A U.S. building-products manufacturer we work with is running a pilot where one AI agent follows each order from the mill certificates through receiving, production, the yard, and dispatch. Most of the time it keeps the record straight. When something needs a decision, such as a mismatch or a delay, it calls in the right person.
There is no robot in that pilot. Every conversation about robots starts with the machine, and we think the more useful place to start is the record the machine will need.
1. The robots are real, and most are still in pilots
A humanoid robot is a robot built roughly in the shape of a person, with arms, hands, and legs or a wheeled base, so it can work in spaces designed for people. As of September 2026, a handful are doing real work. Most of it is labeled a pilot by the companies themselves.
At BMW's plant in Spartanburg, South Carolina, a Figure 02 robot spent ten months of 2025 loading sheet-metal parts for welding in the body shop. BMW says it moved more than 90,000 components over about 1,250 operating hours and supported production of more than 30,000 X3s. BMW calls that a completed pilot. In June 2026, it started a project with the newer Figure 03, sorting parts into sequencing trolleys for the line. In Leipzig, BMW tested a robot called AEON, made by Hexagon Robotics, from December 2025, with its pilot phase planned for summer 2026 in battery assembly and component production.
In warehouses, one deployment is formally commercial. In 2024, GXO signed what Agility Robotics called the industry's first formal commercial deployment of humanoid robots: Digit robots moving totes from collaborative robots onto conveyors at a SPANX facility, rented as a service. Other companies are at an earlier or less clear stage. Apptronik announced partnerships with Mercedes-Benz, GXO, and Jabil in February 2026 without saying which are pilots. Tesla's July 2026 update says it is installing Optimus lines at Fremont and will use the first builds for training data.
So the sober reading is this: real, dated, useful work, in a few large plants, on narrow jobs.
2. The working deployments have narrow, tightly defined jobs
Look at what those jobs have in common. Each one is narrow, repeated, and defined in detail.
At Spartanburg, Figure says the robot had 37 seconds to load parts inside an 84-second cycle, on 10-hour shifts, Monday to Friday. At GXO, Agility says the robots' work is orchestrated through its own software, Agility Arc. BMW describes its humanoids as a complement to existing automation, for work that is monotonous or hard on the body.
None of these robots walked into an undefined job and worked it out. Someone decided what the task was, what counted as done, and how it fit the line around it. That is the same clarity an AI employee needs before it drafts a single cert.
3. A robot needs a record it can read
Our view: the manufacturers that benefit first from humanoid robots will be readable operations. We use that term for an operation where every job, part, and handoff is recorded in a system a machine can read, with a named person who owns each step.
A robot sorting parts into a sequencing trolley has to know which parts belong to which order and in what sequence. That knowledge comes from a system, the same way an AI drafting a CoC has to pull the heat number from a record rather than from someone's memory. If the sequence lives on a whiteboard, or in one supervisor's head, the robot has nothing to read.
This is why we tell clients the groundwork for paperwork and the groundwork for robots are the same groundwork. A digital traveler, a receiving form with the heat number in its own field, and one record per job are what an AI employee works from today. They are also what a robot, or the software that directs it, would work from later.
4. A named owner decides what happens when a machine stops
Every machine stops sometimes. A robot drops a part, a sensor misreads, an AI flags a heat number that does not match the receiving log. What matters is who gets the call and what they can see.
Our view: the rule we use for AI employees carries over. Each one has a named person who owns its work, a limited set of things it is allowed to touch, and a log of every step it took. When our agent flags a mismatch, the quality lead sees exactly what it read and where it stopped. A company that already runs its AI this way has the habits a robot program will need: clear owners, logged actions, and a person who decides.
What to do this week
- Pick one repetitive physical task on your shop floor, such as kitting or moving parts to the line, and write down where the instructions for it come from today: a system, a paper list, or a person's memory.
- For one job, check whether you can trace a part from its mill cert to the shipment in your records without asking anyone. Note every step where you had to ask.
- Name one owner for each handoff in that job (receiving, cut, weld, inspect, ship) and write the names on the traveler template.
- Hold any robot vendor meeting until steps 1 to 3 have clear answers. The records you build for this are the same ones an AI employee needs, so they pay off whether a robot ever arrives or not.
For the full argument, read Manufacturing's ChatGPT Moment Hasn't Happened Yet: Why AI Fits Your Paperwork and How to Prepare. To start the records a robot would need, read From Paper Traveler to Traceability: Why Digital Forms Come Before AI.
Questions people ask
Are humanoid robots working in factories yet?
A few are, mostly in pilots. BMW reports a Figure 02 robot supported production of more than 30,000 X3s during a 10-month pilot in 2025, and Agility Robotics reports more than 65,000 hours of operation for its Digit robots. Most deployments are narrow, repeated tasks at large companies.
What should a mid-size manufacturer do about humanoid robots now?
Get the work digital, traceable, and owned: one record per job, heat numbers in their own fields, and a named owner for each handoff. Those records pay off with AI today and are what a robot would need later.
Terms in this piece
- Humanoid robot
- a robot built roughly in the shape of a person so it can work in spaces designed for people.
- Pilot
- a limited trial of a new system on real work, with a set end date, before a company decides whether to use it widely.
- Robots as a service
- renting robots for a monthly or hourly fee instead of buying them.
- AI employee
- an AI set up like a new hire, with its own login, one defined job, and a named person who reviews its work.
- Readable operation
- an operation where every job, part, and handoff is recorded in a system a machine can read, with a named owner for each step.
- Traceability
- the ability to follow any part back through every step of the job to the material and documents it came from.
