Amazon plans to automate the work of its robot-battered human employees, aiming to have machines take over more and more of the work in its massive logistics machine, according to that purveyor of truthiness, The New York Times. The shift is one of the most aggressive applications of automation by a major retailer, and it could serve as an indicator of how large companies are capable of navigating the pandemic. Internal estimates show that more than a half-million roles will ultimately be taken over by machines and software.
The company already operates one of the world’s largest private logistics networks and is the second-largest employer in the United States. By closing the loop between advanced robotics, computer vision and software to handle everything from packing boxes in fulfillment centers to long-haul transport, Amazon is wagering that it can fulfill more orders with fewer new hires while also tightening delivery windows and reducing costs.

Inside Amazon’s 75 Percent Robotics Goal for Warehouses
Workers and internal documents described to the Times detail a process that works to automate much of the high-volume, repetitive work across inbound, picking, packing and sortation. In a showcase facility in Shreveport, La., that was designed from the ground up to be automated, some thousand robots have replaced much of the need for large teams of human workers. Managers would estimate the site needs about a quarter fewer people than an equal-sized building without this gear, one in which workers will be increasingly displaced as more workflows are handed off to machines.
Amazon plans to roll out the Shreveport model to dozens of fulfillment and sort centers in the near future. The company already has more than a million robotic devices operating throughout its network, the report said, from autonomous mobile drive units that reposition inventory to robotic arms for picking and packing to automated palletizers.
The executives, led by CEO Andy Jassy, are pushing for efficiency after years of expansion with board-level pressure to increase output per facility and per labor hour. Etc., etc. But the automation drive, as Wall Street analysts, including Bank of America’s Justin Post, quoted by the Times, would have it, is a turn away from growth at any cost to industrial discipline.
What changes on the warehouse floor with robotics
Operationally, the work shifts from people on foot covering miles of aisles to goods-to-person stations where robots ferry shelves to fixed picking pods. Vision systems and robotic grippers similar to those found in Amazon’s Sparrow and Sequoia programs can recognize and grasp a growing proportion of common objects. With increasing capabilities, the ratio of tasks shifts naturally from human skill to oversight, exceptions and maintenance.
Industry benchmarks indicate that these systems can reduce worker travel time sharply and boost throughput per square foot. The thornier problems are on the edges: odd-shaped items, fragile goods and complicated situations in which software has to weigh speed against safety. And this is where Amazon’s scale gives it an edge; every pick and scan contributes to the training data, so that the next decision takes even less time.

Jobs, equity, and messaging around automation
There are clear labor implications to the automation plan. The Times has reported that Amazon is gearing up for a communications onslaught highlighting “advanced technology” and even, yes, “cobots” working alongside humans, but also new technical roles to ensure machines keep running. Those are jobs that often require certification or specialized training, and could be fewer in number than the roles they replace.
Workforce equity is also on the rise. Amazon’s fulfillment centers employ a greater proportion of Black workers than is reflected in the national average, so reduced hiring or attrition may hit minority communities disproportionately hard. Labor advocates like the Strategic Organizing Center have previously singled out higher injury rates in warehouses than their peers, leaving them to wonder whether automation will take strain off workers or make work faster.
There is also a strategic aspect — robots don’t organize. Amazon’s longstanding conflicts with unions — fights with delivery drivers and warehouse organizers in New York highlight this tension — help explain why investors might see automation as not only a productivity weapon but also a way to mitigate the risk of labor.
What to watch next as automation expands at Amazon
Several clues will suggest whether the strategy works. See sustained gains in items picked per hour, shorter order-to-ship cycle times and steadier on-time delivery rates, even when faced with peak surges. Capital intensity is the other side of that; robot deployments are costly, and you get to pay it all back over time with high usage and low downtime.
Regulators and local officials will be looking for inflection points in warehouse cluster employment baselines; trends on injuries as automation scales; and whether the promised technical upskilling is getting to current workers. If Amazon can demonstrate in practice that mixed teams of humans and robots are both safer and more productive while continuing to provide stable work, the model could become standard across the industry fast. If not, a backlash over job losses and working conditions could spur tighter oversight.
For now, the world’s biggest online retailer seems to have something else in mind: a future where many of its packages are handled by robots and smaller centers — but with people making sure everything is properly handed off. The extent of that shift — to 75 percent robotic operations — would remake not just Amazon’s warehouses, but the prospects for an entire industry.