How AI’s efficiency can be optimised in the gastronomy

Oulu / FI. (atl) In May 2024, the University of Oulu in Finland has accepted the dissertation on data-driven decision-making (DDDM) by Researcher Nada Sanad, PhD. The Finnish lunch restaurant company Antell was involved in the study, further developing ATSO, Antell’s own AI tool, in particular with regard to its decision-making development and feedback collection.

(Photo: Antell Oy)

Antell’s ATSO AI tool was originally taken into use in 2022. ATSO uses restaurant-specific data, such as the consumption of different dishes, portion sizes, and special diets, to create menu suggestions that are then reviewed and accepted by the restaurant managers. ATSO has also been previously used for planning more environmentally friendly menus in Turku, Finland.

As part of her dissertation, Sanad has developed a decision-making model that helps organizations plan and use AI-powered decision-making systems, enhance collaboration between humans and AI, and support decision-making and evaluation. Decision evaluation plays a crucial role, as it enables continuous learning and takes feedback into account when making future decisions.

When it comes to collaboration between humans, learning from mistakes is relatively easy, but when AI comes into play, collecting feedback and taking multiple different variables into account becomes more difficult. The decision-making model created by Sanad helped Antell understand the different phases of the decision-making process better than before. Participating in the study also made the role and use of the ATSO tool clearer. According to Antell, observing and analyzing causal relationships is now easier, too.

«Thanks to the study, we were able to clarify which things play a role in decision-making, what the role of data and our AI tool is, and how important decision evaluation is at different parts of the process. We implemented feedback surveys that helped us further develop our own AI tool,» says Kaisa Puolitaival, one of Antell’s Development Managers who took part in the study. «When implementing AI tools, companies often face overly optimistic expectations. In the end, all the data needed isn’t always available, and AI isn’t able to forecast every little thing. We now consider ATSO to be more of an AI-powered assistant, and the final decisions are still made by restaurant managers and product development managers.»

So far, the possibilities of AI have only been studied in fields like healthcare, IT, and the manufacturing industry, and the restaurant industry’s vast possibilities and potential have previously been overlooked.

«The benefits that AI can bring into the restaurant industry are unique and concrete. Compared to humans, AI is able to process a lot more data, which makes planning, for example, varied and versatile menus easier than before. In addition, Antell’s AI tool ATSO is able to choose dishes with a smaller carbon footprint and keep a closer eye on the restaurant’s budget,» explains Nada Sanad.

Puolitaival also emphasizes the impact that new tools have on the professional identity of restaurant industry employees; «Companies need to be aware of the fact that implementing new technological tools gives rise to preconceptions and emotions from all sides. Restaurant managers and product developers often have a very strong professional identity based on their skills, which should not be forgotten.»

Antell also encourages other restaurant businesses to think realistically about the benefits of AI. Puolitaival recommends adopting the decision-making model developed by Sanad in the early stages of development.

«Different decision-makers have different aspirations when deploying AI tools. Businesses should not be afraid to adopt AI, but it must be carefully planned. The decision-making model can be used to break a project down into smaller, more easily understood parts. This makes executing the project clearer and more effective,» Puolitaival explains.