Stockers and Order Fillers
53-7065.00Median wage $37,330/yr2,833,810 employed (US)Rank #100 of 923 scored · top 11% by substitution
Receive, store, and issue merchandise, materials, equipment, and other items from stockroom, warehouse, or storage yard to fill shelves, racks, tables, or customers' orders. May operate power equipment to fill orders. May mark prices on merchandise and set up sales displays.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution47
Exposure40
Augmentation49
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
30 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
17%
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%41
panel mean rating 2.6/5 → substitution pressure 41/100
Technical feasibility todayw 20%39
panel mean rating 2.6/5 → substitution pressure 39/100
Cost vs. human wagew 15%39
panel mean rating 2.5/5 → substitution pressure 39/100
Adoption barriersw 20%inverted — strong barriers lower the score72
panel mean rating 2.1/5 (barrier strength) → substitution pressure 72/100
Sector adoption velocityw 10%42
panel mean rating 2.7/5 → substitution pressure 42/100
Task breakdown (30 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.
Keep records of out-going orders.
99CI 97–100 · exposure 100 · augmentation 75 · importance 4.2/5 · click for rater detail
Keep records of out-going orders.
99| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Warehouse, logistics, and retail sectors have extensively adopted automated order tracking systems over the past decade; manual order logging is now rare in digitized operations. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Warehousing and logistics have rapidly adopted digital inventory and order-tracking systems, though smaller operations may still rely on manual or semi-manual methods. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | While the core task is automated, AI-powered systems can assist warehouse workers by predicting demand, suggesting batch optimizations, or flagging shipping errors before dispatch, meaningfully raising productivity in the broader fulfillment workflow. |
| Augmentation potential | claude-sonnet-5 | 4/5 | Even where humans remain involved in fulfillment, AI-backed scanning and tracking systems significantly speed up and reduce errors in maintaining outgoing order records. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Recording outgoing orders is a purely data-entry and logging task that modern systems handle automatically—scanning barcodes, pulling data from inventory databases, and creating logs requires no human judgment and achieves >50% time savings at equal quality with standard warehouse management systems. |
| Task automatability | claude-sonnet-5 | 5/5 | Recording outgoing orders is a structured data-entry task easily handled by warehouse management systems, barcode/RFID scanning, and inventory software with automatic logging, meeting the time-saving threshold with off-the-shelf tools. |
| Adoption barriers | claude-haiku-4-5-20251001 | 1/5 | There are no licensing, regulatory, or legal requirements that mandate a human record outgoing orders; the task is purely transactional with no liability asymmetry preventing automation. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or human-contact requirements exist for order record-keeping; it's a purely administrative/logistics function already widely automated. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | The cost of barcode scanners, RFID readers, or simple API integrations is orders of magnitude cheaper than paying a human wage to manually record each outgoing order. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated scanning and logging systems cost pennies per transaction compared to manual record-keeping labor, representing an order-of-magnitude cost reduction. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Barcode scanning, RFID tracking, and automated warehouse management systems (WMS) are deployed at scale across retail and logistics companies today, reliably logging outgoing orders in production environments. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | WMS and ERP systems (e.g., SAP, NetSuite, warehouse scanners) already perform automated order-record-keeping reliably in production across large-scale logistics operations. |
Compute prices of items or groups of items.
97CI 95–100 · exposure 100 · augmentation 63 · importance 4.4/5 · click for rater detail
Compute prices of items or groups of items.
97| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Price computation automation is already deeply embedded in retail, warehousing, and logistics sectors through POS systems and automated ordering platforms. Adoption is mature and widespread across the industry, not experimental. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | Retail and warehousing sectors have near-universal adoption of automated pricing/POS systems, representing one of the most mature and pervasive automation use cases. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI systems assist workers by providing instant, accurate price lookups and promotional pricing logic, allowing stockers to focus on physical placement and inventory accuracy rather than manual calculation. This augmentation is well-established and productivity-enhancing. |
| Augmentation potential | claude-sonnet-5 | 3/5 | For workers who still perform manual price calculations or verification, calculators and mobile scanning apps provide moderate assistance, though the task itself is largely already automated rather than augmented. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Computing prices of items or groups of items is a purely computational task that current AI and systems can perform end-to-end with minimal human intervention. Barcode scanning, database lookups, and price calculation are fully automatable with existing point-of-sale and inventory systems, easily exceeding 50% time savings. |
| Task automatability | claude-sonnet-5 | 5/5 | Computing prices is a simple arithmetic/lookup task easily handled by POS systems, barcode scanners, and software calculators with full reliability and speed advantage over manual computation. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While some organizations may require human verification for certain transactions or regulatory compliance (e.g., age-restricted items), most retail and warehousing operations have already automated price computation without legal restriction. Adoption barriers are minimal, though some customer-facing contexts may prefer human oversight. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or regulatory barriers prevent automated price computation; it is already standard practice in nearly all retail environments. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | The cost of system inference and integration for price computation is negligible compared to the loaded wage of a human performing the same calculation. Once systems are in place, marginal cost per transaction is near zero. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated price computation via scanners/software costs fractions of a cent per transaction compared to any human time spent calculating prices manually. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Deployed point-of-sale systems, barcode scanners, and inventory management software already reliably perform price computation at scale in retail and warehousing environments worldwide. This capability is mature, production-ready, and used daily across millions of transactions. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | POS and inventory systems already compute prices automatically and reliably at massive scale in virtually every retail and warehouse operation today. |
Read orders to ascertain catalog numbers, sizes, colors, and quantities of merchandise.
84CI 79–89 · exposure 80 · augmentation 63 · importance 4.0/5 · click for rater detail
Read orders to ascertain catalog numbers, sizes, colors, and quantities of merchandise.
84| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Warehouse and logistics sectors have rapidly adopted robotic process automation (RPA) and intelligent order processing systems over the past 5+ years, driven by e-commerce pressure and labor scarcity; deployment is now common in larger and midsize operations. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Warehousing and logistics have rapidly adopted automated order processing, barcode/RFID scanning, and WMS software as part of broader e-commerce fulfillment digitization. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by pre-populating order fields, flagging ambiguities, and suggesting corrections, meaningfully speeding up human verification in cases where orders are complex or require judgment; the human remains in the loop for validation. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-assisted order scanning and digital pick-lists significantly speed up and reduce errors in the human worker's task of identifying order details before fulfillment. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current OCR and data extraction systems can reliably parse orders (digital or scanned) to extract catalog numbers, sizes, colors, and quantities with high accuracy. A workflow combining document parsing with structured data extraction achieves >50% time savings with minimal human oversight, though some edge cases (handwritten orders, ambiguous formats) may require fallback. |
| Task automatability | claude-sonnet-5 | 4/5 | Parsing order text/documents to extract catalog numbers, sizes, colors, and quantities is a straightforward information extraction task well within current OCR/NLP and multimodal AI capabilities., though integration with warehouse systems is needed. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory or licensing barriers exist for automating order reading. Main friction is legacy system integration and warehouse management system compatibility, not legal or liability constraints; adoption remains largely voluntary and economically driven. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or regulatory barriers prevent automating this data-reading subtask; it's a routine clerical/logistics function. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | End-to-end cost per order (OCR + API calls + minimal oversight) is substantially lower than human time at warehouse wages; cloud-based extraction and automation reduce per-unit friction to a fraction of loaded labor cost. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated order parsing via software is extremely cheap compared to manual reading and transcription by a human worker, especially at scale in distribution centers. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Mature products including intelligent document processing (e.g., UiPath, Automation Anywhere, cloud-native solutions) and barcode/SKU scanning systems are deployed at scale in logistics and warehouse operations, reliably extracting order details in production environments. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Warehouse management systems and OCR-based order processing tools already extract structured order data reliably in production, though edge cases (handwritten orders, unusual formats) still require occasional human correction. |
Itemize and total customer merchandise selection at checkout counter, using cash register, and accept cash or charge card for purchases.
79CI 75–84 · exposure 80 · augmentation 50 · importance 4.1/5 · click for rater detail
Itemize and total customer merchandise selection at checkout counter, using cash register, and accept cash or charge card for purchases.
79| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Self-checkout and automated POS systems are in rapid, broad deployment across retail—major chains (Target, Walmart, Kroger) have scaled these substantially. Adoption is especially deep in large, digitized retail environments where this task is prevalent. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Retail has adopted self-checkout and automated POS technology broadly and rapidly over the past two decades, representing one of the more mature automation deployments in a traditionally lower-digitization sector. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | When cashiers remain, modern POS systems assist by automating lookup, calculation, and payment routing, reducing manual effort. However, the task itself lends to full replacement, so augmentation is secondary to automatability rather than a primary value driver. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-assisted registers, computer vision for item recognition, and fraud/theft detection tools help human cashiers work faster and more accurately, though the task itself is largely being automated rather than augmented. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Modern point-of-sale systems and self-checkout kiosks already automate 80%+ of itemization, scanning, and payment processing. Computer vision can identify products and a human or fully automated system can complete the transaction, meeting the ≥50% time-saving threshold with minimal setup. |
| Task automatability | claude-sonnet-5 | 4/5 | Self-checkout kiosks and automated POS/scanner systems already perform itemization, totaling, and payment processing with minimal human involvement in many retail settings, though some human backup remains for exceptions. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While some regulatory oversight of payment systems exists (PCI-DSS for card handling), there are no hard legal barriers preventing automated checkout. Customer preference for human cashiers and shrinkage/fraud concerns create moderate adoption friction, but not legal requirements. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for handling checkout, though some friction exists from theft/loss-prevention concerns, customer preference for human checkout, and age-verification requirements for restricted items like alcohol. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | A self-checkout kiosk or integrated POS system costs far less per transaction than a full-time cashier wage; the infrastructure cost amortizes across hundreds of daily transactions, making AI-driven checkout orders of magnitude cheaper per task. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Self-checkout hardware/software amortized over high transaction volumes is substantially cheaper per transaction than a cashier's wage, though upfront capital and monitoring staff reduce the savings somewhat. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Self-checkout systems with barcode scanning and automated payment processing are deployed at scale in thousands of retail locations today. Mature POS software reliably performs itemization, totaling, and card/cash payment acceptance in production environments. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Self-checkout and automated payment systems are deployed at scale in grocery, big-box, and convenience retail today, though loss-prevention issues and exception handling still require staff oversight. |
Complete order receipts.
79CI 75–82 · exposure 75 · augmentation 75 · importance 4.7/5 · click for rater detail
Complete order receipts.
79| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Warehouse and logistics sectors show rapid, deep adoption of automated order fulfillment and receipt systems. Major retailers, 3PLs, and e-commerce companies have deployed barcode scanning and WMS automation at scale, making this one of the most automated tasks in the supply chain. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Retail, warehousing, and logistics have rapidly adopted barcode/RFID scanning and automated inventory systems over the past decade, making this one of the more digitized physical-sector tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-assisted order receipt (real-time error flagging, item mismatch alerts, automated exception routing) significantly augments human productivity and accuracy when humans remain in the loop for validation and problem-solving. |
| Augmentation potential | claude-sonnet-5 | 4/5 | Even where full automation isn't complete, scanning devices, mobile apps, and AI-assisted data entry substantially speed up and reduce errors in the human process of completing receipts. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Order receipt completion involves structured data entry, matching, and verification—tasks that current AI systems can automate with high accuracy. Scanning barcodes, matching items to orders, and recording quantities are routine processes where AI can achieve ≥50% time savings, though some human oversight and exception-handling remain common in real deployments. |
| Task automatability | claude-sonnet-5 | 4/5 | Completing order receipts (recording items received/filled, quantities, matching to orders) is largely structured data entry that barcode scanning, warehouse management systems, and OCR/AI-driven inventory software can already handle with minimal human input for most transactions. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory or legal barriers prevent automation of order receipt completion. Physical infrastructure investments and integration with existing WMS are required, but these are organizational/technical friction rather than hard regulatory or liability blockers. |
| Adoption barriers | claude-sonnet-5 | 1/5 | There is no licensing, legal sign-off, or human-contact requirement for completing order receipts; it's a routine clerical/operational task with minimal regulatory protection. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Barcode scanning, OCR, and automated data entry are extremely cost-effective compared to warehouse labor. Once infrastructure (scanners, WMS software) is amortized, per-receipt cost is a fraction of manual entry by a human stocker, easily achieving 5–10× cost advantage. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated scanning and inventory software cost a small fraction per transaction compared to a human loaded wage for manual receipt completion, though initial system integration adds some cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed warehouse management systems (WMS) and barcode-scanning automation already perform order receipt functions reliably in production at scale across retail, logistics, and e-commerce. Error rates on routine receipt completion are low, though edge cases (damaged goods, mismatches) still often require human judgment. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Warehouse management systems, RFID/barcode scanning, and automated receipt reconciliation software are deployed at scale in retail and logistics today, reliably generating and completing order receipts in production environments. |
Receive and count stock items, and record data manually or on computer.
65CI 55–75 · exposure 62 · augmentation 63 · importance 4.0/5 · click for rater detail
Receive and count stock items, and record data manually or on computer.
65| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Large retailers, e-commerce fulfillment centers, and logistics providers have widely adopted automated scanning and inventory systems in production; smaller operations lag but sector-wide adoption is deep and accelerating. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Warehousing and retail have adopted barcode/RFID and WMS technology substantially, though many smaller operations still rely on manual counting and paper-based tracking. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-assisted inventory systems (exception detection, anomaly flagging) help human stockers prioritize problem items and improve accuracy, though humans remain responsible for final verification and resolution in most current deployments. |
| Augmentation potential | claude-sonnet-5 | 4/5 | Handheld scanners, mobile apps, and automated data entry significantly speed up and improve accuracy of counting and recording while a human remains in the loop for physical handling. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Robotic systems and computer vision can reliably count and identify stock items; data entry to inventory systems is fully automatable. The manual counting and recording components meet the 50% time-saving threshold with current warehouse automation and barcode/RFID scanning systems. |
| Task automatability | claude-sonnet-5 | 3/5 | Counting and recording stock can be substantially automated via barcode/RFID scanning and warehouse management systems, but physical receiving and handling of items still requires human or robotic manipulation not covered by 'AI' alone. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | No licensing or regulatory barriers prevent automation; warehouses routinely deploy these systems. Minor friction exists around initial system integration and legacy process changeover, but no legal requirement mandates human performance. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but physical warehouse environments, existing infrastructure investments, and need for exception-handling by humans create moderate organizational friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Barcode scanners, RFID systems, and inventory management software have dropped significantly in cost and require minimal ongoing human oversight, making the all-in cost well below a human worker's loaded wage for equivalent throughput. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Scanning/RFID hardware plus software licensing costs are moderate; while cheaper per-unit at scale, upfront integration and hardware costs keep this roughly comparable to labor cost in many smaller operations. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed warehouse automation (barcode scanners, RFID readers, automated conveyor systems) demonstrably perform counting and data recording in production at scale across major retailers and logistics firms, though some manual verification steps may remain. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Barcode scanners, RFID systems, and WMS software are widely deployed and reliably handle data recording, but full automation of physical counting/receiving still often involves manual scanning or human verification. |
Compare merchandise invoices to items actually received to ensure that shipments are correct.
65CI 55–75 · exposure 62 · augmentation 75 · importance 3.4/5 · click for rater detail
Compare merchandise invoices to items actually received to ensure that shipments are correct.
65| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Retail, logistics, and warehouse operations—core sectors for this task—are rapidly deploying automated receiving and inventory systems. Smart warehouses and e-commerce fulfillment centers already use vision-based verification at scale. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Retail and warehousing have moderate digitization with barcode/RFID systems increasingly common, but many smaller stockrooms and distribution points still perform this manually, giving a middling adoption pace overall. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI assists significantly by pre-flagging discrepancies and auto-matching correct items, allowing stockers to focus on exceptions and problem-solving. This raises productivity substantially while keeping humans in oversight roles. |
| Augmentation potential | claude-sonnet-5 | 4/5 | Scanning apps, mobile devices with OCR, and inventory management software significantly speed up the comparison process and flag discrepancies, meaningfully boosting worker productivity while a human still confirms final accuracy. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Comparing invoices to received items is highly structured and repeatable work; computer vision can identify items, OCR can extract invoice data, and rule-based matching can verify correctness. With proper setup, this task achieves >50% time savings with equal or better accuracy than manual checking. |
| Task automatability | claude-sonnet-5 | 3/5 | Matching invoice data to received items is a structured data-comparison task well-suited to automation, but physically verifying items received still requires scanning/counting hardware or human action, limiting full end-to-end automation with off-the-shelf AI alone. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory or legal barriers exist; no licensed professional signature is required. Some organizations prefer human oversight for high-value shipments or retain manual spot-checks, but nothing prevents automation from being the default workflow. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement and minimal regulatory barrier exists, though organizational friction around integrating with existing inventory systems and error-cost concerns for shipment discrepancies create some resistance. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Computer vision and OCR inference costs are now very low; integration into existing warehouse systems is straightforward. The cost per invoice-check is orders of magnitude cheaper than paying a human to manually compare items, especially at scale. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Automated scanning and reconciliation systems can be cheaper per transaction at scale, but require capital investment in scanners/software plus integration, making costs roughly comparable to human labor for smaller operations. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Production systems combining OCR, barcode scanning, and inventory matching exist in warehouses today (e.g., AI-powered receiving platforms, automated scanning workflows). Some manual verification remains but the core comparison is reliably automated in mature deployments. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Barcode/RFID scanning systems and warehouse management software with automated invoice reconciliation exist and are deployed, but many smaller operations still rely on manual counting and matching, so reliability varies by scale of investment. |
Stamp, attach, or change price tags on merchandise, referring to price list.
55CI 44–66 · exposure 53 · augmentation 38 · importance 4.1/5 · click for rater detail
Stamp, attach, or change price tags on merchandise, referring to price list.
55| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption is concentrated in large-scale retail, logistics, and e-commerce warehouses; small and mid-size retailers and physical stores still rely heavily on manual labor. Overall penetration remains moderate, with automation growing but not yet dominant. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Retail and warehousing are moderate-to-slow adopters of physical automation compared to purely digital sectors; some large chains use electronic shelf labels, but most order fillers still manually tag merchandise. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI and automation offer minimal augmentation to human stockers performing this task; it is either replaced or manual. Unlike tasks where AI assists judgment, price-tag application is purely mechanical and offers little room for human-AI collaboration. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Mobile apps and handheld scanners with digital price lookup meaningfully speed up the process of finding and applying correct prices, though the physical tagging action itself still requires the worker. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Current computer vision and robotic systems can identify merchandise, read price lists, and apply or update price tags with high reliability. This is a largely mechanical, repetitive task with clear visual inputs and outputs; modern vision systems and label applicators can perform most of this end-to-end, though handling diverse item geometries adds complexity. |
| Task automatability | claude-sonnet-5 | 3/5 | Referring to a price list and applying tags is a simple, repeatable data-lookup-and-physical-action task; automated pricing/labeling systems (electronic shelf labels, robotic tagging) can handle much of it, but physical merchandise handling still requires setup or robotics not universally deployed. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | There are no legal or licensing barriers to automating price-tag application. The main friction is capital investment, workflow integration, and retailer preference for human flexibility; these are business choices rather than hard regulatory blocks. |
| Adoption barriers | claude-sonnet-5 | 1/5 | There is no licensing, legal, or safety barrier to automating price tagging; it's purely a matter of cost and physical infrastructure. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | Automated labeling equipment has high upfront and maintenance costs, but per-unit operational cost is low once deployed. The total cost-of-ownership is roughly comparable to human labor in large-scale operations; smaller retailers may find it uneconomical. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Electronic shelf labeling or robotic tagging systems require significant capital investment (hardware, integration) that often exceeds the cost of low-wage manual labor for this simple task, especially in smaller retail settings. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Robotic systems and automated label applicators exist in production environments (e.g., warehouses, distribution centers), but they require setup and work best on standardized items; they struggle with irregular shapes, fragile goods, or dense shelving. Deployed solutions show material limitations in scope and adaptability. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Electronic shelf labels and automated pricing systems exist and are used in some large retailers, but manual tag stamping/attaching on physical merchandise is still overwhelmingly done by humans in most stores; robotic solutions remain narrow and costly. |
Obtain merchandise from bins or shelves.
51CI 35–66 · exposure 45 · augmentation 50 · importance 4.1/5 · click for rater detail
Obtain merchandise from bins or shelves.
51| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | E-commerce, logistics, and large retail chains are rapidly deploying warehouse automation and robotic picking systems; adoption is accelerating in high-throughput facilities and is measured in production deployments across the sector. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Adoption is concentrated in large-scale e-commerce fulfillment centers; most retail and warehouse operations still rely heavily on manual picking, reflecting slower diffusion in physical, low-digitization settings. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-powered picking assistants (bin locators, item recognition overlays, route optimization) meaningfully assist human stockers by reducing search time and optimizing workflows, though the human remains the primary executor in many settings. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-driven inventory systems, pick-path optimization, and handheld scanning tools meaningfully improve worker efficiency in locating and retrieving items, even though the physical retrieval itself remains human-performed. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Robotic picking systems and autonomous mobile manipulators can retrieve items from bins and shelves with ≥50% time savings in controlled warehouse environments. Human oversight for exception handling remains, but the core retrieval task is largely automatable with existing technology. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical retrieval of items from bins or shelves requires mobile manipulation and navigation in unstructured environments, which current general-purpose AI/robotics cannot yet do reliably at scale or cost-effectively for most retail settings. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | No licensing requirement or legal mandate for human labor exists; barriers are primarily organizational (capital expenditure, system integration friction, union labor agreements in some facilities) rather than regulatory or liability-based. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal barriers prevent automation, but physical infrastructure constraints, safety requirements around human-robot coexistence, and variable environments create real friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | Robotic systems have high upfront capital costs and ongoing maintenance, offsetting labor savings in some contexts; at scale and for high-velocity items, automation becomes cost-competitive with human labor, but sensitivity to facility layout and SKU variability keeps the ratio near parity. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Automated picking systems require substantial capital investment in robotics and facility redesign, often exceeding the cost of human labor except in very high-volume, specialized fulfillment centers. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Deployed warehouse automation (e.g., collaborative robots, automated storage systems) handles this task in production at major retailers and logistics firms, but performance varies with item fragility, bin density, and environmental chaos; error rates and integration friction remain material. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Warehouse robotics (e.g., Amazon's Kiva/robotic arms, AutoStore) perform limited picking in highly structured, purpose-built facilities, but general shelf/bin picking across varied SKUs and store layouts remains largely unsolved in deployed products. |
Issue or distribute materials, products, parts, and supplies to customers or coworkers, based on information from incoming requisitions.
49CI 35–62 · exposure 42 · augmentation 63 · importance 4.4/5 · click for rater detail
Issue or distribute materials, products, parts, and supplies to customers or coworkers, based on information from incoming requisitions.
49| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Large-scale adoption is rapid in e-commerce, logistics, and retail (Amazon, Walmart, DHL). Smaller firms and less-digitized sectors lag, but the highest-volume fulfillment operations have already implemented or are actively deploying automation. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Warehousing and retail sectors are adopting automation unevenly and slowly compared to information/professional services, with robotics adoption concentrated in a few large players like Amazon. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | Inventory systems, real-time picking guidance, barcode scanning, and robotic assist significantly amplify human pickers' throughput and accuracy. Humans remain central but with AI/automation handling search, routing, and verification tasks in parallel. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Software systems (WMS, barcode scanning, pick-list optimization) meaningfully assist workers in locating and tracking materials, improving speed and accuracy even though the physical task remains human-performed. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | Order matching and routing can be automated (scanning requisitions, identifying inventory locations), but physical picking, packing, and final distribution to varied locations still requires human intervention. Partial automation achieves 30–40% time savings with current warehouse management systems. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical retrieval and distribution of materials requires manipulation and navigation in unstructured warehouse/store environments, which current general-purpose AI cannot do end-to-end without robotics; software can automate the requisition-matching logic but not the physical handoff. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory or licensing barriers; most friction is capital availability and operational complexity of integrating systems into existing warehouses. No legal requirement for human sign-off on order fulfillment, enabling substitution. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but physical infrastructure constraints, liability for misdelivered goods, and existing facility layouts create meaningful adoption friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | Capital-intensive warehouse automation (conveyors, sorters, pick-to-light systems) costs tens of millions per facility with long integration cycles. For individual requisition processing and picking, deployed systems are cost-comparable to human labor after amortization, not substantially cheaper. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Robotic fulfillment systems require large capital investment in hardware, maintenance, and facility redesign, often costing more than human labor unless at very high volume/scale. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Warehouse automation systems, sortation robotics, and automated picking aids are deployed at scale in major logistics and retail operations. Error rates remain material for complex orders, but production use is widespread and mature for high-volume standardized tasks. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Automated storage/retrieval systems and warehouse robots exist in some large distribution centers, but most stocking/order-filling in retail and smaller warehouses is still manual; deployed full automation is narrow and capital-intensive. |
Keep records on the use or damage of stock or stock-handling equipment.
47CI 35–60 · exposure 38 · augmentation 63 · importance 3.8/5 · click for rater detail
Keep records on the use or damage of stock or stock-handling equipment.
47| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | While warehouse automation is growing, damage/wear logging remains largely manual across small and mid-sized operations; adoption of autonomous damage tracking is limited to large retailers experimenting with computer vision, not mainstream practice. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Warehousing and logistics sectors are adopting automation and digital tracking at a moderate pace, with large players deploying at scale while smaller operations lag, reflecting middling overall adoption. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-assisted forms, automated photo logging with classification suggestions, or mobile apps that pre-populate damage records based on images could meaningfully speed up the stocker's record-keeping task while the human retains judgment on liability and severity. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI and digital systems significantly ease and speed up record-keeping tasks—auto-populating logs, flagging anomalies, and generating reports—while workers still verify and handle physical damage assessment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Recording straightforward inventory damage or equipment usage could be partially automated via image recognition or form-filling, but requires judgment about damage severity and causation that current AI handles inconsistently. Setup overhead and need for human verification limit meaningful time savings below the 50% threshold. |
| Task automatability | claude-sonnet-5 | 3/5 | Record-keeping of stock/equipment use or damage is largely structured data entry that AI-enabled systems (barcode/RFID scanning, IoT sensors, forms with OCR) can capture and log automatically, but flagging and describing damage still often requires human observation and judgment. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Minimal regulatory barriers exist for automating damage records, but organizational friction and the need for human judgment on liability/warranty claims create practical friction; most firms rely on human observation rather than automated logging systems. |
| Adoption barriers | claude-sonnet-5 | 1/5 | There are no licensing or legal requirements mandating a human perform this record-keeping; it's a routine administrative task with minimal regulatory barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI systems for damage detection and record-keeping (vision APIs, warehouse software) still require human oversight, setup, and hardware integration, making per-task costs comparable to or exceeding the time cost of manual entry by a stocker. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Digital record-keeping systems reduce labor cost per record substantially, but hardware, integration, and sensor deployment costs keep overall cost roughly comparable to manual logging in many facilities, especially smaller ones. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Computer vision for damage detection exists in research/demo form, but deployed warehouse systems typically require manual entry or structured forms rather than reliable end-to-end automation. Existing warehouse management systems handle status updates but not autonomous damage assessment at production scale. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Warehouse management systems with automated inventory tracking and damage-reporting apps are deployed in many operations, but many smaller or lower-tech operations still rely on manual logs, and damage assessment automation is narrower in scope. |
Store items in an orderly and accessible manner in warehouses, tool rooms, supply rooms, or other areas.
43CI 35–51 · exposure 38 · augmentation 63 · importance 3.9/5 · click for rater detail
Store items in an orderly and accessible manner in warehouses, tool rooms, supply rooms, or other areas.
43| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Major e-commerce and logistics firms are actively piloting and deploying robotic stocking solutions, but adoption remains concentrated in large, digitized facilities. Smaller warehouses and supply rooms continue relying on manual stocking, indicating uneven and middling sector-wide velocity. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Warehousing and logistics have seen automation investment (e.g., Amazon robotics) but overall sector adoption of full automation for stocking/placement remains limited to large players; broad industry adoption is slow. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-powered picking assists (visual guidance, inventory routing, real-time bin location) significantly boost human stocker productivity and accuracy. Augmentation via computer vision and automated routing is increasingly deployed while humans remain responsible for placement decisions. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-assisted inventory systems, barcode/RFID scanning, and optimized placement software help workers decide where to store items and track locations, improving efficiency while humans still perform physical placement. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | Robotic systems and automated guided vehicles can handle portions of stocking and order fulfillment in controlled warehouse environments (e.g., bin placement, retrieval), but require significant setup and integration with inventory systems. Full end-to-end automation with 50% time savings at equal quality remains challenging for unstructured or mixed environments with variable item types and placements. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical placement and organization of items requires manipulation, navigation, and spatial reasoning that current general-purpose AI cannot perform end-to-end; robotic solutions exist only in narrow, highly structured settings. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few licensing or regulatory barriers exist for warehouse automation. Primary friction comes from organizational deployment complexity, customer expectations for human oversight, and need for human intervention in edge cases or unstructured inventory areas. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal barriers, but physical infrastructure, safety requirements around forklifts/heavy items, and the need for adaptable manual dexterity create moderate practical friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Automation requires substantial capital investment in robotics, software, and warehouse infrastructure integration, plus ongoing maintenance. For modest-wage stocking roles in smaller facilities, total cost of automation often exceeds the human labor cost, though large-scale facilities approach parity. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Robotic storage systems require significant capital investment, integration, and maintenance, making them costlier than human labor except at very large scale operations with high throughput. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Deployed robotic stocking systems (e.g., collaborative robots, autonomous mobile robots) work in real warehouses but typically require tailored setup and operate within constrained parameters. Error rates and scope limitations mean they augment rather than fully replace human stockers in production today. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Automated storage and retrieval systems (AS/RS) and warehouse robots exist in some large distribution centers, but they are narrow-purpose, expensive, and not generally deployed for tool rooms, supply rooms, or varied warehouse layouts. |
Mark stock items, using identification tags, stamps, electric marking tools, or other labeling equipment.
43CI 35–51 · exposure 33 · augmentation 38 · importance 3.9/5 · click for rater detail
Mark stock items, using identification tags, stamps, electric marking tools, or other labeling equipment.
43| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Large-scale distribution centers and e-commerce fulfillment centers increasingly deploy automated marking systems, but adoption remains concentrated in high-volume, capital-intensive settings; smaller logistics and retail operations lag significantly. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Warehousing and retail stocking are physical, lower-digitization sectors where robotic/automation adoption for granular tasks like manual marking is still limited and slow-moving compared to information sectors. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | Handheld electric marking tools and vision-assisted placement systems already augment worker productivity, and emerging pick-and-place robots with integrated marking capabilities could further assist human packers and stockers in identifying and labeling items more quickly. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with generating labels, printing barcodes, or suggesting tagging via handheld scanners/software, but it doesn't materially transform the physical marking action performed by the worker. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | Automated systems can mark items in controlled warehouse environments, but the task requires physical manipulation, vision-based item recognition, and precise placement—current robots handle structured, high-volume repetition but struggle with variability in item size, shape, and fragility, achieving partial automation rather than full replacement. |
| Task automatability | claude-sonnet-5 | 2/5 | This is a physical labeling task requiring manipulation of items, tags, and tools in a warehouse; current AI systems cannot perform the physical marking action itself, though robotic labeling systems exist in narrow contexts.dispatch |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | There is no legal requirement that a human must perform marking, and no strong licensing barriers, but organizational inertia, existing manual workflows, and the cost of robotic infrastructure create moderate friction to substitution. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, legal, or human-contact requirement restricts this task; it's a purely operational physical task with minimal regulatory barriers. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Automated marking equipment requires significant capital investment, maintenance, and integration costs that often exceed the loaded wage of warehouse workers performing marking manually, especially for smaller retailers or non-high-volume operations. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Robotic or automated labeling systems require significant capital investment in fixed equipment, which is often costlier than flexible human labor for variable stock unless at very high volume with standardized items. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some warehouse automation systems perform marking in specialized contexts (e.g., barcode application in large distribution centers), but widespread reliable deployment remains limited; most inventory marking still relies on human workers with handheld equipment due to cost and edge-case complexity. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Automated labeling/marking exists in some highly structured manufacturing lines, but general warehouse stock marking with mixed items and tools is not reliably handled by deployed general-purpose AI or robotics products today. |
Take inventory or examine merchandise to identify items to be reordered or replenished.
42CI 35–49 · exposure 38 · augmentation 63 · importance 3.9/5 · click for rater detail
Take inventory or examine merchandise to identify items to be reordered or replenished.
42| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption is slow outside large, capital-intensive retailers and warehouses; small and mid-market retail remains heavily manual. Even large chains pilot rather than deploy full automation at scale due to complexity, cost, and the need for human floor presence for other duties. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Retail and warehouse sectors show slow, uneven adoption of physical inventory automation compared to fast-digitizing white-collar sectors; RFID and vision systems remain niche outside large distribution centers. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | Mobile inventory apps with barcode scanning, AI-assisted demand forecasting, and automated alerts for low stock levels meaningfully amplify human worker productivity and reduce manual counting time. This assistive layer is widely deployed and transforms efficiency while the worker remains in control of replenishment decisions. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-powered inventory management software, demand forecasting, and mobile scanning apps meaningfully assist workers in identifying items to reorder, even though the physical task remains human-performed. |
| Task automatability | claude-haiku-4-5-20251001 | 3/5 | AI vision systems can identify items and stock levels via cameras or integration with inventory systems, achieving partial automation of the counting and identification component. However, the decision logic for reorder thresholds and replenishment rules often requires human judgment about seasonal demand, supplier lead times, and business priorities, limiting full end-to-end automation to roughly half the task. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical inventory counting and visual inspection of merchandise on shelves requires physical presence and manipulation that current AI systems cannot perform end-to-end without robotic hardware.','Some sub-steps (barcode scanning, data logging) are automatable but the core physical task is not. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | There are modest barriers: retailers value human presence on the sales floor for customer service, and many smaller retailers lack the digitization infrastructure to support automated inventory systems. Liability for stock-outs falls on management, reducing legal barriers, but organizational inertia and existing labor contracts create some friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but physical store environments, existing labor structures, and capital costs for automation hardware create moderate organizational friction against full replacement. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Computer vision and automated counting systems have substantial upfront capital costs (hardware, software licensing, integration) and require ongoing maintenance and human oversight. For a low-wage task like stocking, the all-in cost per inventory check often exceeds the loaded wage of the human worker, making the ratio unfavorable. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Robotic/vision inventory systems require significant capital investment (hardware, sensors, integration) that often exceeds the cost of hourly stocker wages for equivalent coverage, especially in smaller retail settings. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Computer vision systems and barcode-scanning robots exist and are deployed in some warehouses and retail environments, but error rates remain material for complex SKU environments and real-world occlusion. Most implementations require significant setup, human verification, and fallback oversight rather than operating fully autonomously at production scale. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some retailers deploy computer-vision shelf-scanning robots or smart shelf sensors, but these are narrow, expensive pilots not broadly deployed across the occupation's typical work environments. |
Provide assistance or direction to other stockroom, warehouse, or storage yard workers.
41CI 13–70 · exposure 33 · augmentation 50 · importance 3.3/5 · click for rater detail
Provide assistance or direction to other stockroom, warehouse, or storage yard workers.
41| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Large logistics, e-commerce, and manufacturing firms have been actively deploying automated task management, routing, and AI-driven monitoring systems for several years, with measurable displacement of coordinator roles. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Warehousing is a physically-oriented, moderately digitized sector with growing use of software for logistics but slow adoption of AI for direct people-management tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI tools assist human supervisors by offering real-time analytics, predictive bottleneck detection, and automated task suggestions, substantially raising productivity while keeping humans in decision-making roles. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI tools like scheduling software, task-assignment apps, or communication platforms can help coordinate workers, but they don't materially transform this interpersonal directing task. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI systems can manage task assignment, route optimization, and performance monitoring in real-time, delivering >50% time savings compared to manual oversight, though complex interpersonal conflict resolution may still require human judgment in edge cases. |
| Task automatability | claude-sonnet-5 | 1/5 | This is interpersonal supervision and coordination among physical workers in a warehouse setting, requiring real-time physical presence, judgment, and social interaction that current AI cannot replicate end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While some organizations prefer human supervisors for morale and complex judgment, there are no licensing or regulatory mandates requiring a human to direct warehouse workers, and adoption friction is organizational rather than legal. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement, but organizational and physical-presence friction is significant since directing people on a warehouse floor requires trust, accountability, and situational awareness. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Automated task dispatch and coordination via existing WMS infrastructure costs a fraction of a full-time supervisor or lead worker's loaded wage, delivering substantial cost-per-task advantage. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this supervisory/assistance function, so cost comparison favors the human by default. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Warehouse management systems with AI-powered task scheduling and worker monitoring are in production use, but reliable real-time guidance and dynamic reassignment still depend on camera integration and safety validation that varies across deployments. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product provides direction or assistance to human warehouse workers in the way a lead worker or supervisor does; this remains outside current AI product capabilities. |
Requisition merchandise from supplier, based on available space, merchandise on hand, customer demand, or advertised specials.
39CI 32–46 · exposure 30 · augmentation 75 · importance 3.6/5 · click for rater detail
Requisition merchandise from supplier, based on available space, merchandise on hand, customer demand, or advertised specials.
39| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Large retailers use demand forecasting and automated ordering for high-volume items, but most small to mid-market retailers rely on manual or semi-automated processes. Adoption is uneven and pilots are more common than full replacement systems. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Retail and warehousing have adopted demand forecasting and automated replenishment tools at a moderate pace, with large chains further ahead than small independent stores, reflecting a middling industry-wide adoption pattern. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI-driven demand forecasting and inventory analytics can substantially improve a stocker's or order filler's decisions about what and how much to requisition, flagging fast movers and slow stock while the human retains final judgment. This represents genuine productivity gain. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-driven demand forecasting and inventory analytics significantly help workers decide what and how much to requisition, improving speed and accuracy while humans still verify physical space and finalize orders. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can analyze inventory data, demand patterns, and supplier information, the task requires judgment about available space, seasonal specials, and real-time business priorities that vary by store. Current systems can assist with recommendations but cannot end-to-end requisition merchandise reliably enough to meet the 50% time-saving bar without substantial human oversight. |
| Task automatability | claude-sonnet-5 | 2/5 | AI/inventory-management systems can generate reorder suggestions from data, but the task as stated involves synthesizing physical space, on-hand stock checks, and judgment calls about promotions that still require human verification and physical presence.dictionaries |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Retailers maintain strong operational control over inventory decisions due to supplier contracts, merchandising strategies, and customer-facing consequences of stockouts. However, no legal requirement mandates human sign-off, creating moderate friction rather than hard barriers. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates a human perform requisitioning, but organizational trust, supplier relationships, and inventory accountability create moderate friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Requisitioning involves API integration with supplier systems, demand forecasting models, and ongoing maintenance. The total cost competes with or exceeds the hourly wage of a stocker or order filler, especially when accounting for integration and oversight. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Inventory software licensing and integration costs are moderate, and while forecasting algorithms are cheap to run, the physical verification of shelf space and stock still requires paid labor, keeping overall costs roughly comparable to human-driven processes. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Inventory management and demand forecasting tools exist, but reliable end-to-end requisitioning requires integration with multiple legacy systems, real-time stock verification, and supplier communication. No mature product reliably handles the full task across diverse retail contexts at scale. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Automated replenishment and demand-forecasting systems are widely deployed in retail (e.g., Walmart, Kroger), but these typically generate recommendations that a human reviews or approves rather than fully autonomous requisitioning tied to physical space constraints. |
Pack and unpack items to be stocked on shelves in stockrooms, warehouses, or storage yards.
36CI 35–38 · exposure 25 · augmentation 38 · importance 3.9/5 · click for rater detail
Pack and unpack items to be stocked on shelves in stockrooms, warehouses, or storage yards.
36| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Large logistics and e-commerce companies (Amazon, etc.) have deployed significant warehouse automation, but adoption remains uneven. Many smaller retailers and regional warehouses continue relying heavily on human stockers. Pilots are common but production adoption is concentrated in high-volume, capital-rich sectors. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Warehousing has growing but still uneven robotics adoption; most facilities remain human-labor dominant with automation concentrated in large e-commerce operations. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI/robotics can assist with item location and shelf placement optimization, but current systems offer limited real-time assistance to humans performing the physical task. Augmentation here is mainly in planning and coordination rather than in-task support. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-driven inventory systems, scanners, and route optimization tools help workers locate and organize stock more efficiently, though the physical packing/unpacking remains manual. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Packing and unpacking items requires physical manipulation, dexterity, and spatial reasoning that current robotic systems struggle with consistently across diverse item types, weights, and packaging. While some specialized automation exists in controlled warehouse environments, general-purpose automation for this task at scale with equal quality remains infeasible. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical picking, packing, and shelving of varied items requires dexterity, mobility, and adaptability to irregular objects that current robotics cannot yet handle reliably or cheaply at scale.'}, |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While there are no hard licensing barriers, organizational friction is moderate: existing warehouse layouts favor human workers, safety liability concerns slow adoption, and smaller retailers lack capital for expensive robotic systems. Physical space constraints and legacy operations create friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirements, but physical infrastructure changes, safety regulations for robots near workers, and variable item handling create real operational friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Warehouse automation systems are capital-intensive and require infrastructure investment, maintenance, and integration overhead that often exceeds the wages of multiple order fillers, especially in small to mid-size operations or less-structured facilities. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Robotic picking/packing systems require significant capital investment, integration, and maintenance that often exceeds low-wage warehouse labor costs for general tasks. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Robotic pickers and palletizers exist in narrow domains (uniform boxes, structured environments), but they require significant setup and fail regularly on irregular shapes, fragile items, or crowded spaces. No mainstream deployed product handles the full range of packing/unpacking tasks reliably without human oversight. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Warehouse robots (e.g., piece-picking arms, AMRs) exist in limited deployments but still struggle with diverse SKUs, packaging, and unstructured stockroom environments, requiring heavy human backup. |
Stock shelves, racks, cases, bins, and tables with new or transferred merchandise.
35CI 35–35 · exposure 25 · augmentation 25 · importance 4.2/5 · click for rater detail
Stock shelves, racks, cases, bins, and tables with new or transferred merchandise.
35| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Retail and warehouse sectors show slow adoption of stocking automation relative to information/finance; most retail stores still rely on human stockers, and pilots remain limited to large chains and controlled warehouse environments. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Retail and warehousing are adopting automation unevenly; large fulfillment centers deploy robotics but most retail stores still rely heavily on manual stocking with slow uptake of physical automation. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI/robotics offer minimal assistance to human stockers on this task today; computer vision for inventory tracking can help with planning, but the core physical stocking action itself sees little augmentation in deployed systems. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with inventory tracking, restock alerts, and optimized placement suggestions, but it does not materially transform the physical act of stocking itself. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Current robotic systems can handle simple repetitive stocking in controlled environments (e.g., warehouses), but real-world shelf stocking requires navigating varied store layouts, handling fragile items, managing limited shelf space, and avoiding damage—capabilities that remain unreliable and constrained to very narrow scenarios today. |
| Task automatability | claude-sonnet-5 | 2/5 | Physically placing diverse merchandise on shelves, racks, bins, and tables requires manipulation skills, spatial judgment, and mobility that current general-purpose robots cannot yet perform reliably across varied retail environments. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Minimal regulatory or licensing barriers exist for automation; the main friction is organizational inertia, capital requirements, and customer familiarity with human stockers rather than legal or liability constraints. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates human stockers, but physical environment variability, safety considerations around moving equipment, and customer-facing store environments create some practical friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Robotics hardware, software integration, and ongoing maintenance remain expensive; the loaded cost per shelf-stocking task still exceeds the wage of low-cost human stockers in most retail settings, especially accounting for setup and oversight overhead. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Robotic stocking systems require significant capital investment, maintenance, and integration, generally not cheaper than low-wage stocking labor except in very high-volume specialized warehouse contexts. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | A few pilot robot systems exist (e.g., shelf-scanning, simple bin-picking in warehouses), but no mature product reliably performs full end-to-end retail shelf stocking in production across diverse retail environments; most deployments are narrow or research-stage. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some automated shelf-stocking and warehouse robots exist (e.g., in Amazon fulfillment centers) but these are narrow, capital-intensive deployments, not general products handling arbitrary retail/warehouse stocking tasks. |
Determine proper storage methods, identification, and stock location, based on turnover, environmental factors, and physical capabilities of facilities.
35CI 32–38 · exposure 25 · augmentation 75 · importance 3.7/5 · click for rater detail
Determine proper storage methods, identification, and stock location, based on turnover, environmental factors, and physical capabilities of facilities.
35| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Warehouse and logistics sectors are actively piloting AI-driven inventory and storage optimization tools, but production-scale autonomous deployment remains limited; most adoption occurs in larger, digitally mature operations rather than across the sector broadly. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Warehousing and logistics have moderate digitization with WMS and slotting tools increasingly adopted, but full AI-driven decision-making for storage location strategy is still in pilot/hybrid stages in most facilities. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI systems can meaningfully augment human decision-making by analyzing turnover data, suggesting optimal locations based on environmental factors, and flagging facility constraints, enabling warehouse staff to make faster, more data-informed storage decisions while retaining final oversight and judgment. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-driven slotting analytics and demand forecasting tools significantly help human planners optimize storage decisions by processing turnover data and constraints faster than manual analysis, even though humans still finalize decisions. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | AI can assist with data-driven analysis of turnover rates and environmental factors to recommend storage locations, but the task requires real-time physical assessments of facility layout, condition, and spatial constraints that current AI systems cannot perform autonomously without human verification. The decision-making involves complex spatial reasoning and facility-specific knowledge that falls short of the 50% time-saving bar for end-to-end automation. |
| Task automatability | claude-sonnet-5 | 2/5 | This involves physical facility assessment, spatial reasoning about warehouse layout, and judgment about turnover patterns integrated with physical constraints, which current AI cannot execute end-to-end without heavy human involvement in the physical environment. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | While there are no strict legal licensing barriers, facility managers and warehouse supervisors typically retain responsibility for storage decisions due to liability concerns around misplacement, damage, or compliance with safety regulations. Organizational friction around changes to established storage protocols also moderates substitution. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but organizational friction exists since these decisions affect safety, workflow, and require physical facility knowledge that's hard to fully digitize and trust to an algorithm alone. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Implementing AI systems for storage optimization requires substantial upfront investment in data infrastructure, integration with legacy WMS platforms, and ongoing training, making the all-in cost comparable to or exceeding the loaded wage of warehouse workers who perform this task with existing systems. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Slotting optimization software has upfront and ongoing licensing/integration costs and still requires human oversight and validation, making it not dramatically cheaper than experienced staff making these calls, especially for smaller operations. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed products reliably perform this entire task autonomously; existing warehouse management systems require significant human input for storage decisions and facility assessments. AI-based recommendation systems exist but are narrow in scope and typically serve as decision-support tools rather than autonomous decision-makers in production environments. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Warehouse management systems and slotting optimization software exist and are deployed, but the actual determination often still requires human judgment integrating physical facility quirks, so full autonomous decision-making is not yet standard. |
Answer customers' questions about merchandise and advise customers on merchandise selection.
34CI 30–38 · exposure 25 · augmentation 63 · importance 4.6/5 · click for rater detail
Answer customers' questions about merchandise and advise customers on merchandise selection.
34| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Large retailers have deployed conversational AI and recommendation systems in pilots and limited production, but adoption is mixed and often supplementary rather than replacement. Smaller retailers lag significantly; the task remains largely human-driven across much of the sector. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Retail is a moderately digitizing sector with self-checkout and chatbot pilots, but frontline stocker/order filler roles see slow AI adoption due to physical task requirements. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI product databases, recommendation engines, and search tools meaningfully assist human stockers and customer service staff in providing faster, more informed advice. These systems enhance productivity and knowledge access while keeping humans in the advisory loop. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI tools like inventory lookup apps and product information systems can help workers quickly answer customer questions, improving accuracy and speed while the human remains the primary interface. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can answer basic, templated merchandise questions via chatbots, the task requires contextual product knowledge, customer preference understanding, and personalized advice that current systems handle inconsistently. Full end-to-end automation with 50% time savings at equal quality is not reliably achievable today. |
| Task automatability | claude-sonnet-5 | 2/5 | AI chatbots can answer basic product questions but in-store customer interaction combining physical navigation, real-time inventory awareness, and personalized advice remains largely human-dependent.It requires physical presence and situational judgment that current AI can't replicate end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Customer preference for human interaction, liability concerns around product recommendations, and retailer brand reputation create meaningful friction. However, no strict legal requirement mandates human advice, allowing gradual automation where organizational risk tolerance permits. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirements, but customer preference for human interaction and the integrated nature of the job with physical tasks creates moderate organizational friction against pure automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Chatbot infrastructure and integration costs are moderate, but oversight and error correction by humans is necessary, especially for advice that affects purchase decisions. The total cost per interaction remains comparable to or higher than a human stocker for quality-equivalent outcomes. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI chat costs are low, this task is bundled with physical stocking work performed by a low-wage worker, so replacing just the advisory component doesn't yield major cost savings without also automating the physical labor. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Deployed chatbots and FAQ systems exist in retail, but they struggle with nuanced customer inquiries, complex product comparisons, and genuine advice. Real-world deployments have material error rates and narrow scope; humans remain essential for non-trivial customer interactions. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Retail chatbots and kiosks exist for online product Q&A, but reliable in-person merchandise advice combined with physical stocking duties is not deployed at scale in production. |
Pack customer purchases in bags or cartons.
33CI 31–35 · exposure 25 · augmentation 13 · importance 3.9/5 · click for rater detail
Pack customer purchases in bags or cartons.
33| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of packing automation is slow outside large logistics hubs; most retail and local fulfillment remain human-staffed. Pilots exist but production deployment is limited, and small retailers show minimal AI adoption in this task. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Retail and warehouse fulfillment are adopting robotic picking/packing at a measured pace, mostly in large-scale distribution centers, but adoption in customer-facing retail packing remains slow.rn |
| Augmentation potential | claude-haiku-4-5-20251001 | 1/5 | AI provides minimal assistance to humans packing bags or cartons; the task is primarily manual dexterity and physical judgment with limited opportunities for algorithmic augmentation while the human remains in the loop. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI/robotics can assist with packing optimization (e.g., box-size selection, item sequencing) in warehouse settings, but there is little direct augmentation for a worker manually bagging customer purchases at a register.rn |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While some structured packing in controlled environments (e.g., warehouses with standardized items) could be partially automated with robotics, general customer purchase packing requires variable item handling, fragile-goods awareness, and bag/carton selection that current AI and robotics struggle with at scale. End-to-end automation meeting the 50% time-saving threshold remains limited to narrow scenarios. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical bagging/packing requires dexterous manipulation of varied item shapes and sizes, which remains difficult for current robotics to do reliably at retail speed and cost.tp Some automated bagging systems exist in narrow contexts (e.g., automated checkout bagging lanes) but are not general-purpose.rn |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Physical retail and fulfillment environments have some technical barriers (capital cost, integration complexity), but no licensing or hard legal requirement mandates human involvement. Customer preference for human checkout/bagging and organizational inertia provide moderate friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates human packing, but physical infrastructure changes, item variability, and customer-facing environments create moderate practical friction to automation.rn |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The upfront capital cost of robotic packing systems (hardware, installation, integration) significantly exceeds the loaded wage of entry-level stockers in most settings, making the all-in cost per task substantially higher than human labor. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Robotic packing systems capable of handling diverse, irregular items require significant capital investment (grippers, vision systems, integration) that often exceeds the cost of low-wage human labor for this task.rn |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Robotics exists for specialized packing (e.g., uniform boxes in fulfillment centers), but no deployed products reliably pack arbitrary customer purchases into bags/cartons across general retail contexts. Current systems have high error rates on item orientation, fragility detection, and carton selection. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | A few automated packing/bagging systems exist in warehouses and some grocery self-checkout setups, but they are narrow, limited to certain item types, and not widely deployed for general customer purchase packing.rn |
Recommend disposal of excess, defective, or obsolete stock.
33CI 30–35 · exposure 25 · augmentation 50 · importance 3.6/5 · click for rater detail
Recommend disposal of excess, defective, or obsolete stock.
33| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Retail and warehouse sectors are digitizing inventory management, but disposal recommendations remain largely manual and human-driven. Adoption of AI-assisted tools is emerging in logistics but not yet deep or widespread in production systems across the industry. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Warehousing and retail stock operations are only moderately digitized, with inventory analytics tools deployed but full decision automation for disposal recommendations still rare and pilot-stage. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can usefully flag excess, defective, and obsolete items via inventory data analysis, helping humans prioritize and organize disposal workflows. However, the actual judgment and recommendation still rests with the human, making this assistive rather than transformative. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-based inventory analytics can meaningfully assist workers by highlighting aging, slow-moving, or anomalous stock, improving the speed and accuracy of human disposal decisions. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Recommending disposal requires evaluating product condition, market value, regulatory constraints, and financial impact—judgment-heavy decisions that AI cannot perform end-to-end reliably today. Current systems can flag excess inventory via data, but final disposal recommendations need human assessment of context and risk. |
| Task automatability | claude-sonnet-5 | 2/5 | This requires physical inspection of inventory condition combined with judgment calls informed by local context (space, demand shifts, damage), which current AI can only partially support via data analysis rather than end-to-end execution.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Some regulatory and financial accountability barriers exist (hazardous material disposal, tax implications, audit trails), but they apply to the decision itself rather than blocking automation completely. Organizations may prefer human sign-off for liability reasons, creating friction but not a hard legal requirement. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement exists, but disposal decisions often carry financial/loss accountability and may need managerial sign-off, creating some organizational friction against pure automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI inventory analysis tools exist but are typically used to support rather than replace the human recommendation process, making the cost structure closer to augmentation than substitution. Full automation would require additional oversight and validation, raising total cost relative to the human wage saved. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While data-driven flagging tools are cheap to run, they still require human verification of physical stock condition, keeping the effective all-in cost closer to comparable to human labor for this specific judgment task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably recommends disposal decisions across defective, excess, and obsolete categories in production environments. AI can assist with inventory metrics, but the actual recommendation—which involves legal, financial, and reputational risk—remains a human call in real organizations. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Inventory management software can flag slow-moving or aged stock using data analytics, but few deployed products autonomously recommend disposal decisions incorporating physical condition assessment at scale in warehouse settings. |
Examine and inspect stock items for wear or defects, reporting any damage to supervisors.
33CI 30–35 · exposure 25 · augmentation 50 · importance 3.3/5 · click for rater detail
Examine and inspect stock items for wear or defects, reporting any damage to supervisors.
33| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Retail and warehousing sectors show early-stage adoption of automation but remain heavily manual; most stockers still inspect visually. Pilots exist but production deployment at scale is uncommon outside large, digitally mature operations. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Warehousing and logistics sectors show moderate automation adoption (e.g., conveyor sorting, robotics) but visual inspection tasks specifically lag behind picking/packing automation in deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-assisted defect detection can flag items for human review, speeding up inspection workflows and reducing eye fatigue. The stocker remains the decision-maker, making this a useful augmentation on a routine but essential task. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-powered vision tools can flag potential defects or damage to assist workers, improving detection speed and consistency, though final judgment often remains human. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Computer vision can detect obvious defects and damage in controlled settings, but real-world stock inspection involves nuanced judgment about wear severity, functional impact, and borderline damage that requires human expertise. Current systems handle only narrow, well-defined damage patterns reliably. |
| Task automatability | claude-sonnet-5 | 2/5 | Visual inspection of physical stock items requires robotic manipulation and computer vision integrated with physical handling, which is not off-the-shelf deployable at scale for general warehouse stock today.confidence. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Liability concerns exist if automated inspection misses damage leading to customer harm, and many retailers prefer human accountability for quality decisions. However, no legal mandate requires human inspection, creating moderate friction rather than hard barriers. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but physical handling and judgment about damage severity create practical friction requiring human dexterity and contextual judgment. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Camera hardware, model inference, and integration costs are moderate, but the need for oversight, retraining on new products, and fallback human inspection keeps total cost per inspection close to or above a minimum-wage stocker's inspection rate. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Deploying vision systems and robotics for physical inspection requires significant capital investment in cameras, sensors, and integration, making it costlier than human labor for most warehouse operations except at very high volume. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Vision-based quality inspection systems exist in manufacturing contexts, but deployed retail and warehouse solutions struggle with varied lighting, angles, and the diversity of stock items. Production systems remain limited and typically require human verification. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some automated quality-inspection systems exist for specific defect types (e.g., barcode scanning, conveyor-based vision systems) but general stock inspection across varied item types in warehouses is largely still manual. |
Receive, unload, open, unpack, or issue sales floor merchandise.
29CI 23–35 · exposure 25 · augmentation 25 · importance 4.0/5 · click for rater detail
Receive, unload, open, unpack, or issue sales floor merchandise.
29| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Retail and small-to-medium warehousing remain highly fragmented, low-digitization sectors with labor-cost-driven operations. Adoption of automation here lags behind finance, tech, and large logistics, with most stockers still hired rather than replaced by machines. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Retail and warehousing are adopting automation, but mostly in large-scale distribution centers; sales-floor stocking specifically remains largely manual with slow, uneven adoption. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI offers limited augmentation for stockers today; AR picking guides and inventory apps exist but do not materially transform the physical unloading and unpacking work. The task remains largely manual and labor-intensive with minimal AI-assisted productivity gains. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI-driven inventory management and handheld scanning tools assist with tracking and prioritization, but they offer limited direct support for the physical unpacking and stocking motions themselves. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While unloading and unpacking involve repetitive physical motions, current AI lacks the dexterity, real-time perception, and adaptive problem-solving needed to handle varied packaging, fragile items, and irregular merchandise at scale. Robotics exist for narrow cases but cannot yet match the flexibility and speed of human stockers across diverse retail environments. |
| Task automatability | claude-sonnet-5 | 2/5 | This is a physical manipulation task involving lifting, unpacking, and moving merchandise, which current AI systems (as opposed to specialized robotics) cannot perform end-to-end; robotic solutions exist only in narrow, controlled pilot deployments.aget in some cases the reasoning software could help route or track items but that's a minor fraction of the task. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Retail environments vary widely (narrow aisles, mixed merchandise, safety hazards, irregular layouts), creating significant practical friction. Customer preference for human staff, liability concerns around robot-caused damage or injury, and the physical demands of compliance with health/safety rules create substantial adoption friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates a human for this task, but physical environment variability, safety concerns around heavy equipment, and organizational inertia create moderate friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | High-end warehouse robots cost hundreds of thousands of dollars upfront with ongoing maintenance, software, and integration costs; current systems cannot achieve 50% time savings at equal quality per dollar compared to minimum-wage stockers in most retail contexts. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Robotic unloading/stocking systems require significant capital investment, integration, and maintenance, often exceeding the cost of low-wage human labor for this task, especially at smaller retail scale. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some specialized robotic systems exist in controlled warehouse environments (e.g., palletizing), but production deployments handling general sales-floor merchandise—with unpacking, inspection, and floor placement—remain limited and error-prone. Most retail still relies on human stockers; no mature off-the-shelf product reliably does this task end-to-end. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some warehouse robots exist for unloading/sorting in highly structured environments (e.g., Amazon fulfillment centers), but retail sales-floor stocking with variable packaging remains largely unautomated in deployed products. |
Dispose of damaged or defective items, or return them to vendors.
28CI 21–35 · exposure 20 · augmentation 38 · importance 3.7/5 · click for rater detail
Dispose of damaged or defective items, or return them to vendors.
28| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Stocker roles are concentrated in lower-wage, lower-digitization settings (warehouses, retail). While some large retailers experiment with automation, meaningful adoption of damage-assessment and return-logistics automation remains limited and spotty. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Warehousing and retail stocking are only moderately digitized in physical operations; while inventory software adoption is growing, robotic handling of damaged goods disposal is still rare in production. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | Vision-based damage flagging systems can assist workers by highlighting suspicious items for inspection, reducing the time spent on manual visual scanning and potentially improving consistency in damage detection decisions. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI-driven inventory management systems can flag damaged items or automate vendor return paperwork, offering some assistance, but the core physical task itself sees little augmentation. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While identifying and sorting items could be partially automated with computer vision, the decision to dispose versus return to vendors and the actual handling/packaging steps require judgment and physical manipulation that current AI systems cannot reliably perform end-to-end at 50% time savings. |
| Task automatability | claude-sonnet-5 | 2/5 | This task requires physical handling, inspection, and movement of goods, which current AI systems and robots cannot reliably perform end-to-end in typical warehouse conditions.; software could assist with tracking/documentation but the physical disposal/return process remains manual. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Vendor return agreements typically require documented evidence and human sign-off; disposal of certain items (electronics, hazardous materials) may have regulatory requirements. However, these are procedural rather than absolute legal barriers to automation of the inspection itself. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but there is some organizational friction around vendor return processes, quality judgment calls, and liability for correctly identifying defective versus salvageable goods. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Computer vision inspection systems exist but are expensive to deploy and maintain, and the cost of incorrect damage assessment (sending good items back, disposing returnable items) makes the all-in cost often exceed simple human inspection and handling. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Physical robotic manipulation and mobility for identifying and moving damaged items is currently far more expensive than paying a warehouse worker to do this routine task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Current computer vision can detect some damage, but classification of items as truly defective versus acceptable, vendor return logistics, and documentation require human oversight. No deployed product reliably handles the full workflow without human intervention. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously identifies, sorts, and physically disposes of or returns damaged items at scale; this remains a manual warehouse floor task with only adjacent software support for logistics tracking. |
Design and set up advertising signs and displays of merchandise on shelves, counters, or tables to attract customers and promote sales.
24CI 19–30 · exposure 16 · augmentation 38 · importance 3.6/5 · click for rater detail
Design and set up advertising signs and displays of merchandise on shelves, counters, or tables to attract customers and promote sales.
24| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Retail stocker roles remain largely low-tech and physically localized; adoption of automation in this specific merchandising task is minimal, with most organizations still relying on human stockers for hands-on display work and aesthetic judgment. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Retail stocking is a low-digitization, physical-labor-heavy sector where AI adoption for in-store tasks remains slow and largely limited to back-office or design-support functions. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist by generating display design suggestions or layout templates, but current tools offer limited value since most stockers rely on store guidelines, planograms, and supervisor direction rather than independent design work that AI could enhance. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Generative AI tools can assist in creating sign designs, layouts, and promotional copy, improving efficiency of the creative/planning portion of the task even though physical execution remains manual. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI could theoretically assist with design layouts or suggest display arrangements, the physical setup on shelves, counters, and tables requires embodied robotics that is not yet deployed at scale in retail environments. Only design conceptualization could be partially automated; actual execution remains manual. |
| Task automatability | claude-sonnet-5 | 2/5 | Physical placement of signage and merchandise displays requires manual manipulation in real-world retail space, which current AI cannot perform; AI can help design the sign graphics but not execute the physical setup. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | There are moderate adoption barriers: retailers prefer human creative judgment for merchandising to reflect brand identity and local customer preferences, and display setup requires physical presence and real-time adjustment that customers often appreciate as a human touch. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or regulatory barrier exists, but physical execution and retailer-specific merchandising standards create practical friction against full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Current AI tools for design assistance are negligible in cost but cannot replace the human physical labor and spatial judgment required; the total cost including human oversight and setup labor far exceeds pure AI inference costs, making AI not cost-competitive for this task. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI can cheaply generate sign designs/templates, but the physical setup still requires paid human labor, so overall cost savings versus a human worker are limited. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed retail automation system currently performs end-to-end advertising sign design and merchandise display setup in production. Visual merchandising remains a skilled human task with minimal commercial AI product penetration in this specific application. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously designs and physically installs retail displays; this remains a human physical task with only design-support tools available. |
Clean and maintain supplies, tools, equipment, and storage areas to ensure compliance with safety regulations.
20CI 5–35 · exposure 13 · augmentation 38 · importance 3.7/5 · click for rater detail
Clean and maintain supplies, tools, equipment, and storage areas to ensure compliance with safety regulations.
20| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Adoption of automation in stocking and warehouse maintenance remains uneven; pilots and partial automation (conveyors, simple robots) are more common than comprehensive AI-driven compliance systems. Most small and mid-sized distribution centers still rely heavily on human stockers for these tasks. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Warehousing and logistics have low digitization for physical upkeep tasks, and robotic cleaning/maintenance solutions are not being deployed at meaningful scale in this occupation. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-powered inspection tools and alerts (computer vision for hazard detection, compliance checklists) can assist stockers in identifying issues and tracking maintenance, improving their productivity and safety awareness without removing them from the loop. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could assist with scheduling maintenance, tracking supply inventories, or generating safety checklists, but it offers minimal help with the physical cleaning and upkeep itself. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI-driven robots can move items and basic cleaning is automatable, the task requires judgment about safety compliance, hazard identification, and knowledge of varying regulations that are difficult to fully automate end-to-end without human oversight. Current systems cannot reliably inspect and maintain diverse storage conditions to meet safety standards without significant human intervention. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical cleaning and maintenance task requiring manual labor and mobility in a warehouse environment; no off-the-shelf AI system can perform the physical work itself. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Safety compliance and regulatory adherence (OSHA, storage regulations) often require documented human accountability and sign-off. Many facilities require a licensed or trained human to certify that storage areas meet safety standards, creating a legal barrier to full automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement exists, but physical dexterity, safety compliance checks, and workplace liability create moderate friction against pure automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Autonomous cleaning and maintenance robots remain capital-intensive with high integration costs, while stockers have relatively low loaded wages. For small-to-medium warehouses, human labor is still cheaper than full robotic solutions; even at scale, the cost gap is not yet an order of magnitude. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute for physical cleaning/maintenance, so any hypothetical robotic solution would be far more costly than a warehouse worker performing this task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Robotic systems exist for some cleaning and material movement, but deployed products rarely handle the full compliance inspection and maintenance task reliably. Production deployments are narrow (e.g., warehouse floor sweeping) and don't cover the judgment-intensive compliance aspects at scale. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product cleans equipment or storage areas; this remains purely a physical, human-performed task with only nascent robotics research in adjacent areas. |
Clean display cases, shelves, and aisles.
19CI 15–24 · exposure 8 · augmentation 0 · importance 3.9/5 · click for rater detail
Clean display cases, shelves, and aisles.
19| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Retail cleaning automation is minimal outside large-scale warehouses. Small and mid-market retailers—which dominate the sector—continue manual cleaning due to cost and workflow integration barriers; adoption remains nascent. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Retail stocking and cleaning tasks are physical and low-digitization; robotic adoption for cleaning tasks in retail settings is minimal and slow-moving. |
| Augmentation potential | claude-haiku-4-5-20251001 | 1/5 | Current AI and robotics offer minimal augmentation to human cleaners performing this task. There are no AI tools that measurably improve a stocker's cleaning productivity while keeping them in the loop. |
| Augmentation potential | claude-sonnet-5 | 1/5 | AI offers little to no direct assistance for the physical act of cleaning shelves or display cases in current practice. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Cleaning tasks require navigation of cluttered retail environments, handling fragile items, and adapting to varied layouts. While robotic cleaners exist in controlled settings, this task in real retail—with obstacles, merchandise placement variation, and breakable goods—remains largely beyond reliable end-to-end automation today. |
| Task automatability | claude-sonnet-5 | 1/5 | Physical cleaning of shelves and display cases requires manipulation, mobility, and dexterity that current AI systems (software-based) cannot perform; this requires robotics, not AI in the deployed sense. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Retail businesses prefer human cleaners to maintain product placement oversight and customer service. Some stores require cleaning standards enforcement, though no legal mandate prohibits robotic cleaning; adoption friction remains moderate. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal barriers, but practical barriers exist: physical environments, breakable merchandise, and irregular shelf layouts create friction for automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Robotic cleaning systems capable of navigating retail environments are capital-intensive and require significant integration. The all-in cost per task execution substantially exceeds the wage of a stocker performing routine cleaning. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Specialized cleaning robots or automation for shelving/display cases would be costly to deploy and maintain compared to low-wage human labor already performing this task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably cleans retail display cases, shelves, and aisles in unstructured, high-variety environments. Niche robotics exist for narrow warehouse scenarios but do not meet production-scale expectations in customer-facing retail spaces. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No mature robotic products are deployed at scale performing general retail cleaning of shelves and aisles reliably; cleaning robots exist mainly for floors, not shelving or display cases. |
Operate equipment such as forklifts.
19CI 7–30 · exposure 13 · augmentation 25 · importance 4.2/5 · click for rater detail
Operate equipment such as forklifts.
19| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Warehouse automation is growing but autonomous forklift adoption remains limited and largely confined to large, capital-rich firms (e.g., major fulfillment centers). Most small and mid-sized stockers still rely on human operators; the sector shows pilot activity but not deep, fast production deployment. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Warehousing and logistics are adopting automation more than average physical sectors, but autonomous forklift adoption remains limited to a small number of large-scale operators, mostly in pilot or narrow deployment phases. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI can assist forklift operators through load recognition, path planning suggestions, or collision alerts, but these are narrow augmentations. Humans remain firmly in control of the vehicle itself, and current systems offer modest productivity gains compared to the need for full task ownership. |
| Augmentation potential | claude-sonnet-5 | 2/5 | Some assistive technologies exist (e.g., collision-avoidance sensors, guided navigation aids) but these offer only marginal productivity gains rather than transforming how the task itself is performed. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Forklifts require real-time navigation, obstacle avoidance, and precise positioning in dynamic warehouse environments. While autonomous forklifts exist in research and limited pilots, they require extensive infrastructure changes and fail frequently in unstructured settings; no current off-the-shelf system achieves 50% time savings over a human operator today. |
| Task automatability | claude-sonnet-5 | 1/5 | Physical operation of forklifts requires real-world manipulation, spatial navigation, and dexterity that current AI systems cannot perform end-to-end without specialized robotic hardware, which is not the 'AI system' being evaluated here. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Liability and safety regulations are significant barriers: forklift operation is heavily regulated by OSHA, and autonomous equipment in shared warehouse spaces faces legal and insurance scrutiny. Many facilities require human oversight or formal certification; this regulatory coverage of automation itself creates meaningful friction. |
| Adoption barriers | claude-sonnet-5 | 3/5 | Workplace safety regulations (e.g., OSHA forklift certification requirements) and liability concerns around automated heavy machinery create moderate friction, though not a strict licensing requirement for a human to operate it. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Autonomous forklift systems (hardware + software + integration + maintenance) cost $150k–$300k+ per unit, while a warehouse worker earns $25k–$35k annually. The upfront capital cost and ongoing maintenance make the total cost of ownership comparable to or higher than human labor for most operations. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Autonomous forklift systems require expensive specialized hardware, facility retrofitting, and maintenance, making them costlier than a human operator in most warehouse contexts today. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Autonomous forklift products exist but are predominantly confined to controlled, mapped environments like warehouses with fiducial markers or GPS-denied settings. Most deployed systems require significant infrastructure investment and still have notable failure rates in real-world variability, falling short of reliable production-scale deployment. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | While autonomous forklifts exist in narrow, controlled pilot deployments (e.g., some large warehouses), they are not generally available or reliable products that replace human operators broadly across this occupation. |
Transport packages to customers' vehicles.
15CI 15–15 · exposure 0 · augmentation 25 · importance 3.1/5 · click for rater detail
Transport packages to customers' vehicles.
15| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Adoption remains negligible; retail and logistics sectors continue relying on human labor for last-mile and customer-facing delivery due to cost and technical barriers, with no production-scale autonomous systems in typical stocker/order-filler workflows. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Retail and warehouse fulfillment physical tasks show very low AI/robotics adoption in practice; this remains a low-digitization physical labor sector. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI offers minimal assistance to the human performing this task; potential augmentation is limited to route optimization or package tracking, which is peripheral to the core physical transport requirement. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with route optimization or notifying customers of pickup timing, but offers little direct assistance to the physical act of transporting packages to a vehicle. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires physical manipulation in unstructured outdoor environments, navigation to individual vehicles, and adaptive handling of varied package sizes—capabilities well beyond current AI/robotics deployed at scale in retail/logistics contexts. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical retrieval and transport of packages to a vehicle, a task current AI systems cannot perform without embodied robotics, which are not generally available for this function. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | While no explicit licensing bars automation, safety liability for autonomous systems operating around customers, premises liability, and organizational friction in customer-facing settings create meaningful but not insurmountable adoption friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but physical dexterity, liability for damage/loss, and customer interaction create moderate friction against automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Mobile manipulation robots capable of this task cost hundreds of thousands of dollars per unit with significant operational overhead, vastly exceeding the hourly wage of stockers. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI/robotic substitute deployed at scale for this task, so the human remains the only cost-effective option today. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product reliably performs end-to-end autonomous package transport to customer vehicles in real retail settings; robotics solutions exist only in controlled warehouse environments. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product reliably performs curbside package delivery to customer vehicles; this remains a human physical labor task with occasional robotic pilots but no production-scale deployment. |
Related occupations — Transportation & Material Moving
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.