"We are busy enough with the production side of things," Inna Hilgenberg told DW. She's the deputy head of the packaging department at Dr. Wolff, a Bielefeld-based midsize cosmetics and pharmaceutical company. Behind her, black and red shampoo bottles are filled while they glide down the conveyor belt. "Writing instructions while busy with these tasks is not that easy," she added.
That's why Hilgenberg has come to appreciate her company's very own artificial intelligence (AI)tool WolffGPT, which she uses practically every day. Her job involves scheduling and instructing staff in the filling and packaging department, as well as ensuring hygiene and safety standards are met. She relies on WolffGPT to write workflows, design presentations edit Word files and Excel spreadsheets. "I only see advantages in AI," she told DW.
Many of the company's employees had already started using ChatGPT in their free time and would have liked to use it at work too. However, no company data can be shared with public large language models, or LLMs. That's why Dr. Wolff developed an in-house AI model on the basis of existing LLMs. It's connected to selected company systems and data and, therefore, sensitive company data remains protected.
In Germany, large companies are more likely to use AI than smaller ones. Manufacturers most commonly make use of AI to generate texts, images and code, according to a future technologies study conducted by ifaa, a Düsseldorf-based institute of applied work science. The study found that over 40% of companies surveyed already use AI, with 37% planning to do so.
Increasing competitive pressure is forcing many small and medium-sized enterprises (SMEs) in the manufacturing sector to consider using AI to boost productivity and efficiency, according to the IfM, a German institute that looks into small and medium-sized enterprises. It found that reservations in the workforce had slowed down the widespread adoption of AI.
Humanoid robots steal the show at German trade fair
To view this video please enable JavaScript, and consider upgrading to a web browser that supports HTML5 video
Dr. Wolff staffers embrace AI
Hilgenberg was one of first staffers to embrace AI at Dr. Wolff. She completed an online training course at the company's AI Academy and now supports others in her team who want to use the technology. The company has about 110 such AI pioneers among its 930 employees worldwide.
"They are ordinary colleagues who don't necessarily have to have a technical background," the company's AI manager, Zhuo Li, told DW.
These AI pioneers spend around 10 hours watching short instructional videos and completing small tasks to get up to speed.
Topics covered include how to craft good prompts and what happens to data in LLMs. They then spend part of their time identifying potential AI applications in their day-to-day work, and developing and testing prototypes for them.
They also assist colleagues who want to learn more about AI. "We have been able to find people in each team who have good ideas, a keen interest in the topic and a desire to pass on their knowledge," Li told DW. "We encourage everyone to try out what AI can do."
A silver bullet?
SMEs are tapping into AI to offset the skilled workers shortage and attract young talent. According to the ifM, LLMs allow companies to complete certain tasks and thereby reduce the workload on employees, as well as the need for specialists.
The institute found that 25% of surveyed companies already apply AI. One painting contractor, for example, used a robot to prime large surfaces, while digital assistants help to plan transport routes in the refrigeration sector.
Across Europe, only about 6% of SMEs currently use AI, according to OECD figures. A 2024 OECD study found that sluggish digitalization, poor connectivity and the misunderstanding of the potential benefits of AI were some of the barriers for why companies were not embracing LLMs. But the IfM also said some employees may already be using AI chatbots without informing their employers.
Is this Germany's most advanced AI factory?
'AI agents in use in almost all areas'
Dr. Wolff also teaches its employees how to create AI agents using the in-house WolffGPT platform to perform specific tasks. These include answering typical IT support queries, standardizing international invoices and creating social media posts.
"We already have AI agents in use in almost all areas," Zhuo Li told DW. "While some agents are only useful for two people, others are shared with entire departments," Karoline Bauch, an AI communication expert in the company's human resources department, told DW. "They make work much easier." She added that she had already built several AI agents herself.
"There is a high degree of acceptance," Bauch said. "When WolffGPT was down for several hours, we all realized how much we already rely on it for work."
Almost 90% of all employees who work at a computer workstation have completed the AI training modules since fall, according to company figures. "These courses are completed during working hours," Bauch explained. "That's why we have kept the lessons short so everyone can complete them at their own pace."
Administrative, creative tasks are ideal for AI applications
So far, AI has mainly been used at Dr. Wolff to support administrative and creative tasks, though LLM applications in the production sector are in the works.
Anyone in the company can take AI classes, as a basic understanding of LLMs benefits everyone, whether this knowledge is used at work or in a private context. Factory workers and warehouse employees just have to find a computer workstation.
Dr. Wolff is keen to see its staff learn about and apply AI, as the family-run business continues to face off competition from powerful corporations in Europe, Asia and the US.
This article was originally written in German.
