Cobots for small manufacturers: a realistic entry point into automation
Collaborative robots have changed the economics of factory automation for small and mid-size production companies. What is realistic to expect - and what the common mistakes are.
For most of the history of industrial robotics, automation was a large-company game. The capital cost of a traditional industrial robot - the arm, the safety cage, the integration, the programming - was justified only at high volumes and long production runs. A company making a few thousand parts a week had no real entry point.
Collaborative robots, or cobots, have changed that economics in the last five years. The pandemic has accelerated the conversation: the robots-kept-running story from spring 2020 is now part of board-level discussions about production resilience that were not happening two years ago.
What a cobot actually is
A cobot is a robotic arm designed to work alongside people rather than in a separated, caged area. The safety approach is different from traditional industrial robots: instead of preventing contact entirely, cobots are designed to detect contact and stop or slow down. This allows them to operate in shared workspaces without extensive physical guarding.
The major manufacturers - Universal Robots, FANUC, ABB, KUKA, and several others - all have cobot lines. The leading platforms (Universal Robots' UR series in particular) have large ecosystems of end-effectors, vision systems, and integration software built around them.
Where they fit for small manufacturers
The tasks that cobots handle well:
- Pick and place - moving parts between positions, loading and unloading machines, palletising. Repetitive, well-defined, high-volume-per-shift.
- Screw driving and fastening - consistent torque, consistent cycle time, no fatigue.
- Gluing and dispensing - path following with consistent speed and pressure.
- Simple assembly - inserting components, pressing fits, where the tolerance allows for the positioning accuracy of the arm.
- Machine tending - loading a CNC machine, waiting, unloading. A task that ties up a skilled worker unnecessarily.
The tasks they do not handle well: anything requiring fine dexterity and real-time adaptation, highly varied parts without vision, or process steps where the physical setup changes frequently.
The cost picture in 2020
A UR5 or similar mid-range cobot arm costs roughly 25,000-35,000 EUR. Add an end-effector (gripper, screwdriver, etc.) for 3,000-15,000 EUR, and integration and programming for 15,000-40,000 EUR depending on complexity. A complete, working installation at a single workstation is typically in the 50,000-90,000 EUR range.
Payback periods for well-chosen applications are typically 12 to 24 months. For a task that currently requires one operator per shift across three shifts, the arithmetic usually works.
What goes wrong in practice
Three failure patterns are common.
First: the task selected for the pilot is not actually repetitive enough. If the product changes weekly and setup takes a day, the cobot sits more than it runs.
Second: integration is underestimated. The arm is the small part of the budget. The gripper design, the fixturing, the vision system if needed, the safety assessment, the programming, the operator training - this is where the real cost lives.
Third: no internal owner. A cobot that breaks or needs reprogramming needs someone on-site who knows how to handle it. Outsourcing all of that to the integrator creates a dependency that is expensive and slow.
The right way to start
Before calling an integrator, identify one repetitive task that currently ties up a person for at least four hours per shift. Document the task in detail: cycle time, part variability, tolerances, what changes between batches. That documentation is the basis for a realistic vendor conversation - and it protects you from being sold a solution to a problem you do not actually have.