Humanoid Pilots Are Coming to the Factory Floor: How to Scope One

AGCO and Agility Robotics have signed a commercial agreement to pilot Agility's Digit humanoid robot at AGCO's Hesston, Kansas production facility, with the U.S. pilot expected to begin in 2027, according to AGCO's October 20 announcement and recent coverage in High Plains Journal. The initial tasks are narrow and specific: sorting and stacking parts and placing totes into storage. That narrow scope is the most useful part of the story for any operations leader weighing a humanoid program.

Start with the task, not the robot

AGCO describes the goal as moving repetitive, physically demanding work off people to improve ergonomic safety and efficiency. Notice what is not in the announcement: no robot count, no line takeover, no promise of a lights-out cell. The pilot is built around material handling tasks that are well defined, repeat many times per shift, and are easy to measure.

That is the right shape for a first humanoid deployment. Before a robot arrives, a team should be able to write down the exact task boundary: where the part or tote comes from, where it goes, what counts as done, and what the robot should do when something is out of place. If that description runs longer than a page, the task is probably too broad for a pilot.

Define success before the pilot starts

A pilot without agreed metrics tends to end in opinions. Useful measures for a tote or parts handling task include:

• Cycle time and throughput compared against the current manual baseline
• Intervention rate: how often a person has to step in, and why
• Uptime across full shifts, including charging and recovery time
• Error rate: misplaced totes, dropped parts, or mis-sorts
• Safety and ergonomics outcomes for the people who previously did the work

Capturing the manual baseline before the robot shows up is easy to skip and hard to reconstruct later.

The facility has to be ready too

Humanoids are pitched as fitting into spaces built for people, but a production floor still has to support them. Agility manages Digit fleets through its Agility Arc cloud platform, according to High Plains Journal, which means network coverage, IT approvals, and data access need to be settled early. Charging locations, travel paths, tote and rack dimensions, floor conditions, and how the robot interacts with existing safety zones all affect whether a pilot runs smoothly or stalls in its first week.

Thomas Enghof, AGCO's director of manufacturing technology, framed the robots as teammates paired with safety governance. In practice, that governance means a documented risk assessment, clear rules for shared spaces, and a training plan for the operators and maintenance staff who will work next to the robot.

Plan for what happens after the pilot

A pilot that works in one area raises the next questions quickly: who supports the robots on second shift, how failures get diagnosed and logged, how spare parts and software updates are managed, and how lessons transfer to a second building. Teams that answer those questions during the pilot, not after it, are the ones that move to production on schedule.

What it means for a project

For manufacturers watching the AGCO announcement and considering their own humanoid or mobile manipulation pilot, the work splits into a few practical tracks. Emtech's Testing engineers build the test plans and baseline data that turn a pilot into a clear go or no-go decision. Lab support covers site readiness, including IT and network setup, plus in-house 3D printing for fixtures or tote adapters. Reliability engineers run root-cause analysis across controls and embedded systems when interventions start to pile up. When the decision is to scale, Emtech's Deployment and Engineering & Field Support teams handle installation, commissioning, and multi-site rollout, with senior engineers measured against the customer's KPIs rather than billed hours.