Artificial Intelligence-Supported Decision Support in Institutional Foodservice: A Plan-Produce-Monitor-Improve Framework
14. Uluslararası ACHARAKA Tıp, Hemşirelik, Ebelik ve Sağlık Bilimleri Kongresi, Baku, Azerbaycan, 16 - 18 Temmuz 2026, ss.655-660, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Basıldığı Şehir: Baku
- Basıldığı Ülke: Azerbaycan
- Sayfa Sayıları: ss.655-660
- Erciyes Üniversitesi Adresli: Evet
Özet
Institutional foodservice systems
are expected to provide nutritionally adequate, acceptable, feasible and safe
meals while also controlling production volume and reducing avoidable food
waste. In practice, decisions about menu design, purchasing, preparation,
portioning, consumption and leftovers are often handled as separate steps. This
separation limits the use of real consumption data in future menu planning and
may weaken the connection between nutritional quality and operational
sustainability.
This paper proposes a conceptual
framework for positioning artificial intelligence-supported applications as
decision-support tools in institutional foodservice. The framework is organized
around a four-stage cycle: plan, produce, monitor and improve. It connects menu
planning and optimization, AI-supported meal planning, image-based food
recognition, portion or volume estimation, multimodal nutrient-intake
estimation, demand-related production planning, plate-waste monitoring,
food-intake assessment. The aim is not to present these technologies as
independent innovations, but to show how they may contribute to a continuous foodservice
decision process.
In the planning stage, AI-supported
tools may assist the comparison of alternative menus according to nutritional
composition, menu feasibility, cost-related constraints, consumer acceptability
and environmental considerations. These tools do not replace professionals'
decision; rather, they can make trade-offs more visible when several goals need
to be considered at the same time. In the production stage, previous service
records and demand-related variables may support more accurate preparation
decisions and help reduce overproduction. In the monitoring stage, image-based
recognition, portion or volume estimation and multimodal assessment approaches
may provide more objective information on actual food intake, liquid leftovers
and plate waste than routine visual observation or retrospective records alone.
The improvement stage is the point
at which the framework becomes a decision-support cycle rather than a
collection of digital tools. Data obtained after service can be used to revise
menus, standardize portions, improve staff training, strengthen simulation-based
learning and identify workflow points where automation may support consistency.
In cafeterias and other institutional settings, it may support food-waste
reduction by showing where avoidable waste occurs and which menu or service
characteristics require revision.
The proposed framework places the
dietitian at the center of interpretation and implementation. Artificial
intelligence should be considered an auxiliary infrastructure that organizes
data and supports decisions, not an autonomous authority in menu planning or
nutrition care. Safe implementation requires local food and menu datasets,
validation under real service conditions, attention to portion diversity, data
privacy, institutional feasibility and compatibility with existing foodservice
workflows. The framework also underlines that different institutional settings
require different levels of evidence: a hospital meal tray, a school lunch menu
and a university cafeteria service cannot be evaluated with the same
operational assumptions. Within these conditions, AI-supported systems may help
institutional foodservice move from fragmented decision-making toward a
continuous plan-produce-monitor-improve model that links nutritional adequacy
with operational efficiency and waste reduction.
Keywords: Artificial intelligence;
institutional foodservice; menu planning; intake monitoring; food waste;
conceptual framework