Swiss pragmatism, AI and tech.

In today’s Heroes of Innovation series on the main stage we have a small manufacturing company from Switzerland that is taking very bold steps in implementing AI and other tech solutions in their operations. I’m always looking to find examples of companies like this one and showcase them to the world for others to see the benefits of technology on manufacturing businesses especially when it is done in the right way and when at the help of the company sits a great operator and CEO like Achim Schroeter who impressed me with his Swiss pragmatism combined with a forward looking approach and vision that waiting is not the answer when it comes to AI and other tech so you have to start somewhere, you can’t just sit and wait until the use cases are defined, you have to explore and find the right use cases for your team and company.

Buckle up and be prepared to take notes because there are a lot of things to learn from this company and for many manufacturing companies of similar size like Urimat this is exactly what you need to read especially now during the times we’re going through.

So let’s go 🚀

When the market won’t wait.

“Do it in small steps.” “The market is not waiting for us.”

That line said by Achim carries almost the whole story. URIMAT is a small Swiss company in waterless urinal technology, with around 16 employees in Switzerland, another 12 in Germany, and exports into roughly 40 to 50 countries through long-term partners. When Achim talks about change, he talks like a CEO who sees margin pressure, slow internal work, and bigger competitors who can still be beaten if a smaller company moves faster.

You can hear that pressure clearly. The euro had weakened against the Swiss franc, most of the company’s costs are in Swiss francs, and export margins into Europe were getting squeezed. Achim’s answer was lean manufacturing, cost savings, supplier work, and digitalization, because he did not want to keep pushing price increases onto long-term distribution partners.

That matters because URIMAT is not a software company trying to look modern for the sake of it. It is a manufacturing business with its own tooling, outside production partners, end assembly, and all the little delays that pile up when information sits in too many places and people keep doing work the same way because that is how it has always been done.

The work changed before the tools did.

Achim had already been carrying lean thinking for years before AI entered the picture. He said that background came from earlier work after his studies, and when he arrived at URIMAT in 2018 he started pushing for leaner processes, lower finished stock, and more just-in-time thinking.

He described that early phase in very human terms. He went to the shop floor, talked with employees one by one, tried to understand each process as it was actually being done, and asked people to look at their own work and ask whether it was really lean, really efficient, really the best use of time and resources.

Some people changed quickly. Some did not. Achim says that is normal in a smaller company where some employees have been doing the same work for 10, 20, or 25 years, and where routines can harden into something that feels natural simply because it is familiar.

That same step-by-step rhythm shaped the company’s AI adoption. Achim said they started with ChatGPT, began with just a couple of people testing it, gave employees time to try the free version, paid for access when there was a real use case, and brought in outside training so the team could understand what was coming. His language around it stayed very plain the whole time: just test it, see what comes out, and learn as you go.

The first wins came from work that had been sent outside the company. Achim said URIMAT had been spending around CHF 10,000 to 15,000 a year on outside marketing agency work, and later estimated that AI was saving around 50 to 70 percent of the time on social media text, with some marketing tasks landing in the 50 to 80 percent range. In the follow-up interview he said a social media text that used to take about 30 minutes could now take around 10 to 15.

He also kept the limits in view. AI could do a lot of the first-pass work on text and product images, but he still talked about the human side, the taste, the feeling, the fact that you still have to think about what you want to say. That tone is part of why the story feels believable. He is not talking about magic. He is talking about work.

A quote should not take longer than the customer’s patience.

The next layer was ERP.

Achim said URIMAT had been running separate ERP and CRM systems, which meant two sources of customer information and too much friction around inquiries, offers, confirmations, stock checks, and production handoffs.

During our recent discussion he mentioned that the main Odoo implementation was done and stable, with smaller improvements still being worked through. Achim said the internal sales process was saving around 20 to 30 percent, and the time to prepare an offer had dropped from roughly 20 to 25 minutes, sometimes 25 to 30, down to about 5 to 10 minutes.

That sounds like a small administrative gain until you hear how he talks about speed. If a quote takes too long, you can lose the deal. In a company like URIMAT, speed is part of the commercial edge, and Achim comes back to that point more than once.

The rollout itself followed the same pattern as the earlier mindset work. The team had a test environment before the live system, people learned by doing, they asked questions inside the sandbox, gave feedback on processes, and reached the point where most questions could be answered internally rather than by outside IT support.

Achim also sees the ERP as a base for what comes next. He talked about future forecasting, automatic purchase orders, price comparisons, and the possibility of using company data more directly in later AI tools once the system has enough history and the smaller improvements are finished.

Hiring changed too.

One of the strongest parts of the discussion sits in HR because it shows AI inside an actual hiring workflow. Achim talked about an internal sales position that had been posted around 1.5 years earlier with a standard job description and drew about 10 applications, then posted again after the wording was reworked with AI and drew around 100.

He did not overclaim the result. He said the market had shifted too, more people were on LinkedIn, and the jump was probably not only about AI, yet he was still surprised by the amount and by the quality of the people who came in. His phrasing there matters, because it sounds like a manager describing what he actually saw, not someone forcing a lesson into the story.

Then they pushed one step further into the process. After choosing 5 or 6 people for first interviews, the team asked AI what kind of assessment fit the role, got back a multiple-choice structure, turned it into a Google Form, sent candidates the link, and used the responses as part of the screening.

That detail says a lot about how Achim thinks. He used AI to sharpen the job post, attract more applicants, and build a simple test tied to the role, but he kept the interviews and the judgment in human hands.

There was another small signal in the job post itself. Achim said they also wrote that candidates should be open to new artificial intelligence and internal trends, because for him that openness is already part of what a company needs from the people it hires.

What happens when the know-how walks out the door.

The deeper part of the conversation came when the subject turned to age, experience, and what happens when long-tenured employees retire. Achim said URIMAT has people in their 50s and 60s, and in one open position the company had already planned a handover where the outgoing employee stayed 3 more months to pass on the work to the newcomer.

Even there, he did not talk about transfer as simple copying. He said the new person should get the introduction, hear how things were done, and still stay open-minded enough to do some things in a different way, maybe even a better way. “Don’t copy and paste everything” is more or less the management logic underneath the whole discussion.

That is where the idea of an internal knowledge base starts to make sense. Achim talked about an internal knowledge center, maybe with something like Copilot later on, where procedures, reference documents, customer-specific changes, and old problem-solving knowledge could stay inside the company instead of leaving with the person who held it. In May, he described the future version in more detail: documents in a library, frequently asked questions, maybe bots, maybe a connection to ERP data, and one central place where people can ask what happened with a customer, what changed in a design, or how a similar issue was solved before.

He is not pretending that this is finished. He says the ERP and website work still come first, the database has to be in better shape, and the company still needs to think carefully about model choice, cost, and whether heavy-use cases belong on public systems or on something more local later on.

Start small.

Still, the direction is clear and Achim’s advice for other manufacturing companies is to start small, learn by doing, train people, keep the useful parts, and move fast enough that the company stays ahead of the drag that builds up inside old processes.
That is what makes this story worth reading for other manufacturers because Achim is not selling a theory here but he is describing what it looks like when a traditional business actually begins to change.

Thank you Achim for your insights!

Follow Achim Schroeter
Follow URIMAT

Subscribe to our newsletter https://notifiqation.beehiiv.com/subscribe
Follow us for latest updates and posts about upcoming articles on
https://www.linkedin.com/company/notifiqation/
https://x.com/Notifiqation_