Dau, D., Steuck, P.-F., Wartenberg, M., & Knackstedt, R. (2026). Toward a Multi-Agent Architecture for GenAI-Powered Conference Companions. Proceedings of the 28th International Conference on Human-Computer Interaction, HCI International 2026, Montreal, Canada, July 26–31, 2026, Lecture Notes in Computer Science. HCI International 2026.

Abstract. Scientific conferences face a critical challenge: participants must nav-igate fragmented information across proceedings, schedules, and programs under severe time constraints. Existing Generative AI (GenAI) solutions often address isolated tasks but lack integrated support for diverse participant needs. This paper presents a technology-agnostic multi-agent architecture for GenAI-powered con-ference companions derived from two Design Science Research cycles. Based on analyses of conversation logs, questionnaire responses, and participant discus-sions, we identify seven recurring clusters of participant information needs and map them to specialized agents within a five-layered architectural framework. The core of this system is a Multi-Agent System (MAS) in which a central or-chestrator agent coordinates a network of specialized agents, each optimized to access specific hybrid data sources—ranging from structured relational databases for logistics to unstructured vector stores for scientific content. We contribute abstract design knowledge enabling developers to implement systems tailored to specific contexts while remaining independent of rapidly evolving technologies. By focusing on the underrepresented perspective of participants, this work ex-tends Human-Computer Interaction (HCI) knowledge for information-dense en-vironments and provides practitioners with a systematic blueprint to prepare nec-essary data infrastructures, significantly lowering barriers to adoption.
 

Keywords: Generative AI, Human-AI Interaction, Scientific Conferences, De-sign Science Research.

--

Fröhlich, N. A., Gottschewski-Meyer, P., & Strohmann, T. (2026). Design Features for Virtual In-Vehicle Assistants in Autonomous Public Transport. Proceedings of the 28th International Conference on Human-Computer Interaction, HCI International 2026, Montreal, Canada, July 26–31, 2026, Lecture Notes in Computer Science. HCI International 2026.

Abstract. The transition towards autonomous public transport (APT) is considered a key lever for achieving ecological and social sustainability. However, passenger acceptance is a critical prerequisite for the successful deployment of APT, particularly given the absence of a human driver. A potential to foster acceptance is the use of virtual in-vehicle assistants (VIVAs), as they offer a promising means of compensating for a possible lack of communication, support, and passengers’ subjective sense of safety, particularly in situations involving unexpected system behaviour. This study addresses this potential by developing and evaluating concrete design features (DFs) that operationalize existing design principles (DPs) for comprehensive VIVAs in APT. Following a design science research approach, the DFs were validated through an iterative evaluation process including six think-aloud sessions with experts and potential users. The study identifies in sum 32 DFs, 17 fundamental DFs and 14 supplemental DFs, emphasising personalized yet restrained communication, transparency to foster trust, clear emergency guidance, adaptive support for vulnerable users, and cybersecurity measures. A key finding is the distinction between fundamental and supplementary DFs, with users consistently favoring functionally simple and non-intrusive designs. This work contributes prescriptive and reusable design knowledge by systematically translating abstract DPs into actionable DFs, thereby bridging the gap between theory and implementation.

Keywords: Autonomous Public Transport, In-vehicle Assistants, Conversational Agents, Design Knowledge, Design Science Research, Design Features.

--

Gottschewski-Meyer, P., Lang, F., Auf der Landwehr, M., & Knackstedt, R. (2026). Assessing the Influence of Personality on the Effectivity of Digital Nudges. Proceedings of the 28th International Conference on Human-Computer Interaction, HCI International 2026, Montreal, Canada, July 26–31, 2026, Lecture Notes in Computer Science. HCI International 2026.

Abstract. Slight adjustments to the digital choice environment, so-called digital nudges, can change the behavior of users in the digital sphere. Individual personality can also impact user behavior, but the personality-related effectiveness of digital nudges is yet understudied, and existing research shows divergent results. The objective of this exploratory study is to investigate whether individual personality influences the efficacy of default, anchoring, and hyperbolic discounting nudges with and without pro-environmental assertions. It contributes to the existing body of knowledge on the interplay between digital nudges and personality and demonstrates that specific personality facets influence the efficacy of digital nudges. Furthermore, our study demonstrates the importance of considering personality facets rather than personality traits, as this avoids potential problems of mutual cancellation and could lead to inconclusive study results.

Keywords: Digital Nudges, Personality Traits, Big Five, E-grocery.

--

Steuck, P.-F., Bierschwale, D., Schmidt, R., Shidujaman, M., & Knackstedt, R. (2026). LLMPT: A Template for Prompt Documentation in Scientific Publications. Proceedings of the 28th International Conference on Human-Computer Interaction, HCI International 2026, Montreal, Canada, July 26–31, 2026, Lecture Notes in Computer Science. HCI International 2026.

Abstract. The rapid adoption of generative AI in scientific research creates a need for transparent documentation of prompts, model configurations, and devel-opment decisions. Publication practices are inconsistent, with many studies providing fragmented insight into how large language model (LLM) interactions shape research outcomes. This study addresses this gap by introducing LLMPT, a standardized template for documenting prompts and configuration parameters in publications. Following a Design Science Research approach, we conducted a systematic literature review on existing documentation practices and a structured screening of commercial conversational agent interfaces and open-source LLMs to identify relevant technical parameters. The resulting template supports both one-off prompt documentation and iterative refinement, enabling researchers to track changes in prompts, model versions, hyperparameters, and additional func-tions over time. LLMPT accommodates diverse usage contexts, ranging from web-based interfaces to API-based deployments, and remains compact enough for practical integration into publications. A focus-group evaluation with re-searchers experienced in GenAI-based work demonstrates the template's utility, highlights the need for such structured documentation, and confirms its potential as an interdisciplinary tool for improving transparency and methodological rigor.
 

Keywords: Prompting, Large Language Models, Generative AI, Scientific Documentation.

--

Walter, D., Steuck, P.-F., & Knackstedt, R. (2026). Exploring Informal Caregivers’ Experiences with a Context-Sensitive Voice-Based AI Companion. Proceedings of the 28th International Conference on Human-Computer Interaction, HCI International 2026, Montreal, Canada, July 26–31, 2026, Lecture Notes in Computer Science. HCI International 2026.

Abstract. Informal caregivers provide the majority of long-term care for older adults and individuals with chronic conditions, often managing complex care tasks alongside emotional support and coordination responsibilities. This role is associated with substantial emotional, physical, and psychological strain, and caregivers often struggle to access timely guidance. Digital support for informal care is heterogeneous and fragmented, and existing tools typically focus on dis-crete functions and often fail to address caregivers’ evolving and diverse needs. Building on this gap, this paper focuses on the evaluation of an initial voice-based generative AI prototype designed to provide context-sensitive, low-threshold support. The system leverages natural voice interaction and anthropomorphic features to deliver tailored guidance. Semi-structured interviews with informal caregivers were conducted to explore perceptions of usefulness, understandabil-ity, and emotional experience. The findings inform the human-centered design of AI companions, illustrating how generative AI can provide empathetic, situation-ally relevant support and enhance the everyday caregiving experience.

Keywords: Informal Care, Anthropomorphism, Design, Voicebot.