Why the paperwork moves faster than the dock
This job splits into two halves. One half is records: matching packing slips to purchase orders, keying receipts into a warehouse management system, preparing bills of lading, and computing freight charges. Software has been eating into that half for two decades, and recent AI tools read messy documents far better than older systems did.
The other half happens with your hands and eyes. Someone has to open a crushed carton and judge whether the goods are still sellable. Someone has to walk the rack when the count on screen does not match the count on the shelf, pull the pallet apart, and find where the error started. Someone has to sign for a short delivery, call the carrier, and hold the line on a claim.
So the honest reading is task erosion rather than a job vanishing. The Bureau of Labor Statistics counts 816,870 of these jobs in the US, with median pay of $45,260 and a projected 7.6% decline in employment between 2025 and 2035 (BLS, 2025). That looks like fewer clerks per dock and fewer openings for beginners, which is the pattern behind our shrinking roles list.
What software runs, what it assists, and what stays on the floor
Start with the work AI can run end to end (share of task time: 0%). That is mostly document and data work: producing shipping labels and bills of lading from order data, and reconciling received quantities against purchase orders in the system. These tasks have clean inputs, clear rules, and an audit trail, which is exactly what current tools handle well.
Next comes assisted work (share of task time: 61%). Here the tool does the first pass and a person decides. Scanners and RFID readers keep a running count while the clerk confirms the odd reads. Tracking software flags a late or misrouted shipment, then the clerk works out whether to chase the carrier, split the order, or warn the customer.
The rest of the work stays with people (share of task time: 39%). Inspecting inbound freight for damage and deciding what to reject is one. Physically investigating a count discrepancy, then correcting the record and the stock, is another. Our Can AI do it? score for this occupation is 30 out of 100; how that share of task time is measured is set out in the coverage method.
How strong the evidence is
There is no published head-to-head test of an AI system against a working shipping, receiving, and inventory clerk. Our Is it better than a person? grade is D, which means the comparison has not been measured, so we give no parity number at all. Vendor case studies about scan accuracy or forecast error are not the same thing: they test a component, not the whole role.
What would settle it is narrow and testable. Run a receiving dock for a quarter with the same inbound volume, and compare a clerk-led process against an automated one on damage caught at intake, inventory accuracy after cycle counts, chargebacks avoided, and time to resolve a disputed delivery. Until something like that exists in public, treat any confident claim about parity here with caution. The grading scale is explained on the quality parity page.
When the balance could shift
Most likely between 2036 and 2056 (8 in 10 of our scenarios). What that window represents, and how we build it, is set out on the replacement year page.
Two things could pull it earlier. Mobile robots are the robotics tier that matters here, and they are already common in large distribution centers for moving and counting stock; wider fleets would cut the walking-and-checking part of the role. The second is cost: running a model over shipping documents is far cheaper per task than paying for the same hour of clerical labor, so the record half faces steady pressure.
Two things hold it back. Physical variety is the big one, because damaged, mislabeled, and mixed pallets do not arrive in a format a robot expects, and older buildings were not laid out for machines. The second is accountability: someone has to sign for goods, certify a count for an audit, and stand behind a claim against a carrier. That sign-off has stayed with people even in heavily automated sites.
What to do: get yourself attached to the exception work, the counts, claims, and damaged freight, because that is where the human share of this job sits.
Staying needed on the dock
Three tasks are worth owning. First, inbound inspection and damage calls, including the photos and notes that make a claim stick. Second, discrepancy investigation: finding why the system and the shelf disagree and fixing both. Third, carrier and supplier coordination when a shipment goes wrong and a decision has to be made quickly.
Two skills travel well. One is data hygiene in the warehouse management system, including item setup, location accuracy, and knowing why a scan failed. The other is automation oversight: running cycle counts around robots and scanners, spotting where the tool is quietly wrong, and training new staff on the exceptions.
If you are weighing a move, these roles sit close to yours: Production, Planning, and Expediting Clerks, Cargo and Freight Agents, and Weighers, Measurers, Checkers, and Samplers. You can put any two side by side with the job comparison tool, see neighboring roles on the material recording and distributing family page, or read the wider picture for warehousing jobs. People often ask whether AI will replace shipping, receiving, and inventory clerks outright; the task split above is a better guide than any single answer, and the scoring method shows how we got there.