoring tools require significant infrastructure investment (sensors, cloud integration, integration with existing systems) and still need human technicians on-site. The all-in cost per repair event likely exceeds the loaded wage of a farmworker, especially for small operations.
Cost vs. human wage	claude-sonnet-5	2/5	Sensor/monitoring software is cheap, but the repair labor itself still requires a paid technician, so overall cost savings versus a human worker are limited.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Deployed products exist for sensor monitoring and predictive maintenance alerts, but no current system reliably diagnoses and repairs irrigation/climate systems autonomously. Production systems are limited to data collection and alerting; humans still perform 80%+ of actual repair and maintenance work.
Technical feasibility today	claude-sonnet-5	2/5	Smart sensors and monitoring dashboards exist to flag anomalies, but actual repair work is still done by human technicians; no deployed product performs the physical maintenance task.

Inform farmers or farm managers of crop progress.

CI 30–35 · exposure 25 · augmentation 50 · 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	Agriculture remains a laggard sector for digital automation. While some large-scale, capital-intensive operations adopt precision agriculture, most small and mid-sized farms—where farmworkers are concentrated—have slow adoption of remote monitoring. Current uptake is pilot-heavy, not production-standard.
Sector adoption velocity	claude-sonnet-5	2/5	Agriculture, especially crop/nursery/greenhouse labor, is a low-digitization sector with slow AI adoption relative to information or finance industries.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI-assisted crop monitoring (image recognition on photos, automated alerts from sensors, summary report generation) usefully augments a farmworker's daily observations, helping them prioritize which fields to scout and communicate status more thoroughly. The human farmworker remains the decision-maker and observer, but AI can reduce time spent compiling and communicating routine information.
Augmentation potential	claude-sonnet-5	3/5	AI tools like crop-monitoring apps, image recognition, and reporting dashboards can help workers document and communicate progress more efficiently, though human field observation remains central.
Task automatability	claude-haiku-4-5-20251001	2/5	This task requires observing crop conditions in situ and communicating findings to decision-makers. While AI can assist with image analysis of crops or generate summary reports from sensor data, it cannot reliably replace the on-site inspection, judgment about timing, and contextual communication that experienced farmworkers provide. Achieving 50% time savings with equal quality would require near-perfect remote sensing—currently unavailable for most small to mid-scale farms.
Task automatability	claude-sonnet-5	2/5	Reporting crop progress requires field observation and physical assessment that AI cannot independently perform; AI could help summarize or communicate data a human has already gathered, but not replace the observation-to-communication task end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	3/5	Moderate barriers exist: farmers are accustomed to human-based communication, trust face-to-face reporting, and may distrust algorithmic summaries of crop health. However, no legal requirement mandates human observation, and agricultural technology adoption is growing. Organizational friction in adopting new monitoring systems is material but not insurmountable.
Adoption barriers	claude-sonnet-5	2/5	No licensing or legal requirement mandates human reporting, but practical barriers exist: workers already on-site have context, judgment, and trust relationships that make full automation organizationally unlikely.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Current AI crop-monitoring solutions (sensors, imagery, cloud processing, analytics) cost hundreds to thousands of dollars annually and still require human interpretation and oversight. For a farmworker wage ($25k–$35k/year loaded), this remains comparable or more expensive, especially for small operations that dominate the sector.
Cost vs. human wage	claude-sonnet-5	2/5	Deploying sensors, drones, or imaging systems to replace human observation and reporting entails meaningful capital and integration costs compared to a low-wage worker simply relaying observations.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Deployed products exist for crop monitoring (e.g., satellite/drone imagery, crop sensors with AI analysis), but they typically narrow to single crops, controlled environments, or require manual labeling. Real-world farm communication remains human-centered. No mature production system autonomously observes, interprets, and reports crop progress across diverse farm conditions with demonstrated reliability at scale.
Technical feasibility today	claude-sonnet-5	2/5	Some sensor/drone-based crop monitoring products exist, but routine verbal or written reporting from farmworkers to managers is still done informally and manually in most operations, not via deployed AI products.

Harvest plants, and transplant or pot and label them.

CI 28–35 · exposure 25 · augmentation 25 · 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	Agricultural automation is progressing but remains concentrated in large-scale commodity operations; most farmworker roles in nurseries, greenhouses, and diverse crop harvesting remain manual, with pilot projects outpacing production deployment.
Sector adoption velocity	claude-sonnet-5	1/5	Agriculture is a historically low-digitization, physically-demanding sector with slow robotics adoption; crop/nursery labor automation remains mostly pilot-stage outside a few high-value specialty crops.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI and robotics offer minimal assistance to human farmworkers performing this task today; tools are neither widely available in field/greenhouse settings nor designed to augment rather than replace manual harvesting and transplanting workflows.
Augmentation potential	claude-sonnet-5	2/5	AI-assisted tools like plant health sensors, robotic aids, or labeling automation can support workers in some steps, but the physical dexterity required for most harvesting and transplanting means augmentation is currently limited.
Task automatability	claude-haiku-4-5-20251001	2/5	While some harvesting and transplanting steps could theoretically be roboticized, current systems lack reliable dexterity and environmental adaptability to perform the full task end-to-end with 50% time savings at equal quality across diverse crop types and field conditions.
Task automatability	claude-sonnet-5	2/5	Some harvesting/potting/transplanting robots exist for specific crops (e.g., strawberries, nursery pots) but the general task across diverse plants, terrain, and delicate handling remains largely manual with no off-the-shelf system achieving 50% time savings broadly.
Adoption barriers	claude-haiku-4-5-20251001	3/5	While there are no hard licensing barriers to automation, adoption faces organizational friction (capital investment for small farms), quality and liability concerns around plant damage, and customer preferences in specialty nurseries for hand-selected stock and labeling accuracy.
Adoption barriers	claude-sonnet-5	2/5	No licensing or legal requirement for human performance, but organizational friction is real: capital cost, need for customized equipment per crop type, and physical/environmental variability slow substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Current agricultural robots and robotic systems remain expensive to acquire, integrate, and maintain, making their cost per task-equivalent higher than the loaded wage of farmworkers in most regions, especially in labor-abundant markets.
Cost vs. human wage	claude-sonnet-5	2/5	Specialized harvesting/potting robots are capital-intensive and often cost more per unit output than low-wage farm labor, though some large-scale nursery automation can approach cost parity for specific repetitive tasks.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Prototype agricultural robots exist for harvesting specific crops (e.g., strawberries, apples), but deployed systems are narrow in scope, have material error rates, require controlled environments, and are not yet reliable at production scale across the diverse crop and nursery operations implied by the task.
Technical feasibility today	claude-sonnet-5	2/5	Deployed agricultural robots handle narrow use cases like automated potting lines in large nurseries, but harvesting and transplanting most crops in field/greenhouse settings still rely on human labor due to variability in plant size, ripeness, and fragility.

Identify plants, pests, and weeds to determine the selection and application of pesticides and fertilizers.

CI 25–35 · exposure 30 · augmentation 63 · importance 3.5/5 · 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	Farmworkers operate in lower-digitization, fragmented sectors; while precision agriculture technology is growing, actual deployed AI decision agents for pesticide selection remain rare in production, with adoption primarily in large-scale, data-rich operations.
Sector adoption velocity	claude-sonnet-5	2/5	Agriculture is a historically low-digitization sector; precision ag and AI scouting tools are in pilot/early adoption phases rather than deep production use.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI-assisted plant and pest identification tools can speed visual inspection and flag issues a worker might miss, moderately raising productivity, but the worker typically retains judgment over chemical selection and application.
Augmentation potential	claude-sonnet-5	4/5	AI-powered pest/weed identification apps and drone imagery meaningfully assist workers in spotting issues faster and more accurately, improving decision quality even though humans still perform physical application.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI can identify plants and pests via image recognition with reasonable accuracy, determining appropriate pesticide/fertilizer selection and dosage requires site-specific judgment (soil conditions, weather, crop stage, regulations) that current systems cannot reliably perform end-to-end without substantial human oversight.
Task automatability	claude-sonnet-5	2/5	Vision-based identification of plants/pests/weeds is technically feasible via image recognition apps, but the full task—linking identification to correct pesticide/fertilizer selection and application in field conditions—still requires human judgment and physical execution.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Pesticide selection and application are heavily regulated (EPA, state pesticide applicator licensing); liability for crop damage, ecological harm, or residue violations falls on the operator, creating strong legal barriers to fully autonomous AI decision-making without licensed human sign-off.
Adoption barriers	claude-sonnet-5	3/5	Pesticide application often requires licensed applicators and regulatory compliance, and misapplication carries crop/environmental liability, creating moderate barriers even if identification itself is automatable.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI vision systems and decision-support platforms have deployment and integration costs; however, farmworkers are typically low-wage labor, and the overhead of maintaining systems with oversight often does not yet reach economic parity with seasonal human labor in small-to-medium operations.
Cost vs. human wage	claude-sonnet-5	2/5	Deploying imaging/sensor systems plus agronomic decision software has meaningful setup and hardware costs that are not clearly cheaper than a trained worker's judgment for small-to-mid operations, though large farms may see gains.
Technical feasibility today	claude-haiku-4-5-20251001	3/5	Computer vision models for plant and pest identification are deployed in some agricultural products and research tools, but they operate with material error rates on field images and typically require expert confirmation before application decisions are made.
Technical feasibility today	claude-sonnet-5	2/5	Consumer plant/pest ID apps and some precision-ag scouting tools exist but have material error rates outdoors and are rarely integrated into full application decisions in production farm operations.

Inspect plants and bud ties to assess quality.

CI 23–35 · exposure 25 · augmentation 38 · importance 3.3/5 · 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	Agricultural technology adoption is slower in smaller farms and nurseries; greenhouse operations are more capital-intensive but still largely rely on human inspectors. Pilots exist but production deployment of automated plant inspection remains limited across the sector.
Sector adoption velocity	claude-sonnet-5	1/5	Agricultural field labor is a low-digitization sector with slow, limited AI adoption, mostly confined to isolated pilots in precision agriculture rather than widespread deployed inspection systems.
Augmentation potential	claude-haiku-4-5-20251001	3/5	Computer vision tools can flag suspect plants or defects for faster human review, reducing eye-strain and improving consistency, but the farmworker must ultimately make the quality judgment. This assistive use improves productivity modestly without replacing human expertise.
Augmentation potential	claude-sonnet-5	2/5	AI-based imaging tools can flag potential issues or anomalies to support worker attention, but given the specific dexterous and contextual judgment needed for bud-tie inspection, the productivity boost remains limited.
Task automatability	claude-haiku-4-5-20251001	2/5	Computer vision can detect some visible defects and anomalies, but assessing overall plant quality requires nuanced judgment about growth stage, subtle health indicators, and bud tie condition that varies by crop type. Current systems struggle with this level of horticultural expertise and cannot achieve the 50% time savings threshold end-to-end.
Task automatability	claude-sonnet-5	2/5	Visual inspection tasks require fine motor coordination with plants and physical handling of bud ties in variable field/greenhouse conditions, which current AI vision systems can partially assist but not reliably replace end-to-end without significant robotic hardware investment.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Quality assessment is legally and contractually tied to human expertise for produce export standards, food safety compliance, and customer specifications. Buyers typically require human sign-off on plant quality grading, creating strong regulatory and liability barriers to full automation.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement exists for this task, but physical environment variability, plant fragility, and the need for dexterous handling of bud ties create practical organizational friction against automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	High-quality agricultural imaging systems, labor for camera setup/deployment, and the need for human validation of borderline cases make the total cost comparable to or exceeding a farmworker's loaded wage for equivalent inspection output.
Cost vs. human wage	claude-sonnet-5	2/5	Deploying specialized vision sensors and robotic inspection systems in variable outdoor/greenhouse environments carries high capital and integration costs relative to low-wage farm labor, making AI more expensive on a per-task basis today.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	While image classification models exist for plant disease and defect detection, deployed agricultural inspection products remain narrow in scope and have material error rates. No mainstream production system reliably assesses holistic plant quality and bud tie integrity across diverse crop types at scale.
Technical feasibility today	claude-sonnet-5	2/5	Some agricultural computer vision products exist for crop health monitoring and defect detection, but these are narrow-scope solutions not widely deployed for the specific combined task of plant/bud-tie quality inspection in production settings.

Clean work areas, and maintain grounds and landscaping.

CI 24–33 · exposure 20 · augmentation 25 · 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	Agricultural labor remains in laggard sectors for automation adoption: small farms, outdoor/physical work, variable conditions, and low capital investment appetite slow deployment of AI and robotics for grounds and cleaning tasks.
Sector adoption velocity	claude-sonnet-5	1/5	Agriculture and landscaping are low-digitization, physical-labor-heavic sectors with minimal AI/robotics adoption at scale for general grounds maintenance.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI could assist with task planning and area monitoring (e.g., identifying zones needing attention via imagery), but augmentation is limited because the core work is physical manipulation in unstructured outdoor environments where human oversight and decision-making remain essential.
Augmentation potential	claude-sonnet-5	2/5	AI-enabled tools like GPS-guided mowers or app-based scheduling can assist some aspects of grounds maintenance, but the core physical cleaning and landscaping work sees limited productivity gains from AI.
Task automatability	claude-haiku-4-5-20251001	2/5	Cleaning and landscaping involve highly variable physical environments, obstacle detection, and dexterous manipulation that current robots and AI agents struggle with at scale. While some narrow subprocesses (e.g., identifying visible debris piles via computer vision) are automatable, end-to-end grounds maintenance and cleaning work remains impractical for most farm and nursery settings today.
Task automatability	claude-sonnet-5	2/5	Cleaning work areas and maintaining grounds/landscaping requires physical manipulation across varied, unstructured outdoor terrain, which current AI and robotics cannot fully replace; some robotic mowers exist but don't cover general cleanup and landscaping.'
Adoption barriers	claude-haiku-4-5-20251001	2/5	These tasks are physical and site-specific, with modest regulatory barriers but significant practical friction: farm owners must invest in specialized equipment, ensure safety protocols, and manage integration with existing workflows. No legal requirement mandates human presence, but organizational and capital hurdles are real.
Adoption barriers	claude-sonnet-5	2/5	No licensing or legal requirement mandates a human for this task, though physical unpredictability of outdoor environments and liability for equipment damage create some practical friction.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	Labor costs for farmworkers and landscapers in the U.S. remain low relative to the capital, maintenance, and operational complexity of robots or AI systems capable of performing this task outdoors in unstructured environments.
Cost vs. human wage	claude-sonnet-5	2/5	Specialized robotic mowers can be cost-competitive for large flat lawns, but general cleaning and landscaping tasks still require human labor or expensive custom robotics, making overall cost comparable or higher than human labor.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed commercial systems reliably perform general farm or nursery cleaning and grounds maintenance at production scale. Robotics exist for specialized tasks like weeding in controlled environments, but practical, integrated solutions for varied outdoor cleaning and landscaping work remain research-stage.
Technical feasibility today	claude-sonnet-5	2/5	Deployed products (robotic mowers, some autonomous cleaning devices) handle narrow subtasks like lawn mowing, but general grounds maintenance and cleanup are still done by humans in production settings.

Tie and bunch flowers, plants, shrubs, and trees, wrap their roots, and pack them into boxes to fill orders.

CI 24–28 · exposure 16 · augmentation 25 · importance 3.2/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Nurseries and greenhouse operations are typically small to mid-sized, geographically dispersed, and capital-constrained; agricultural technology adoption lags information and finance sectors significantly, and this particular task remains largely manual even in larger operations.
Sector adoption velocity	claude-sonnet-5	1/5	Crop and nursery labor is a low-digitization, physical-task sector with minimal AI/robotics adoption for this kind of delicate manual work.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI-assisted design tools or yield optimization could help upstream, but during the hands-on tying, wrapping, and packing phase itself, current AI offers minimal real-time assistance—computer vision might flag plant defects, but human dexterity and judgment remain primary.
Augmentation potential	claude-sonnet-5	2/5	Some assistive tools (conveyor systems, basic sorting aids) can support workers, but AI offers little direct augmentation to the tying/bunching/wrapping process itself.
Task automatability	claude-haiku-4-5-20251001	2/5	While robotic systems exist for simple repetitive tasks, tying, bunching, wrapping roots, and packing flowers into boxes requires dexterous manipulation, adaptive force control, and visual judgment of plant condition—capabilities that current off-the-shelf systems struggle with reliably. Partial automation of packing might be feasible, but the full task is far from the 50% time-saving threshold at equal quality.
Task automatability	claude-sonnet-5	2/5	This requires fine motor manipulation of variable, delicate organic materials (tying, bunching, wrapping roots) which remains difficult for robots to do reliably at human speed and quality; only partial automation exists for uniform packing steps.
Adoption barriers	claude-haiku-4-5-20251001	2/5	There are modest barriers: no strict licensing requirement for the work itself, but customer expectations for plant quality and handling, organizational friction around capital investment, and the physical complexity of the task create friction to full automation adoption.
Adoption barriers	claude-sonnet-5	2/5	No licensing or legal barriers, but physical handling of fragile perishable goods requiring dexterity and quality judgment creates practical friction against automation.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	The capital and integration costs for robotic systems capable of handling delicate plant manipulation, combined with ongoing maintenance and oversight, far exceed the loaded wage of farmworkers performing this labor, particularly in seasonal or variable-volume operations.
Cost vs. human wage	claude-sonnet-5	2/5	Specialized robotic grippers and vision systems for handling fragile, variably shaped plants are expensive to develop and deploy relative to low-wage manual labor typically used for this task.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed commercial product reliably performs the complete task of tying, bunching, wrapping roots, and boxing plants at scale. Research prototypes exist for narrow components (e.g., robotic packing), but production systems handling the full workflow with live plants are not in widespread operational use.
Technical feasibility today	claude-sonnet-5	1/5	No deployed products handle flower/plant bunching and root-wrapping in production; agricultural robotics for delicate handling of irregular plant material is still largely research or pilot stage.

Load agricultural products into trucks, and drive trucks to market or storage facilities.

CI 14–35 · exposure 20 · augmentation 25 · importance 3.3/5 · 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	Agriculture remains a low-digitization, geographically dispersed sector with high operational variability. Autonomous adoption in farming is in early pilot stages; most crop transport relies on human-driven trucks and relies on cost stability of migrant labor.
Sector adoption velocity	claude-sonnet-5	1/5	Agriculture is one of the least digitized, most physically-oriented sectors, with minimal AI/robotics adoption for field logistics tasks like this.
Augmentation potential	claude-haiku-4-5-20251001	2/5	Current AI and automation tools offer minimal augmentation for the core task of loading and transport; telematics and route optimization help marginally, but do not meaningfully amplify worker productivity in the loading or driving itself.
Augmentation potential	claude-sonnet-5	2/5	Some route-planning or logistics software can assist scheduling deliveries, but the physical loading and driving itself receives little direct AI augmentation today.
Task automatability	claude-haiku-4-5-20251001	2/5	While autonomous trucks exist in research and limited deployment, loading agricultural products into trucks remains a largely manual, dexterous task requiring adaptation to varied product types, fragile items, and spatial reasoning. End-to-end automation (loading + driving) is not yet reliably performable at scale with current deployed systems, though the driving portion has seen incremental progress.
Task automatability	claude-sonnet-5	2/5	Loading is manual physical labor requiring dexterity in variable outdoor/greenhouse conditions, and driving to market involves navigating unstructured rural roads and facility logistics; current AI/robotics cannot do this end-to-end reliably.dolor
Adoption barriers	claude-haiku-4-5-20251001	4/5	Significant barriers exist: autonomous trucks face regulatory approval gaps (especially in rural areas and mixed traffic), safety liability asymmetries, FMCSA oversight, insurance limitations, and farms' preference for flexible, low-capex labor. These create material friction against substitution.
Adoption barriers	claude-sonnet-5	2/5	A commercial driver's license is needed for truck driving, but there's no strict licensing barrier against automation itself; physical loading has minimal regulatory barriers though liability for road driving matters.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Autonomous vehicle systems and robotic loading remain capital-intensive; total cost of ownership and integration for a small-to-mid-scale farm operation still exceeds the wage burden of seasonal farmworkers, particularly given underutilization outside peak seasons.
Cost vs. human wage	claude-sonnet-5	1/5	Autonomous loading robotics and self-driving trucks for this use case would require far greater capital investment than current low-wage farm labor, making AI more expensive all-in.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Autonomous truck projects exist in controlled settings (highways, mining), but full-task feasibility for this job—including farm loading logistics, last-mile delivery, and varied agricultural cargo handling—lacks demonstrated production-scale reliability in real farm-to-market workflows.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product autonomously loads farm produce and drives trucks to market today; self-driving trucks remain in limited highway pilots, not farm-to-market operations with loading.

Plant, spray, weed, fertilize, water, and prune plants, shrubs, and trees, using gardening tools.

CI 15–29 · exposure 13 · augmentation 25 · 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	Agricultural automation, especially for labor-intensive nursery and crop work, remains in pilot/early-adoption phases in most regions. Small and medium farms dominate the sector and lack capital and digitization for rapid AI/robotic deployment; adoption is laggard relative to information and finance sectors.
Sector adoption velocity	claude-sonnet-5	1/5	Agriculture, especially crop and nursery labor, remains one of the least digitized and automated sectors, with physical robotics adoption still in pilot phases.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI and robotics offer limited real-time assistance to field laborers today. Computer vision systems could identify pests or disease, but integration into handheld tools remains rare and unproven at scale, so meaningful in-task augmentation is minimal.
Augmentation potential	claude-sonnet-5	2/5	Some precision agriculture tools (sensors, irrigation scheduling apps, pest detection via imaging) can assist decision-making, but they don't meaningfully augment the physical execution of planting, weeding, or pruning.
Task automatability	claude-haiku-4-5-20251001	2/5	While some individual operations (spraying, watering) could be partially automated with robotics, the full task requires dexterous manipulation, environmental adaptation, and judgment (e.g., identifying disease, selective pruning). Current robots lack the reliability and generalization across diverse plants and conditions needed for ≥50% time savings at equal quality.
Task automatability	claude-sonnet-5	1/5	This is a physical, multi-step manual task requiring dexterity, mobility across uneven terrain, and fine motor manipulation of plants and tools that current AI systems cannot perform end-to-end today.
Adoption barriers	claude-haiku-4-5-20251001	2/5	Few regulatory or licensing barriers specifically restrict automation of these tasks, but adoption is hindered by technical immaturity, high upfront capital, and the practical difficulty of retrofitting systems across diverse crop types and farm layouts.
Adoption barriers	claude-sonnet-5	2/5	There are few legal or licensing barriers, but practical barriers exist such as variable field conditions, plant fragility, and the need for adaptable dexterity that hardware struggles with.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	Agricultural robotics remain capital-intensive with high integration costs, maintenance, and limited utilization rates. For seasonal or small-to-medium operations typical of crop and nursery work, the all-in cost per task-equivalent exceeds the loaded wage of farmworkers.
Cost vs. human wage	claude-sonnet-5	1/5	Specialized agricultural robots and automation systems require high capital investment, maintenance, and are far more expensive per task-equivalent than low-wage manual farm labor in most contexts.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Some narrow agricultural robots exist for specific tasks (e.g., harvesting, weeding), but no deployed product reliably performs the full suite of planting, spraying, weeding, fertilizing, watering, and pruning with consistent quality across diverse crop types and nursery environments at commercial scale.
Technical feasibility today	claude-sonnet-5	1/5	While some agricultural robots exist for narrow subtasks (e.g., precision spraying or automated pruning in trials), there is no deployed product that performs the full range of planting, weeding, fertilizing, watering, and pruning reliably at scale.

Sell and deliver plants and flowers to customers.

CI 5–35 · exposure 13 · augmentation 38 · 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	1/5	Agricultural and nursery sectors have low digital infrastructure and digitization rates; AI adoption in this domain remains minimal, with no evidence of meaningful production-level automation of sales and delivery tasks.
Sector adoption velocity	claude-sonnet-5	2/5	Agriculture and nursery/retail sectors are low-digitization, physical-labor-intensive industries with slow AI and automation adoption compared to information or finance sectors.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI could assist with minor tasks like inventory tracking or customer order management, but provides limited augmentation for the core activities of physical delivery and face-to-face sales interaction.
Augmentation potential	claude-sonnet-5	3/5	AI can assist with online storefronts, inventory management, customer inquiries, and route optimization for delivery, meaningfully aiding parts of the sales and logistics process even though physical tasks remain human-performed.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires physical delivery, face-to-face customer interaction, and judgment about plant selection and condition—capabilities well beyond current AI systems. No meaningful automation is possible for the core delivery and sales components.
Task automatability	claude-sonnet-5	2/5	Physical delivery and in-person selling of plants require manual handling, driving, and customer interaction that current AI systems cannot perform end-to-end.While an app or chatbot could handle order-taking, the core sell-and-deliver task remains largely physical and human-executed.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Customer preference for human interaction when purchasing plants, legal liability for product condition upon delivery, and the requirement for physical presence at customer locations create substantial barriers to automation.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement for selling plants, but physical delivery logistics, customer relationship preferences, and handling of perishable goods create moderate practical friction against full automation.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	AI systems cannot physically deliver products or engage in face-to-face sales, making any cost comparison moot. The human labor cost is substantially lower than any hypothetical AI solution involving robotics and logistics.
Cost vs. human wage	claude-sonnet-5	2/5	AI can reduce costs for order processing and customer service, but physical delivery still requires human labor, vehicles, and logistics, keeping overall costs comparable to or only marginally cheaper than human-based delivery.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed product can execute end-to-end sales and physical delivery of plants and flowers. This requires embodied presence, relationship management, and real-time decision-making that current AI cannot reliably perform in production.
Technical feasibility today	claude-sonnet-5	2/5	E-commerce platforms exist for plant sales but actual delivery and physical handoff still rely on human drivers and salespeople; no deployed AI product autonomously sells and delivers plants.

Repair and maintain farm vehicles, implements, and mechanical equipment.

CI 5–35 · exposure 13 · augmentation 50 · 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	2/5	Agricultural automation adoption remains uneven and concentrated in large-scale operations; small and mid-size farms (where most farmworkers labor) have low digital infrastructure and slow adoption of AI diagnostics relative to information and finance sectors.
Sector adoption velocity	claude-sonnet-5	1/5	Agriculture and manual equipment maintenance are low-digitization, low-AI-adoption sectors with minimal production deployment of automation for physical repair tasks.
Augmentation potential	claude-haiku-4-5-20251001	3/5	AI-assisted diagnostics and maintenance scheduling can help farmworkers identify problems faster and optimize repair decisions, but the assistance is partial and requires human validation and hands-on execution in most practical scenarios.
Augmentation potential	claude-sonnet-5	3/5	AI can assist via diagnostic guidance, repair manuals, parts lookup, and troubleshooting chatbots, but the hands-on repair work itself remains unaided by AI.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI vision systems can diagnose some mechanical faults and generate repair instructions, the task requires physical manipulation of equipment in variable outdoor conditions, hands-on troubleshooting, and judgment about equipment lifespan versus replacement—most of which remains beyond current automation.
Task automatability	claude-sonnet-5	1/5	This requires physical dexterity, diagnostic judgment, and manual manipulation of diverse mechanical/hydraulic/electrical systems that current AI cannot perform end-to-end without a human body doing the work.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Farm equipment repair involves liability concerns for safety-critical systems (e.g., heavy machinery), manufacturer warranty restrictions, and often requires licensed technicians or engineers for certification—creating meaningful legal and organizational friction against full automation.
Adoption barriers	claude-sonnet-5	2/5	No licensing generally required for basic farm equipment repair, but physical access, liability for improper mechanical repairs, and equipment variability create practical friction.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	AI vision and diagnostic tools add cost and oversight overhead without replacing the bulk of labor (physical repair work, specialist knowledge); the all-in cost per repair episode would typically exceed the wage of a skilled farmworker performing the task.
Cost vs. human wage	claude-sonnet-5	1/5	There is no AI substitute performing the physical repair labor, so cost comparison favors the human by default since the AI cannot deliver the output alone.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	AI diagnostic tools exist (e.g., image-based fault detection), but production deployment for farm equipment maintenance is limited; most systems lack the robustness for field conditions and still require human technicians to execute repairs and validate diagnoses.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product autonomously repairs or maintains farm vehicles and equipment; at best AI provides diagnostic troubleshooting text, not physical repair.

Harvest fruits and vegetables by hand.

CI 10–29 · exposure 13 · 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	Agricultural automation adoption remains slow outside large commodity producers; most fruit and vegetable harvest (especially perishables requiring delicate handling) still relies on manual labor in small and mid-size farms, which dominate the sector in labor volume.
Sector adoption velocity	claude-sonnet-5	1/5	Agriculture is a notoriously slow-adopting, low-digitization sector for physical automation, with mechanized/robotic harvesting still in early pilot stages for most crops.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI-assisted ripeness detection or yield mapping can provide marginal support, but current systems offer limited real-time guidance that meaningfully raises harvester productivity; most augmentation value remains prospective rather than deployed at scale.
Augmentation potential	claude-sonnet-5	2/5	AI-assisted tools (e.g., yield prediction, ripeness detection apps) can guide human pickers on timing and location, but they don't materially transform the physical act of harvesting itself.
Task automatability	claude-haiku-4-5-20251001	2/5	Selective harvesting of delicate fruits and vegetables requires fine motor control and visual discrimination for ripeness/quality assessment. While crop detection and localization are possible, grasping without bruising and consistent ripeness judgment remain difficult for current robotic systems in unstructured field environments, and end-to-end automation does not yet achieve 50% time savings at equal quality.
Task automatability	claude-sonnet-5	1/5	Hand-harvesting requires fine motor dexterity, gentle handling of variable produce, and mobility across uneven terrain that current general-purpose AI/robotic systems cannot replicate at scale with equal quality or speed.
Adoption barriers	claude-haiku-4-5-20251001	3/5	No specific legal license is required for human harvest labor, but adoption faces organizational friction: infrastructure investment, seasonal labor patterns, technical skill gaps on farms, and established supply chains. Liability for crop damage and worker preference for familiar methods create moderate friction.
Adoption barriers	claude-sonnet-5	2/5	No licensing or legal requirement mandates human harvesters, but practical barriers like field variability, produce fragility, and lack of robust robotic dexterity create significant physical/organizational friction.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	Current robotic harvesting systems incur high capital costs, maintenance, and infrastructure integration, far exceeding the low per-unit cost of seasonal farm labor in most regions where this task dominates.
Cost vs. human wage	claude-sonnet-5	1/5	Specialized harvesting robots require expensive hardware, maintenance, and per-crop customization, making them far more costly than low-wage manual labor in most current contexts.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	Research prototypes and early-stage robotic harvesters exist (e.g., strawberry, apple picking bots), but most have narrow crop/condition scope, material damage rates, or low throughput. No production-scale deployment at broad farm scale has demonstrated reliable substitution; most systems remain in pilots or controlled greenhouse settings.
Technical feasibility today	claude-sonnet-5	1/5	Robotic harvesters exist only as narrow, crop-specific research or pilot deployments (e.g., strawberries, some orchard fruit) with high error rates and limited throughput; no general deployed product harvests diverse fruits/vegetables by hand-equivalent means.

Haul and spread topsoil, fertilizer, peat moss, and other materials to condition soil, using wheelbarrows or carts and shovels.

CI 15–20 · exposure 0 · augmentation 0 · 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	Agriculture remains a laggard sector for AI/robotics adoption due to low digitization, outdoor unpredictability, small-to-medium farm sizes, and capital constraints. Actual field deployment of autonomous hauling/spreading robots is virtually nonexistent.
Sector adoption velocity	claude-sonnet-5	1/5	Agriculture and manual farm labor are among the least digitized sectors with minimal AI/robotics adoption for basic soil-handling tasks.
Augmentation potential	claude-haiku-4-5-20251001	1/5	Current AI and robotics offer minimal assistance to farmworkers performing material hauling and spreading, as the task is primarily physical labor without a strong digital or decision-support component that AI could enhance.
Augmentation potential	claude-sonnet-5	1/5	AI offers essentially no assistance to a worker physically hauling and spreading soil materials with a wheelbarrow and shovel.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires physical manipulation of bulk materials across variable terrain and unstructured farm environments. Current AI/robotics lack the dexterity, environmental adaptation, and cost-effectiveness to replicate shovel work and material spreading at scale in real farm conditions.
Task automatability	claude-sonnet-5	1/5	This is a physical manual labor task requiring hauling and spreading bulk materials with hand tools; no current AI system can perform this physical work, though robotic equipment exists separately from AI reasoning systems.
Adoption barriers	claude-haiku-4-5-20251001	2/5	The primary barrier is technical (no feasible automation exists), not regulatory. Once automation becomes viable, few legal or licensing obstacles would prevent adoption, though some customers may prefer human labor for quality assurance.
Adoption barriers	claude-sonnet-5	1/5	There are no licensing, legal, or human-contact barriers preventing mechanization of this task; adoption is limited purely by technical and economic feasibility, not regulation.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	Specialized agricultural robots capable of this work (if they existed in production) would cost orders of magnitude more than the hourly wage of farmworkers, particularly when factoring in integration, maintenance, and site-specific setup.
Cost vs. human wage	claude-sonnet-5	1/5	Manual labor with wheelbarrows and shovels is cheap relative to any robotic or AI-driven system capable of physical material handling, making AI substitution currently far more expensive.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed commercial products reliably perform the full task of hauling and spreading soil amendments independently. Specialized agricultural robots exist only in narrow, controlled research settings, not in production farming operations.
Technical feasibility today	claude-sonnet-5	1/5	No deployed AI product performs manual soil-conditioning labor; any automation would require specialized agricultural robotics (not general AI), which remains largely research-stage or niche for this specific task.

Direct and monitor the work of casual and seasonal help during planting and harvesting.

CI 5–28 · exposure 13 · augmentation 25 · 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	1/5	Agricultural labor remains among the least digitized sectors, with small farms and seasonal workforce churn limiting investment in AI management infrastructure. Adoption of even basic digital systems is slow outside large agribusiness operations.
Sector adoption velocity	claude-sonnet-5	1/5	Agriculture, especially crop/nursery/greenhouse labor management, is a low-digitization sector with minimal AI adoption for direct worker supervision.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI dashboards showing work progress or safety alerts could assist a supervisor, but the core task—directing and motivating workers, making real-time decisions about task allocation and safety—still depends on human judgment and presence in the field.
Augmentation potential	claude-sonnet-5	2/5	Some scheduling, workforce tracking, or task-assignment software can assist supervisors, but direct monitoring and interpersonal direction of workers sees limited AI augmentation today.
Task automatability	claude-haiku-4-5-20251001	2/5	While AI could monitor work via computer vision and flag task completion, directing diverse human workers in real-time requires dynamic judgment, safety decisions, and interpersonal adaptation that current systems cannot reliably handle end-to-end. Significant manual oversight would remain necessary.
Task automatability	claude-sonnet-5	1/5	Directing and monitoring seasonal field workers requires physical presence, real-time judgment about worker performance, crop conditions, and interpersonal management that current AI systems cannot perform end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Health and safety liability is severe in agricultural work; injury attribution and worker welfare decisions must legally rest with a responsible human overseer. Regulatory expectations and worker compensation frameworks create strong practical barriers to full automation.
Adoption barriers	claude-sonnet-5	3/5	No licensing requirement exists, but the task involves managing people, safety, and labor relations, creating organizational and practical friction against automation.
Cost vs. human wage	claude-haiku-4-5-20251001	2/5	Camera infrastructure, cloud processing, and human oversight for safety and exception handling remain costlier than a single supervisor's wage, especially given the low labor costs in agricultural contexts and the redundancy required for liability.
Cost vs. human wage	claude-sonnet-5	1/5	There is no viable AI substitute performing this supervisory task, so any AI cost is not comparable to a human all-in cost for this specific function.
Technical feasibility today	claude-haiku-4-5-20251001	2/5	No production systems today reliably manage the full supervision, task allocation, and real-time direction of farm crews. Computer vision monitoring exists in research, but integrated worker management products are not deployed at scale in agricultural settings.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product supervises or directs human agricultural laborers in the field; this remains firmly in the human supervisory domain.

Dig, cut, and transplant seedlings, cuttings, trees, and shrubs.

CI 15–15 · exposure 0 · augmentation 0 · 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	Nursery and greenhouse operations remain highly labor-dependent with minimal AI or automation adoption; these are traditionally low-digitization sectors where manual work persists due to cost economics and seasonal labor availability.
Sector adoption velocity	claude-sonnet-5	1/5	Agricultural manual labor sectors are among the slowest to adopt AI/robotic automation due to low digitization, physical terrain variability, and cost structures favoring human labor.
Augmentation potential	claude-haiku-4-5-20251001	1/5	Current AI systems offer negligible assistance for the core manual operations of digging, cutting, and transplanting; no deployed tools meaningfully augment human productivity in these specific physical tasks.
Augmentation potential	claude-sonnet-5	1/5	Current AI offers minimal direct assistance to a worker physically digging, cutting, and transplanting plants; there is no meaningful software-based productivity boost for this hands-on physical task.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires precise manual dexterity, dynamic environmental adaptation, and careful handling of delicate biological material in outdoor conditions. Current AI and robotics lack the fine motor control, real-time sensory feedback, and dexterity needed to dig, cut, and transplant at production scale and quality parity with humans.
Task automatability	claude-sonnet-5	1/5	This requires physical manipulation of plants and soil in variable outdoor/greenhouse conditions—current AI systems (software/LLMs) cannot perform physical digging, cutting, or transplanting; robotic solutions are not generally available off-the-shelf for this varied task.
Adoption barriers	claude-haiku-4-5-20251001	2/5	While there are no strict licensing barriers for automation itself, the work occurs in rural, low-capital-intensity sectors with high labor supply and low substitution pressure. Physical constraints and biological variability present practical friction, though not legal prohibition.
Adoption barriers	claude-sonnet-5	2/5	No licensing or legal requirement mandates a human perform this task, but physical/environmental variability, capital costs, and lack of mature robotic solutions create practical friction to substitution.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	Specialized horticultural robots capable of this task remain prohibitively expensive (hundreds of thousands to millions), with long amortization periods and high maintenance costs, far exceeding the loaded wage of seasonal farmworkers in most regions.
Cost vs. human wage	claude-sonnet-5	1/5	Specialized robotic planting/transplanting equipment is expensive, requires significant capital and maintenance, and is not cost-competitive with low-wage manual labor for this variable, dexterity-intensive task.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	While robotics research has explored horticultural automation, no deployed commercial systems reliably perform the full sequence of digging, cutting, and transplanting seedlings in real greenhouse or field environments at scale. Existing prototypes remain research-stage with high failure rates on fragile materials.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product performs general transplanting of seedlings, cuttings, trees, and shrubs at production scale; agricultural robotics for planting exist only in narrow, research or pilot contexts for specific crops.

Dig, rake, and screen soil, filling cold frames and hot beds in preparation for planting.

CI 15–15 · exposure 0 · augmentation 0 · 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	1/5	Agricultural sectors, especially small and mid-sized nurseries and farms, show slow adoption of automation for soil-preparation tasks; most operations remain labor-intensive and are concentrated in low-digitization, price-sensitive segments.
Sector adoption velocity	claude-sonnet-5	1/5	Crop and nursery labor is a low-digitization, physically intensive sector with minimal AI/robotics adoption for manual soil preparation tasks.
Augmentation potential	claude-haiku-4-5-20251001	1/5	AI offers no meaningful assistance to workers performing manual digging, raking, and screening of soil; this task has no information-processing or decision component where AI tools could enhance productivity.
Augmentation potential	claude-sonnet-5	1/5	Current AI tools offer essentially no assistance for the physical acts of digging, raking, and screening soil in this context.
Task automatability	claude-haiku-4-5-20251001	1/5	This task requires physical manipulation of soil, rakes, and screening equipment in outdoor/greenhouse environments. Current AI systems cannot perform the embodied, dexterous work of digging, raking, and screening soil at scale or speed, and no deployed robotic systems reliably handle the variability of soil composition and cold frame/hot bed preparation.
Task automatability	claude-sonnet-5	1/5	This is a physical manual labor task requiring dexterous handling of soil, tools, and containers in variable outdoor/greenhouse conditions; no off-the-shelf AI system can perform this physical work end-to-end today.
Adoption barriers	claude-haiku-4-5-20251001	2/5	Few formal licensing barriers exist for soil preparation, but physical work on distributed farm sites, weather dependence, and the need for real-time site adaptation create practical friction against full automation, though these are not hard regulatory blocks.
Adoption barriers	claude-sonnet-5	2/5	No licensing or legal barriers exist, but physical environment variability (uneven soil, mixed debris, delicate seedbed prep) creates practical friction against automation adoption.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	The capital cost of specialized agricultural robots, their maintenance, software integration, and the need for site-specific customization far exceed the loaded wage of seasonal farmworkers performing this labor-intensive task.
Cost vs. human wage	claude-sonnet-5	1/5	Any robotic system capable of this task would require expensive specialized hardware far exceeding the cost of low-wage manual farm labor for equivalent output.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed products reliably perform this full sequence of physical soil preparation tasks in agricultural or nursery settings. While some experimental agricultural robots exist, none demonstrably handle the coordinated digging, raking, and screening workflow in production environments.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product performs digging, raking, and screening soil into cold frames; agricultural robotics for such fine manual prep work remains research-stage or highly narrow prototypes.

Move containerized shrubs, plants, and trees, using wheelbarrows or tractors.

CI 15–15 · exposure 0 · augmentation 13 · importance 3.2/5 · click for rater detail

Panel ratings per dimension for this task
Dimension	Model	Rating	Rationale
Sector adoption velocity	claude-haiku-4-5-20251001	1/5	Agricultural automation adoption remains slow, especially in small-to-medium farms and nurseries. Containerized plant movement is labor-intensive but not yet widely targeted by commercial automation vendors at scale.
Sector adoption velocity	claude-sonnet-5	1/5	Agriculture and nursery work is a low-digitization, physically demanding sector with minimal AI/robotics adoption for manual material handling tasks.
Augmentation potential	claude-haiku-4-5-20251001	2/5	AI offers limited augmentation for this physical task; existing tools like GPS guidance on tractors provide modest assistance, but the core movement and handling work remains primarily manual.
Augmentation potential	claude-sonnet-5	1/5	Current AI offers essentially no assistance for the physical act of moving containerized plants using wheelbarrows or tractors.
Task automatability	claude-haiku-4-5-20251001	1/5	Moving containerized plants requires physical manipulation in unstructured outdoor/greenhouse environments with variable terrain, weight distribution, and fragile botanical material. Current AI lacks embodied robotics capable of handling this safely and reliably.
Task automatability	claude-sonnet-5	1/5	This requires physical manipulation of heavy, variably-shaped objects across uneven outdoor/nursery terrain, which current AI systems (software-based) cannot perform; robotic solutions for this remain experimental, not deployable at scale.'
Adoption barriers	claude-haiku-4-5-20251001	2/5	Minimal legal barriers exist, but significant practical barriers include the need for robust hardware, outdoor operation in variable conditions, and integration with existing farm workflows.
Adoption barriers	claude-sonnet-5	2/5	No licensing or legal requirement mandates a human for this task, but physical environment variability, liability for damaging plants/equipment, and lack of infrastructure create moderate practical friction.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	The capital cost of agricultural robotics capable of this task far exceeds the loaded hourly wage of farmworkers, making the all-in cost per task substantially higher than human labor.
Cost vs. human wage	claude-sonnet-5	1/5	Any robotic alternative would require expensive specialized hardware, sensors, and maintenance far exceeding the low wage cost of manual laborers performing this task.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed products perform this task reliably today. Specialized agricultural robots for transplanting exist in research, but none handle containerized shrub/tree movement at production scale with consistent quality.
Technical feasibility today	claude-sonnet-5	1/5	No mature product performs general containerized plant moving in nurseries today; autonomous ag-robots exist for narrow tasks like row navigation but not general wheelbarrow/tractor plant transport.

Repair farm buildings, fences, and other structures.

CI 5–15 · exposure 0 · augmentation 25 · importance 2.9/5 · 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	Agricultural sectors, particularly crop and nursery operations, have low automation adoption overall and remain capital-constrained with high preference for flexible, adaptable human labor. Farm infrastructure repair is a low-priority automation target.
Sector adoption velocity	claude-sonnet-5	1/5	Agriculture is among the least digitized, most manual-labor-dependent sectors, with negligible robotic adoption for structural repair tasks.
Augmentation potential	claude-haiku-4-5-20251001	2/5	While AI-powered inspection tools (drone imagery, damage classification) could assist planning and documentation, current systems offer minimal real-time assistance during the actual hands-on repair work. Augmentation potential is modest compared to the manual dexterity required.
Augmentation potential	claude-sonnet-5	2/5	AI can help via diagnostic apps, repair instructions, or planning material needs, but offers little direct assistance during the physical repair work itself.
Task automatability	claude-haiku-4-5-20251001	1/5	Farm building and fence repair requires physical manipulation of materials, structural assessment in varied outdoor conditions, and real-time problem-solving that current AI systems cannot perform autonomously. The task demands embodied action in unstructured environments beyond what mobile robots can reliably execute today.
Task automatability	claude-sonnet-5	1/5	Structural repair requires physical manipulation of materials, tools, and improvisation in variable outdoor conditions that no current AI system can perform end-to-end.
Adoption barriers	claude-haiku-4-5-20251001	4/5	Structural repair and building maintenance often require licensed contractors in many jurisdictions, and liability for faulty repairs creates strong legal and insurance barriers to unsupervised automation. Liability costs for failed repairs are asymmetric and substantial.
Adoption barriers	claude-sonnet-5	2/5	No licensing requirement blocks automation, but the physical unstructured environment (varied terrain, materials, unpredictable damage) is itself a major practical barrier to any automated system.
Cost vs. human wage	claude-haiku-4-5-20251001	1/5	Autonomous repair systems would require custom hardware, sensing, and integration far exceeding the cost of a farmworker's hourly wage. Current automation in construction/maintenance is expensive and limited to highly structured environments.
Cost vs. human wage	claude-sonnet-5	1/5	There is no viable AI-driven robotic solution for this task, so any hypothetical system would be far more expensive than a human laborer with basic tools.
Technical feasibility today	claude-haiku-4-5-20251001	1/5	No deployed products perform end-to-end farm structure repair autonomously. While vision systems can inspect damage, actual repair—measuring, cutting, fastening, adjusting for site-specific conditions—remains purely manual and human-dependent.
Technical feasibility today	claude-sonnet-5	1/5	No deployed product autonomously repairs fences or farm buildings; robotic construction/repair remains research-stage and limited to controlled, narrow demonstrations.

Related occupations — Farming, Fishing & Forestry

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.