EXPOSED
The work is physical — unloading trucks, breaking down pallets, facing shelves, picking totes, scanning barcodes — so language models can't touch the core of it, but that's not where the threat comes from. The threat is warehouse automation capital: goods-to-person systems, autonomous mobile robots, and improving suction/pinch pickers already handle a growing share of order filling in high-volume distribution centers, and inventory counting is going to shelf-scanning robots and computer vision. Retail store stocking in cluttered aisles with mixed SKUs is far harder to automate economically and will persist much longer than DC picking, so this title splits sharply by setting.
Headcount grew steadily across the period.
Median pay $27,380 → $37,330 +9.1% in real terms
This line is counted by the Bureau of Labor Statistics — the one figure on this page that isn't a judgement of ours. Headcount moves on demand, offshoring, demographics and the business cycle, and automation is one term among several, often not the loudest.
So a falling line is not evidence that AI did it, and a rising one is not evidence that it won't. Both happen in this register: some occupations resist automation and shrink anyway, others are highly automatable and keep growing. The marked year is 2020.
BLS projection, 2024–2034
+8.5% 2,764,800 → 2,999,800 on the projections basis
Exposed, but growing
AI can already do a lot of these tasks, and the BLS still expects +8.5% more of these jobs by 2034. Demand for the output is growing faster than the work is being automated away — the mechanism BLS gives for software developers, and the combination people most often misread as an error.
Different clocks. The score is what current AI could do to this work today. The projection is how many of these jobs will exist in 2034. Everything between the two — how fast employers actually adopt, whether demand grows in the meantime — is why they can point opposite ways without either being wrong.
~472,300 openings a year on average, including replacing people who leave.
MarkerPackerPickerPricerPullerRouterSorterGearmanShipperStockerStubberToolmanReceiverStockmanTicketerCrib ClerkDie KeeperMeat ClerkTag MarkerTool ClerkWarehouserYard ClerkBay StockerCrib Tender
Holding it up: embodiment . Weakest point: liability shield .
Mixed — a routine tier and a judgment tier Breaking down a mixed pallet, judging which dented cans go back, facing a shelf so labels read straight, and hunting a mis-slotted SKU in a cluttered backroom are still hard for machines — but the barcode scan, the pick path, the cycle count, and the tote handoff are already being run by AMRs and goods-to-person walls in high-volume DCs, which is why this sits at 12 and not 17.
Hands-on in uncontrolled environments You are on your feet an eight-to-ten hour shift lifting 25-50 lb cases, climbing ladders, running a pallet jack or order picker through aisles shared with forklifts and shoppers, in trailers that hit 100°F and freezers that hit -10°F — the 15 rather than 19 reflects that the floor is at least a mapped, indoor, flat space that engineers can design robots around.
No licence, no signature requirement No licence, no certification, no registry — a powered industrial truck authorization under 29 CFR 1910.178 is employer-issued paperwork, not a credential you carry to the next job, and when a pick is wrong or a pallet falls the company eats it, so there is nothing here to slow a substitution decision.
Executes defined procedures on defined inputs Pick lists, planograms, FIFO date rotation, and RF-gun prompts define nearly every decision; the discretion left is calling a damaged case, flagging a count variance, or deciding whether to overstock or backroom the excess, which is real but small and reversible — hence 3.
The verdict above describes this occupation as a whole. Almost nobody does the typical version of a job — tick what's actually in your week and see how your own mix sits.
Your task mix speaks to task resistance (12/20 here) — how much of the day's work current AI already does. That is the dimension the boxes above are about.
It cannot move the other three. Liability shield (1/20) is whether the law requires a licensed human to sign. Trust premium (4/20) is whether buyers specifically pay for a person. Judgment and accountability (3/20) is whether the role exists to own consequential calls. Those are facts about the occupation's standing, not about which tasks are in your week — a paralegal who does only trial exhibits still holds no licence. Together they are 8 of this occupation's 35 points (23%).
Embodiment (15/20) is also a property of the work rather than the worker, but we don't tag individual tasks as physical or not, so the picker can't tell you anything about it. That's a limit of this tool, not a claim.
Did we get the list right? Tell us what's missing — the tasks are written from the outside, and you're reading this from the inside.
Orderlies EXPOSED
Postal Service Mail Carriers EXPOSED
Maids and Housekeeping Cleaners EXPOSED
The moves above are yours to make. This is the other half: what would have to change in the world for the occupation itself to score higher. None of it is in any one person's gift, but it is where the floor actually comes from. Scores here are not a one-way ratchet. Only two of the five dimensions — task resistance and embodiment — track what machines can do. The other three track law, what buyers will pay for, and who is answerable, and those move in both directions, often in response to the same pressure AI creates. If every lever below landed, this occupation would score around 47/100, still EXPOSED.
Task-mix shift inside the same title: if goods-to-person and AMR systems absorb routine tote picking and cycle counting in high-volume DCs, the residual job becomes exception work — jammed inductions, damaged/mislabeled cases, deformable and odd-shaped SKUs robots reject, decanting into totes, and recall/quarantine pulls. This tier genuinely exists (Amazon's 'amnesty' pickers, robot-floor exception associates) and current picking hardware fails on it. Applies mainly to automated DCs; retail store stocking has less of a two-tier split.
If the surviving employment concentrates in retail store stocking, grocery backroom/produce, and mixed-SKU cluttered-aisle facing rather than DC picking, the measured embodiment of the title rises purely by composition — narrow aisles, customer traffic, tall ladder work, deformable and crushable goods, and cold/frozen rotation remain economically unautomated. Watch grocery chains' capex splitting between micro-fulfillment (automated) and in-store shelf labor (not).
Certification-gated subsets of the work: DOT/PHMSA HM-232 hazmat handling certification, FDA FSMA 204 traceability recordkeeping for high-risk foods (compliance date already set), and state pharmacy-board rules on who may stock or restock Schedule II-V in retail pharmacies. If FSMA 204 lot-level traceability makes a named, trained receiver personally attest to inbound cold-chain and lot data, a documentary signature attaches to receiving work that currently has none. This shields specific stocker subsets, not the mass of the occupation.
If retailers formalize the receiver/lead role as the accountable party for recall execution, damaged-load rejection at the dock, and shrink/theft escalation — a named person signing off that a pallet was refused or a lot destroyed — the call under ambiguity becomes owned rather than diffuse. Watch job-architecture changes creating 'inventory control specialist' or 'traceability lead' as a distinct accountable title carved out of stocker headcount.
If ANSI/ITSDF B56.5 and OSHA enforcement of mixed human/AMR floors keeps requiring human presence for robot fault recovery, blocked-path clearing, and manual retrieval from automated storage, a floor of physical human tasks stays legally attached to the automation itself. Watch OSHA National Emphasis Program citations on warehouse robotics and any state rule requiring trained human recovery staffing ratios.
The limit. No plausible route to a meaningful trust premium: buyers of stocked shelves and filled totes do not know or pay for who touched the goods, and no consumer-facing signal exists to build one on. The realistic ceiling is roughly the mid-40s, and it comes almost entirely from surviving-subset composition (retail aisles, hazmat, cold chain) plus exception work — not from the 2.8M-worker title as a whole. Warehouse capital intensity keeps compressing headcount even where per-worker scores rise.
| New York-Newark-Jersey City, NY-NJ | 131,080 | $38,600 +3% |
| Dallas-Fort Worth-Arlington, TX | 111,100 | $37,590 +1% |
| Los Angeles-Long Beach-Anaheim, CA | 85,630 | $38,650 +4% |
| Chicago-Naperville-Elgin, IL-IN | 79,850 | $38,450 +3% |
| Riverside-San Bernardino-Ontario, CA | 77,160 | $43,200 +16% |
| Houston-Pasadena-The Woodlands, TX | 72,970 | $36,890 -1% |
| Atlanta-Sandy Springs-Roswell, GA | 50,970 | $36,160 -3% |
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 45,770 | $36,800 -1% |
| Cheyenne, WY | 1,720 | $46,900 +26% |
| Seattle-Tacoma-Bellevue, WA | 30,860 | $46,830 +25% |
| Kenosha, WI | 2,530 | $46,030 +23% |
Has AI actually changed your work? One tap, anonymous, and the running tally is public. Nothing else is asked of you.
Rather than check back: get the digest and we'll tell you what changed — or watch a single occupation from its own page.