I help manufacturers, technology suppliers and market researchers connect robot capability with factory tasks, component requirements and pilot economics.
My perspective combines European market development at UBTECH with experience in robotics, system integration and business building.
I begin with the workflow and compare options against the same requirements. These categories overlap: a humanoid torso can use legs or wheels, and an autonomous mobile robot (AMR) can carry an arm.
Questions I use to structure an application review
Robot type
Application to investigate
Key comparison question
Humanoid
Handling work across stations designed for people.
Does its reach, mobility and manipulation justify the added complexity?
Wheeled mobile manipulator
Moving between stations and handling objects.
Can wheels cover the route while the arm meets the handling requirements?
AMR
Material transport between defined locations.
Can transport be solved separately from loading and unloading?
Fixed arm or cobot
Repeated manipulation at a bounded workstation.
Would fixtures and a stable layout simplify the application?
Legged inspection robot
Inspection along routes with access constraints.
Do terrain and stairs create a real need for legs?
Technology & market questions
Turn broad interest into specific evidence.
These are the questions I can help structure for an expert discussion, supplier review or opportunity assessment.
Where could an industrial humanoid add value?
I look for a complete task where handling, movement or access creates an unmet need. Carton handling, machine tending and material supply are useful applications to investigate. The comparison includes the current process, a dedicated automation solution and a mobile manipulator.
What I would ask for: object samples, workflow footage, station layout, cycle requirements and the exceptions an operator resolves today.
Which components are shared across robot types?
The common building blocks include motors and drives, transmissions and bearings, sensors, computing, control electronics and power systems. Grippers and other end effectors connect the robot to its task. Schaeffler’s component overview illustrates several of these categories.
Shared categories do not make parts interchangeable. I compare torque, precision, mass, duty cycle, thermal limits, interfaces and support requirements for the intended application.
Useful output: a component map separating reusable capabilities from requirements specific to each robot form.
How do you judge maturity beyond a demonstration?
I separate a successful demonstration from repeatable task performance, supervised pilot operation and readiness for the customer’s production conditions. A video alone cannot establish that progression.
Evidence to request: representative run duration, completed cycles, quality, intervention frequency, recovery time, availability and performance across object variations. Record what was autonomous and what an operator or remote assistant did.
Single-purpose or multi-purpose: what changes?
I would define one bounded first task, then test the value of extending the same platform. Additional tasks introduce new tooling, data, integration, validation and support requirements.
The commercial question: does higher utilisation or easier redeployment outweigh changeover effort and added complexity? A flexible platform still needs a justified first application.
Where could growth come from over the next 5–10 years?
I would compare adoption scenarios by task and customer segment rather than rely on a single robot shipment forecast. The variables to test include reliable manipulation, deployment effort, access to task data, service coverage, utilisation and total cost.
For a component supplier, I would also ask which capabilities serve several robot types and which depend on a particular architecture. This makes the growth discussion explicit about its assumptions and the evidence that could change them.
What must a pilot prove commercially?
Agree the baseline and acceptance criteria before the trial: throughput, quality, human intervention, recovery and the conditions under which results count. Include tooling, integration, training, maintenance, energy and support in the cost model.
Decision: continue, revise the task or stop, using agreed evidence. My ROI calculator helps make illustrative economic assumptions visible.
My relevant experience
A factory workflow. A seven-page assessment.
During my European market-development role at UBTECH, I observed a Swedish manufacturing workflow, prepared a seven-page workflow decomposition and data checklist, and coordinated technical discussions with suppliers.
The assessment covered carton dimensions and condition, handling positions, throughput, exceptions and the business objective. It connected customer needs with questions that engineering teams could evaluate.
Experience shown: customer discovery, application assessment and concept evaluation. This case does not establish a completed robot installation or measured productivity gain.
The complete handling sequence
Each step needs an answer.
01
Pick up
Object variation and grasp conditions.
02
Transfer
Travel path and destination position.
03
Open
Packaging, tools and exceptions.
04
Unload
Product handling and placement.
05
Dispose of the carton
Waste flow and readiness for the next cycle.
Anonymised summary of my assessment approach.
Public industry context
Examples to examine critically.
These external sources provide context for the discussion. They are separate from my project experience and do not imply an affiliation or endorsement. Reviewed 12 October 2026.
BMW Group · Factory application
From laboratory evaluation to a factory pilot.
BMW describes staged evaluation of humanoids and integration into manufacturing. Its Leipzig article includes factory photographs and videos of industrial applications.
My question: what changed in the workstation, interfaces and operating process to make the task viable?