Why robots stall in UK manufacturing, and how to get started
Why UK manufacturers stall after buying robots, what the deployment bottleneck actually is, and how to prove a first use case in simulation before committing to hardware.
OLO Robotics Team · · 5 min read
Most British manufacturers aren't held back by the price of robots. The harder part is deployment: programming, integration, and validation, all of which usually require specific ROS 2 specialists. The fastest way to reduce the risk of automation is to prove the idea in simulation, using the software skills you already have, before you commit to hardware.
Where the UK stands
- Three-quarters of UK manufacturing SMEs run without a single robot, according to Make UK.
- The UK has 112 robots per 10,000 manufacturing workers. That's roughly half the EU average and below the global average of 151.
- The gap has barely moved despite years of government support aimed at closing it.
The bottleneck isn't the hardware
Capable, affordable robots already exist. Quadrupeds, rovers, and robot arms are all available off the shelf. As our CEO Nick Thompson puts it, “Getting a robot operational is not primarily a hardware problem. The hard part is everything that comes after it arrives.”
That “everything after” usually includes:
- Programming: writing robot behaviour in ROS 2, the open-source framework most robots use.
- Integration: connecting the robot to your sensors, systems, and network.
- Validation: proving it works safely and reliably before it goes near production.
The tools for all of this were built for robotics specialists, and there aren't many of them. So projects queue up behind the few people who can deliver them, or never start at all. Our ROS 2 quadruped inspection guide walks through that build-versus-buy decision in more detail if the first robot is a quadruped.
The expertise is already on your site
The industry has spent a decade talking about a skills shortage. We see it differently. “When you sit with customers, they rarely talk about a ‘skills crisis’,” says our COO Eleanor Tang-Smith. “They talk about ideas they can't get to.”
Your production managers, engineers, and software developers already know which processes are worth automating. What they've lacked is tooling that lets them act on that knowledge without first becoming ROS 2 experts.
What changes when deployment gets easier
- Simulate before you buy. A cloud simulation of the robot, running realistic physics, lets you test feasibility before any capital spend.
- Write the code once. On OLO, the same SDK and code run in simulation and on the real robot. Work you do against a digital twin carries straight over when the hardware arrives, with no reconfiguration.
- Start before delivery. Teams can build navigation and task logic while the robot is still on order.
- Keep control on site. Code runs on an Appliance next to the robot, so scheduled jobs keep working even if cloud connectivity drops.
- Use robots that work from day one. OLO is ROS 2-native, so you can use any ROS 2 robot, or any of the natively supported robots we have tested thoroughly.
Traditional deployment vs software-first
| Traditional deployment | Software-first deployment | |
|---|---|---|
| Who can build | ROS 2 specialists or an external integrator | Your own engineers and developers, in Python or TypeScript |
| When work starts | After the hardware arrives and is set up | Immediately, in simulation |
| Cost of testing an idea | Hardware purchase plus integration time | A simulation run |
| Moving to the real robot | Re-work and re-configuration | Same code, same interfaces |
| Debugging | On site, often hard to reproduce | Recorded sessions replayed in the browser |
How to get started
- Pick one process. Choose something repetitive, hazardous, or hard to staff, such as inspection rounds, material moves, or monitoring. If the first candidate is a patrol or plant round, our robotic vs manual inspection comparison helps you decide which routes are worth moving first.
- Ask your own team. Find the people who know that process best, plus anyone who can write Python or TypeScript.
- Prove it in simulation. Build and test the behaviour against a digital twin before committing budget.
- Pilot on real hardware. Deploy the same code to one robot and compare the results against how you work today.