AI agents in action: Kufstein Uni is hosting an AI workshop
- 19.02.2025
- General
Alexander Schiegl, a graduate in Business Engineering, explained the use of AI agents in marketing. The workshop at the University of Applied Sciences Kufstein Tirol offered practical insights into the world of artificial intelligence.
AI meets marketing: at a workshop at the University of Applied Sciences in Kufstein, participants learned how AI agents are revolutionizing marketing processes. An insight into the future of AI-supported communication.
At the end of January, participants in the Virtual Marketing Company workshop at the University of Applied Sciences Kufstein Tirol delved deep into the world of artificial intelligence. Organized by the University of Applied Sciences' AI Competence Center, the workshop provided a practical introduction to how AI agent frameworks can be used in combination with modern large language models (LLMs). LLMs are powerful AI programs that understand language and can generate text. They are trained with huge amounts of text and enable applications such as automated text analysis, customer communication or content creation.
Numerous representatives from regional companies that are already part of the new Kufstein AI competence center took the opportunity to find out about the latest developments.
AI AGENTS AS THE KEY TO AUTOMATION
The workshop was led by Alexander Schiegl, a graduate industrial engineer (FH) and research associate at the Josef Ressel Center Vision2Move. In his introduction, he explained the basic concepts behind AI agents: “Basically, AI agents are specialized systems that can be entrusted with certain tasks, similar to human employees. They have a defined role and act according to their specifications.”
AT ITS BASIC CONCEPT, AN AI AGENT IS A SPECIALIZED SYSTEM THAT, LIKE A HUMAN EMPLOYEE, CAN BE ASSIGNED SPECIFIC TASKS.
Dipl. Wirt. Ing. (FH) Alexander Schiegl
Research associate at the Josef Ressel Center Vision2Move
Using a simulated marketing company, Schiegl illustrated how various AI agents can work together in a structured hierarchy. For example, a virtual CEO Managing Director takes over strategic planning, while specialized agents are responsible for social media, PR or radio advertising. The course uses OpenAI technology and also addresses the possibility of using self-hosted AI systems for the tasks.
A key element of this technology is the tight integration of AI agents and LLMs. The AI agent acts as a higher-level entity that guides the thinking process, defines tasks, and sets the context. In this way, the LLM is not considered a mere text generation tool, but an integral part of the thinking process that takes over the content realization – such as formulating a social media post or a press release. The agent formulates requests to the LLM, assesses the generated content and adjusts it if necessary. This creates a dynamic, cognitive work process in which both components interact to ensure that the results meet the desired objectives.
“INFORMATIVE AND INSIGHTFUL”
Diana Baldinger, BA, from Sparkasse Kufstein, was enthusiastic about the workshop: “The workshop was very informative and insightful, and the speaker was able to answer all our questions thanks to his high level of skills / competencies in the field. In our company, I see the concrete application of AI-supported marketing solutions in supporting our marketing department in creating content for social media, although this is initially possible with existing AI applications.”
PRACTICAL INSIGHTS INTO THE DEVELOPMENT OF AI SYSTEMS
The workshop offered a mix of theoretical introduction and practical implementation. After an introduction to the architecture of AI agents, the participants created their own marketing agents in Google Colab – a free programming platform that makes it possible to develop AI-generated content step by step.
A central topic was the cross-platform consistency of advertising messages. “The challenge is that texts and content for different channels – such as radio advertising, video or newspaper articles – have to be structured differently. The AI agents have to learn to adapt to the respective style and the requirements of the platform,” explained Schiegl.
Focused work: Participants implement what they have learned directly and develop their own AI-supported marketing strategies.
Ethical questions and data protection concerns were also addressed. One important aspect for companies is the choice between commercial AI solutions such as OpenAI and open-source models such as LLama, which have to be self-hosted. Schiegl emphasized that well-thought-out data protection strategies are particularly necessary for sensitive company data.
THE AI COMPETENCE CENTER AS AN INNOVATION DRIVER
The workshop also highlighted the role of the recently established Kufstein AI Competence Center. The center enables companies to benefit from the research expertise of the University of Applied Sciences Kufstein Tirol and to develop practical AI solutions for their own business processes.
Diana Baldinger emphasizes the importance of the competence center: “By participating in the AI competence center, we want to actively support the digital transformation of the regional economy. With our involvement as an active partner, we want to help to successfully implement the project and thus jointly generate long-term benefits for the entire region.”
The event impressively demonstrated how AI agents can already automate and optimize business processes today. Participants left the workshop with valuable insights into the practice of AI-supported marketing automation – and with ideas on how they can implement what they had learned in their companies.
TO BE CONTINUED?
When asked whether further workshops would follow, Schiegl was open: “We will listen carefully to the topics that particularly interest the participants and offer further training if there is sufficient demand.” This means that the AI Competence Center of the Kufstein University will remain a central point of contact for companies that want to develop in the field of artificial intelligence.
Links: