AI, SCALE ME UP!
German companies are increasingly turning to AI-driven automation. However, for this to yield economic benefits, a number of hurdles must be overcome. A management and technology consultancy based in Düsseldorf and a professor who conducts research into AI and automation technology at Düsseldorf University of Applied Sciences offer their insights.
Nils Kolwes advises Cassini Consulting AG clients on automation and AI and also supports the company’s internal AI development to optimise its own management and technology consulting processes.
The launch of ChatGPT at the end of 2022 marked a milestone in the development of AI. For the first time, an AI assistant became available to the general public that provided flexible, context-sensitive answers to questions rather than predefined standard information. Tech giants such as Google and Microsoft followed suit with their AI assistants Gemini and Copilot; today, new models are flooding onto the market every day. This rapid progress offers German companies many opportunities, but also presents challenges. According to the latest figures from a Forsa study commissioned by the TÜV Association, 56 per cent of the companies surveyed use generative AI tools, particularly for process automation. However, 50 per cent also see a high or very high need for further training in the field of AI. Most companies are unable to meet this need. Only 27 per cent stated that employees had already attended AI training courses.
Targeted AI integration and empowering employees to make effective use of the new tools – this is where Düsseldorf-based management and technology consultancy Cassini Consulting AG comes in. Since its foundation 20 years ago, it has been guiding medium-sized companies, large corporations and the public sector through the digital transformation. “Automation and AI have really picked up pace. We’re seeing this in customer enquiries too,” explains Nils Kolwes, AI & Automation Consultant at Cassini. “Many are realising that AI isn’t a passing trend, but a technology that’s here to stay. Yet most don’t know where to start.” The challenges are not purely technological, but also cultural: “For many, everyday use represents a major shift. Some are afraid of change and feel overwhelmed. Others are open to it, but need to be guided in order to use the tools effectively,” says Kolwes.
Cassini provides support through consultancy services and workshops. For example, staff are trained in the use of AI. A prerequisite is a regulatory framework that is binding for everyone. Among other things, this framework specifies who is permitted to use which AI tools and how, which data may be accessed, and how the quality of the results is ensured. Cassini also provides support during the implementation of AI solutions: “At the outset, we work together to identify which use cases could actually deliver added value. One example is manual invoice processing – a major time-consuming task when dealing with 500 documents a day. AI is very good at reading handwritten but digitised documents and placing them in the correct context. Together, we then work out what information and instructions are needed so that, for example, an AI agent can take over the task. This reduces the effort and thus the costs enormously.” The aim, says Kolwes, is always to enable companies to develop their own AI solutions. “This requires a central point of contact where employees can test the feasibility of their ideas and then have them implemented if appropriate – because automation starts with each individual,” he explains. This is an approach that Cassini Consulting follows itself, too – in its own AI workshop, numerous ideas have already been turned into small AI tools that have made a significant impact on day-to-day consultancy work. Among other things, incoming customer enquiries are automatically matched against staff profiles to identify the perfect ‘match’ in a matter of seconds.
Dr Dorothea Schwung is a researcher and lecturer in Artificial Intelligence and Data Science at Düsseldorf University of Applied Sciences.
Whether a company and AI are a good fit depends on the economic incentive, says Prof. Dr Dorothea Schwung, who holds a professorship in the newly established field of Artificial Intelligence & Data Science in Automation Technology at Düsseldorf University of Applied Sciences, HSD, (Department of Electrical Engineering & Information Technology). She explains: “Large companies in logistics, mechanical engineering, the semiconductor industry or the automotive sector have it easier, as their investments in AI pay for themselves quickly. Added to this is the fact that production in these sectors is already highly automated and there is sufficient data to train AI.” Although mediumsized companies are also increasingly turning to AI, they have to define their economic objectives for investment all the more clearly given their smaller budgets: “I need something that I want to optimise, and where AI has a tangible effect. This could be shorter lead times, reduced waste or higher product quality,” says Prof. Dr Schwung. It is also important to note that the use of AI is not appropriate for every problem: “Where a problem cannot be analysed using traditional methods, AI can be helpful – particularly if there is a good data set or suitable simulation models available.”
She cites predictive maintenance as an example of AI automation in manufacturing. In predictive maintenance, machines are monitored using AI and sensors to detect and prevent impending failures at an early stage, before they occur. This reduces downtime, lowers repair costs and extends the service life of machines. In her research, Prof. Dr Schwung focuses on AI-supported methods for multi-agent systems. Among other things, these increase efficiency in logistics when drones and mobile robots coordinate themselves, plan routes independently and communicate with one another.
New AI solutions often emerge from collaboration between the research sector and industry. Companies can apply for research funding programmes by submitting proposals for specific projects. This requires a clearly defined research question, project idea and project objective. Consortia are frequently formed between a university and one or more industry partners, who work together on a project. In addition, universities can operate as service providers in the market and undertake contract research projects. For Germany, Prof. Dr Schwung sees the greatest AI potential in traditional sectors such as mechanical engineering and the automotive industry: “We have excellent engineering expertise in these fields. But AI also requires technological and infrastructural prerequisites such as servers and cloud systems. Unfortunately, we are lagging behind in this regard in Europe. And AI education through degree programmes has only recently been launched.” Through her work in research and teaching at the HSD, the professor aims to continue helping Germany catch up in the international AI race. •
Text: Tom Corrinth
Pictures: CASSINI
