Game plan consulting without any effort, complex explanations of funding programs without personnel - sounds too good to be true, but it is a reality. Retrieval Augmented Generation (RAG) is an advanced development of AI that combines generative AI models with extended information retrieval capabilities. RAG also offers promising prospects for theaters, museums, libraries and archives to support creative processes, improve the audience experience and make operational processes more efficient. At the Österreichische Theatertechnische Gesellschaft (OETHG) in Salzburg, project manager Franziskus Linsmann presented the opportunities of RAG for culture.
RAG enables large language models (LLMs) to access relevant internal document collections and generate more precise and context-related answers on this basis. This further development of AI, which is available as a project at ChatGPT or as a “space” at Perplexity, for example, addresses one of the biggest challenges of previous language models: instead of random answers from the internet, RAG provides targeted information from specific, defined data sources. This increased precision significantly increases reliability and applicability.
actori was able to present the possible applications of RAG in the cultural sector at the OETHG management meeting in Salzburg at the beginning of March 2025. The focus was naturally on the theater sector. For example, AI-supported digital theater guides fed with information can accompany viewers through productions, provide interactive background information and thus deepen the understanding and appreciation of what is presented. Dramaturgs and marketing departments also benefit, as the AI can build on texts that have already been generated. For example, RAG could automatically create marketing materials and program descriptions in a consistent style, making editorial work easier.
Potential in operational processes too
In addition to creative and audience-specific applications, RAG also offers considerable potential for optimizing operational processes. For example, an AI-supported internal chatbot can preserve and pass on institutional knowledge in order to counteract the high staff turnover in cultural institutions. It is also conceivable to develop a data-based controlling system with RAG that provides insights into key business figures and thus facilitates well-founded decision-making. actori itself relies on RAG to support the Kulturplan Lausitz project fund, where a trained chatbot answers complex questions about the call for funds and financing.
Challenges and ethical aspects
Despite these opportunities, the integration of RAG into the cultural sector also brings challenges. Data protection remains a key issue: institutions must ensure that the processing of visitor data complies with strict German data protection guidelines. At the same time, it is important to find a balance between technological innovation and the preservation of theatrical traditions. Quality assurance remains essential - RAG-generated content should always be subjected to a critical review by specialist staff.
Used correctly, Retrieval Augmented Generation can enrich the German cultural landscape in the long term. Those who use RAG in a targeted and responsible manner can strengthen their own position as a cultural institution and at the same time build sustainable digital structures. actori has already successfully implemented RAG-driven chatbots in cultural institutions. With tailor-made digitization strategies, strategic consulting and innovative methods, actori supports your institution on its way to becoming an AI-supported institution (team@actori.de).
A contribution from Franziskus Linsmann, project lead