Paper Number
ECIS2026-1619
Paper Type
CRP
Abstract
Cafeteria demand planning requires both algorithmic pattern recognition and human expertise, yet current systems treat these separately, which generates significant food waste. This paper reports on a 9-month action design research (ADR) project at a German financial services firm. Using a practice-driven abductive approach, we developed a collaborative forecasting system that leverages semantic processing using large language models (LLMs) to solve the “cold-start” problem for novel menu items while preserving human agency via override mechanisms. Our evaluation combines algorithmic benchmarking, reducing forecast errors by 30% over naive baselines, with two think-aloud sessions showing that human judgment remains critical for high-uncertainty events. We distill our findings into a meta-design and four design principles (DPs), grounded in kernel theories, for systems where human contextual intelligence and algorithmic recognition must coexist. We contribute to the discourse on human-AI collaboration and sustainable IS by providing a rigorous blueprint for designing synergistic, trustworthy, and diagnostic operational planning tools.
Recommended Citation
Cappel, Justus; Strohmann, Timo; Burger, Mara; Voss, Marleen; and vom Brocke, Jan, "Schnitzel-Prediction: Designing Human-Ai Collaboration For Cafeteria Demand Forecasting" (2026). ECIS 2026 Proceedings. 7.
https://aisel.aisnet.org/ecis2026/is_resil/isresilience/7
Schnitzel-Prediction: Designing Human-Ai Collaboration For Cafeteria Demand Forecasting
Cafeteria demand planning requires both algorithmic pattern recognition and human expertise, yet current systems treat these separately, which generates significant food waste. This paper reports on a 9-month action design research (ADR) project at a German financial services firm. Using a practice-driven abductive approach, we developed a collaborative forecasting system that leverages semantic processing using large language models (LLMs) to solve the “cold-start” problem for novel menu items while preserving human agency via override mechanisms. Our evaluation combines algorithmic benchmarking, reducing forecast errors by 30% over naive baselines, with two think-aloud sessions showing that human judgment remains critical for high-uncertainty events. We distill our findings into a meta-design and four design principles (DPs), grounded in kernel theories, for systems where human contextual intelligence and algorithmic recognition must coexist. We contribute to the discourse on human-AI collaboration and sustainable IS by providing a rigorous blueprint for designing synergistic, trustworthy, and diagnostic operational planning tools.