12 February 2026 - Support & Artificial Intelligence (AI) heiBOT-IT: IT service around the clock
Following a successful test phase, the IT service chatbot “heiBOT-IT” is now expanding the URZ's support and AI offerings. Students and employees can use the open-source chatbot around the clock for IT questions. It has been technically and visually improved and, thanks to a special extension with URZ content, provides precise answers to IT-specific questions in German and English.
The new chatbot heiBOT-IT is based on the qwen3-32b model and has been specially equipped with content from the computer center, such as instructions and websites. heiBOT-IT can be accessed via the browser. The IT service bot is currently accessible via a link in the contact field and in the footer of the URZ homepage (red flag with envelope on the right-hand side of the screen and at the bottom of the website); on the URZ support page; and in the IT service contact form.
The chatbot can provide first-level support and responds exclusively in its role as an IT service agent to inquiries about IT problems. It is not designed for other areas of application. heiBOT-IT also offers a mobile view that is optimized for smartphones and works like the web version.
Digital, open-source-based IT service around the clock
heiBOT-IT is available 24 hours a day and provides answers to simple questions in seconds – without processing time, email correspondence, or visits to an IT service point. It covers URZ-related topics with a focus on IT services and uses content from websites and documents to provide targeted and efficient assistance. As with the university AI platform YoKI, the URZ relies on the use of open-source models. This keeps development transparent, adaptable, and independent of individual providers.
Chatbot features: An overview
What can you expect from heiBOT-IT?
- First-level support: Fast and accurate answers to questions about IT services—with references to relevant sources.
- Automatic support for common issues such as university ID, eduroam, VPN, or software.
- Speech recognition for German and English with responses in the appropriate language.
- Linked sources and instructions in a separate window for more detailed information or discussion of the query and chatbot responses. For example, a question about multi-factor authentication contains not only a short answer but also links to instructions and relevant sources on related topics.
- Focus on URZ services and instructions, no information about individuals or other topics.
- Comprehensible answers with numerical references to the sources used (maximum of five per answer for clarity).
Behind the scenes: How heiBOT-IT works
The university AI solution is based on the qwen3-32b language model, which has been specifically enhanced with URZ websites and instructions using Retrieval Augmented Generation (RAG) technology. A RAG model combines a generative language model with a retrieval system: it retrieves relevant information from precisely defined data sources – such as the URZ service catalog and instructions – and uses this to formulate natural, understandable support responses.
Around 450 pages from sources such as the service catalog and instructions were fed into the system. Duplicates were removed and outdated content was given less weighting to ensure accurate and up-to-date responses.
The chatbot supports students and staff with questions about the university's IT services—the aim is to answer queries simply, quickly, and reliably with relevant links.
Marcus Buchwald, member of the research group at the Engineering Mathematics and Computing Lab (EMCL)
Our RAG system combines a powerful open-source language model with targeted retrieval of existing knowledge on high-performance AI hardware operated by the URZ. This ensures that users always receive the latest information from over 450 websites.
Holger Altenbach, Service Group Manager at URZ
IWR and URZ combine expertise for AI services
heiBOT-IT was developed by Marcus Buchwald from the Engineering Mathematics and Computing Lab (EMCL) research group at the Interdisciplinary Center for Scientific Computing (IWR) and is operated by group leader Holger Altenbach (Core Technologies and Collaboration Services division) at the University Computing Center (URZ). The cross-institutional team is continuously working on further developing the chatbot, optimizing its functions, and replacing models with more powerful versions as needed.
With heiBOT-IT, the URZ is expanding its AI portfolio and supplementing the YoKI AI platform, which was also developed as a joint project, with an efficient open-source-based support service.
