Swedish industry has never competed on low wages. It was never an option. A country of ten million people, with long transport distances and some of the highest labour costs in the world, cannot win on price per hour.
What we have competed on is productivity. How much value each hour worked creates. And every time Swedish industry has made a step change in that figure, it came down to the same thing: a new technology took over work that people had previously done by hand.
It is worth looking at how those steps actually happened, because the pattern is repeating right now — just in a place where most people are not looking.
Three waves, one pattern
Mechanisation. Water power and steam engines turned a forge into an ironworks. The sawmills, ironworks and engineering shops that grew through the second half of the 1800s were not built on more pairs of hands but on power coming from somewhere else. Sweden went from agrarian to industrial in roughly two generations.
Electrification. Around the turn of the century came the second wave, and this is where Sweden moved to the front. ASEA, Separator, SKF, Ericsson: companies built on Swedish inventions, in an era when electricity reshaped the shop floor itself. The electric motor meant each machine could be driven on its own, rather than by belt from a central shaft. The shop floor could finally be laid out around the flow instead of around the power transmission.
Robotisation. In 1974 ASEA introduced the IRB 6, the world’s first microcomputer-controlled electric industrial robot and the first to be sold commercially. It could lift six kilos and was not superior to a human in any way but one: it made the same movement identically, around the clock, without tiring. Today Sweden is among the most robot-dense countries in the world, and a large part of our manufacturing base would not exist here without that development. It would have gone somewhere else.
Three waves, one shared trait. None of them was about doing the difficult work. They were about taking over the repetitive work — and thereby freeing people up for what genuinely required a human.
Another shared trait: all three met the same objection. It will not work here. Our production is too particular. And what will people do instead? The outcome was the same every time. The companies that took the step grew. The ones that waited had to compete against the ones that had.
The robots came to the shop floor — and stayed there
There is an imbalance here that few talk about.
Walk into a Swedish manufacturer today and look at production. It is most likely a well-measured, well-optimised operation. Changeover times have been cut for decades. Material flows have been redrawn. Cycle times are measured in seconds. Every investment was justified on a calculation someone could defend in a board meeting.
Then walk twenty metres across, into the office.
There sits someone reading an order in an email window and typing the same details into the ERP. There sits someone else comparing a supplier invoice against a purchase order, line by line. There is a queue of unconfirmed orders that grows every holiday season. There is the same status question from the same customer taking three minutes to answer, for the twentieth time this month.
Nobody has measured those hours. Nobody has costed them. Nobody would accept the equivalent waste in production.
It is not carelessness. It is that administrative work has been hard to automate when the source material arrives as free text in an email. Structured flows like EDI have worked for decades, but a task where a person has to read an email in German, understand that the customer means the item they bought in March, check the price against the agreement and register the order has required judgement. And judgement has not been available for purchase.
That is the condition that has changed.
Where the hours actually sit
Do the arithmetic on your own company. In a manufacturer with fifty employees it is not unusual for five to ten people to work with orders, purchasing, delivery monitoring, customer support and finance. If a third of their time goes to moving information between email and the ERP, that is two to three full-time roles producing nothing — they are transporting data.
It is the same waste as transporting material between stations. The difference is that nobody has drawn a value stream map over it.
Those tasks look essentially identical in every manufacturing company:
| Task | What it consists of |
|---|---|
| Order entry | Interpret incoming orders, match to the right item, register, confirm |
| Purchasing | Create purchase orders, chase confirmations, follow up on silence |
| Delivery monitoring | Reconcile promised dates against real ones, link delays to the customer order |
| Customer support | Answer the same questions about order status and lead times |
| Finance | Match supplier invoices against orders and goods receipts |
None of it is hard. But all of it is recurring, always running and impossible to set aside.
Monitor ERP — where the work happens
For a large share of Swedish manufacturing, all of this happens in Monitor.
The same year the IRB 6 rolled out of ASEA’s workshop, Monitor was founded in Hudiksvall. It has shipped production planning systems since the early 1980s, and the system is built specifically for manufacturing companies, not as a general ERP platform later adapted to industry. Material planning, manufacturing orders, post-calculation, traceability and stock hang together in a way that mirrors how an engineering business works. That goes a long way to explaining why the system took hold with so many Swedish industrial companies.
It also means something important: the work is already structured. Orders, items, customers, agreements, suppliers, purchase orders and delivery dates already sit in a system with clear rules. The problem is not that the information is missing. The problem is that someone has to key it in, pull it out and compare it.
That is exactly what a digital coworker does in Monitor:
- Registers orders — interprets incoming orders and call-offs in any format, matches them to the right item and customer agreement, and creates the order with the right quantity, price and delivery date.
- Validates — checks price, terms and delivery address against agreements and order history, and stops at anything that deviates instead of guessing.
- Confirms — sends the order confirmation the same day, in the customer’s language.
- Monitors purchasing — chases unconfirmed purchase orders and reconciles promised dates against real ones.
- Connects the chain — sees which customer order is threatened by a delayed inbound delivery, and raises the alarm while there is still time to act.
- Answers customers — pulls order status and lead times straight from the system instead of someone having to go looking.
The work happens in Monitor, with the same permissions and the same audit trail as for a human colleague. No new interface, no parallel database, no migration. Anything calling for judgement is handed to a person, with the background gathered.
Bruzaholms Bruk, in Bruzaholm since 1660, did exactly that with its order flow. Fårbo Manufacturing did the equivalent with three digital coworkers across different parts of the chain. Both in Monitor, both without changing systems. We have collected how it went under customer stories, and the technology behind the connections is described under integrations.
The lesson from the robot cell
There is one more thing to take from robotisation, and it may be the most important.
No Swedish workshop adopted robots by deciding to “become a robotics operation”. They bought a robot for one station. They picked the station where the work was most repetitive, most punishing and easiest to measure. They ran it, measured the outcome, and took the next station once the first one paid for itself.
The same applies now. A company setting out to “introduce AI” has no idea where to start, how to evaluate it or who owns it. A company deciding that order confirmations will go out the same day has a measurable target, a clear boundary and a person who notices the difference.
The second lesson: the robot did not replace the welder. It took over the repetitive seams, and the welder went from running the same seam all day to running the cell, handling deviations and assuring quality. The work moved up the value chain.
The office looks the same. The person registering orders today can spend that time on customer dialogue, quotations and deviations instead — the work that genuinely requires a human and that there has never been time for.
Competitiveness is a ratio
A Swedish industrial company competes with companies in countries where labour costs a fraction. That equation resolves in one way only: every hour here has to create more value than every hour there.
In production, Swedish industry has driven that ratio for 150 years, wave after wave. In administration it has moved a great deal more slowly. Much the same number of people perform much the same manual steps as twenty years ago — just in a newer interface.
That is where the headroom is now. Not because administration matters more than production, but because it is the part of the business that never got its robot.
Companies taking that step will not experience it as a revolution. They will experience it as orders confirmed the same day, no deliveries slipping without warning, and order intake growing without the office having to grow with it.
Just like last time.
Want to know where it would pay off most for you? Start from the operation, not the technology — there is more on that in The operational perspective and How to get started with AI in your business.