AI-powered investment recommendations in the agri-food sector

Purpose This study aims to develop an intelligent, personalized investment recommender system for the agri-food sector by integrating adaptive neuro-fuzzy inference system (ANFIS) with behavioral finance. It focuses on aligning farmers’ financial management traits (FFMT) with suitable agricultural technology (agri-tech) investments, particularly drone technology. The system provides tailored guidance to enhance farmers’ financial decision-making and supports artificial intelligence (AI)-driven food marketing strategies. Design/methodology/approach This study applies a hybrid methodology that integrates fuzzy logic with machine learning. The dataset originates from an online investment questionnaire conducted in Hungary in 2019 (n = 1,542), made available to the authors under the framework of the 1.3.1-VKE-2018–00,007 project. Data were analyzed using JMP (K-means clustering) and MATLAB (for ANFIS modeling). Six financial management traits (FMTs) served as input variables, while investment types were used as outputs to define recommendation classes. After preprocessing, 79 valid input–output pairs were obtained for ANFIS, with 55 allocated for training and 24 for testing. Findings The K-means algorithm grouped investment options into three clusters: Cluster 1 (n = 592, 38.4%, cautious traditionalists), Cluster 2 (n = 610, 39.6%, passive moderates) and Cluster 3 (n = 340, 22.0%, active aggressive). The model generated personalized recommendations based on inputs such as safety perception, excess cash use and saving strategies. Among farmer participants (5.1% of the sample), 56.25% were male and 43.75% female, with 50% residing in Budapest. The FFMT–ANFIS model achieved robust performance on the training set (Root mean square error (RMSE) = 0.78) with ten-fold cross-validation (mean RMSE = 0.80, SD = 0.05). On the held-out test set, the model achieved an RMSE of 0.79 and an R2 of 0.875. After preprocessing and generating 729 fuzzy rules, the model’s effectiveness in producing accurate, behavior-driven recommendations was confirmed. Research limitations/implications The study is limited to self-reported behavioral data from Hungarian respondents and focused on drone investment scenarios. The relatively small share of farmers in the sample (5.1%) also limits external validity, which future research should address through stratified or field-based sampling. Broader validation across geographies and agri-tech domains is recommended. Future work should integrate real-time financial behavior and market responsiveness to increase system adaptability and generalizability. Practical implications This is the first study to integrate FMTs and ANFIS for investment decision support in the agri-food domain. It bridges gaps between behavioral finance, AI and food marketing, offering a replicable framework for behavior-driven agri-tech adoption. The model contributes to smart, data-informed and inclusive agricultural investment ecosystems. Social implications By promoting personalized investment literacy and tech adoption among farmers, this model fosters digital inclusion and supports sustainable food systems. It enables better access to decision-making tools, particularly for smallholders, reducing inequality and enhancing trust in AI systems used in agricultural finance and marketing. Originality/value This study presents a novel framework that integrates financial management traits (FMTs) and ANFIS for investment decision support in the agri-food domain. Although the previous research has explored similar adaptive and fuzzy-logic-based recommender systems in financial and agricultural settings, this research introduces an integrated FFMT–ANFIS framework tailored for investment decision support in the agri-food sector. It bridges gaps between behavioral finance, AI and food marketing, offering a replicable framework for behavior-driven agri-tech adoption. The model contributes to smart, data-informed and inclusive agricultural investment ecosystems.