Introduction

The postharvest management of cut flowers plays a critical role in enhancing the competitiveness of cut flowers and ensuring their quality and longevity (Dahal 2013). Vase life (VL) is the duration that a cut flower maintains its ornamental values and is a critical factor in consumer satisfaction determination (Rihn et al. 2014). The shorter the VL of the flowers purchased, the lower the possibility for repeat buying (Dennis et al. 2005;Rihn et al. 2014). Previous studies have shown that various factors, including water stress, environmental conditions, disease susceptibility, and ethylene damage, significantly influenced the VL of cut flowers (Fanourakis et al. 2013;Ha et al. 2022). Despite these challenges, the flower industry has improved storage conditions, packaging, and transport techniques (Thakur 2020;Rashed et al. 2024). However, the requirement for more advanced solutions remains, especially as consumers demand higher-quality flowers with longer VL (Dennis et al. 2005;Rihn et al. 2014). The cut flower industry has faced challenges in guaranteeing the quality and longevity of cut flowers during export and distribution. The current classification and grading system are time-consuming and labor-intensive and often overlook the main factors influencing the VL of flowers. Recently, artificial intelligence (AI) prediction for VL guarantee of cut flowers has gained attention as a potential technique to improve efficiency and ensure flower quality (Kim et al. 2024;Sun et al. 2021). However, the success of these technologies depends on the acceptance of customers and flower industry professionals. Thus, understanding the consumer perceptions and preferences in VL guarantee based on AI prediction is necessary to develop a practical VL guarantee system for the flower industry.

Consumers’ perceptions of AI technology vary widely (Bhatnagar and Sharma 2024). While some consumers believe AI is an effective tool that can enhance product quality, others are concerned about its reliability, cost, and the potential loss of traditional methods. The flower industry is no exception, with many flower industry professionals recognizing the benefits of AI but hesitating to use it without clear evidence of its effectiveness and reasonable cost. Therefore, this study surveyed and evaluated customer perceptions regarding AI applications in VL guarantees of cut flowers. We focused on general customer perceptions and flower industry professionals’ opinions in Osaka and Fukuoka which are major export markets of Korean cut flowers.

Methods

The survey data were collected from general customers and flower industry-related professionals in Osaka (October 2023) and Fukuoka (May and June 2024). The sample size included 102 general customers from Osaka (51 people) and Fukuoka (51 people), and 51 flower industry professionals in Fukuoka. The survey content related to key factors of customers such as sex, age, cut flower purchase preferences (purchase location and preferred flower type and color), willingness to pay more for flowers with a VL guarantee, and trust in AI application for predicting VL of cut flowers. Preferences regarding the duration of VL guarantees and the possibility of purchasing flowers if VL is guaranteed were also explored in the survey. Data were collected and analyzed using Excel and SigmaPlot version 14.0.

Results

Demographic characteristics of surveyed customers

In Osaka, 60.4% of respondents were female, while 39.6% were male. In Fukuoka, the gender distribution was more balanced, with 48.5% male and 51.5% female (Fig. 1A). A significant difference in age groups was observed. In Osaka, 68.8% of respondents were under 20, while in Fukuoka, 43.3% were in their 50s (Fig. 1B). The gender and age groups also differed between general customers and flower industry-related people (Fig. 2A-B). A higher percentage of flower industry-related professionals were in their 50s, with the majority being male (Fig. 2A-B).

Flower purchase behavior of surveyed customers

Data from Table 1 showed the preferences and behaviors of general consumers and flower industry-related people regarding cut flower purchases. The majority of general customers in both Osaka (90.0%) and Fukuoka (83.2%) purchased flowers from flower shops, with smaller percentages buying from wholesale markets or supermarkets (Table 1). However, flower industry-related people favored wholesale markets (100.0%) due to their professional needs, compared to general customers (Table 1).

The most popular flower type among general customers was roses, with 55.8% in Osaka and 48.8% in Fukuoka indicating a preference for roses. In both cities, pink flowers were the most preferred, and red flowers were the second most popular choice, particularly in Fukuoka. These behaviors were also similar between general customers and flower industry-related people (Table 1).

Consumer’s preferences regarding vase life guarantees

In Osaka, 58.3% of general customers indicated they like to purchase flowers if the VL was guaranteed, with 35.4% willing to pay more for such cut flowers. In Fukuoka, 73.5% of respondents showed interest in purchasing flowers with a VL guarantee, with 47.1% willing to pay higher prices (Fig. 3A-B). These results suggest that while both cities showed interest in VL guarantees, Fukuoka customers were more likely to accept such services. When asked about the preferred duration for the VL guarantee, 54.2% of Osaka consumers favored a 7-day guarantee, whereas 49.4% of Fukuoka consumers preferred a 10-day guarantee (Fig. 3C). This preference for longer guarantee durations indicates that the VL of cut flowers determines the consumer purchasing decisions.

Among flower industry-related people, only 27.4% indicated they would purchase flowers with a VL guarantee, with 19.6% willing to pay more. In contrast, 65.9% of general customers showed interest in such a guarantee. Preferences for VL guarantee duration also differed, with flower industry professionals showing a greater preference for 10-day guarantees (44.7%) compared to general customers, who favored 7-day guarantees (48.4%) (Fig. 4A-C).

Trust in AI technique for vase life prediction

Table 2 shows the differences in the trust in AI for VL prediction and willingness to pay more for AI-guaranteed VL between Osaka and Fukuoka customers as well as between general customers and flower industry-related people. The level of trust in AI technology in VL prediction differed between Osaka and Fukuoka customers. In Osaka, 61.7% of respondents expressed uncertainty about AI's reliability, whereas 52.9% of Fukuoka respondents believed it was reliable, and 20.6% of people highly believed AI in VL prediction (Table 2). Flower industry professionals expressed more trust, with 56.9% considering AI moderately reliable (Table 2).

Willingness to pay for AI-guaranteed vase life system

In Osaka, 57.1% of general customers were willing to pay up to a 10% higher price for flowers with an AI-guaranteed VL, while 22.5% were willing to pay 20% more. In Fukuoka, 38.1% of surveyed people were willing to pay 20% more, while a significant rate (34.3%) indicated they would pay 30% or more. Flower industry professionals were not willing to pay higher prices (≥ 30%), with most preferring minimal price increases (5 - 20%) (Table 2).

Discussion

Recently, with the increase in flower consumption through online markets, the VL of cut flowers significantly affects customer purchasing decisions (Bulgari et al. 2021;Gabellini and Scaramuzzi 2022;Rihn et al. 2014). Therefore, it is necessary to establish an accurate and effective quality guarantee system to ensure VL and postharvest quality of cut flowers. However, the current VL guarantee methods often rely on subjective evaluation, overlooking critical factors such as the susceptibility of cut flowers to postharvest disease and senescence characteristics. While AI offers promising advancements in VL prediction through prediction models based on deep learning machine techniques (Kim et al. 2024;Sun et al. 2021), the success of its application depends on understanding and resolving the varied perceptions and trust levels among different customer groups.

The survey revealed distinct demographic (sex and age), job, and behavioral differences between consumer groups. These differences may influence cut flower purchasing behaviors and perceptions of AI-guaranteed VL systems in the flower industry. General customers, particularly in Fukuoka, demonstrated a strong interest in VL guarantees, a willingness to pay higher prices, and higher trust levels in AI-based prediction systems. In contrast, flower industry-related people exhibited lower interest in VL guarantees and were less willing to pay higher prices regarding AI systems despite moderate trust in AI’s potential. This difference may be due to their reliance on traditional expertise and established practices for guaranteeing flower quality. Thus, to encourage adoption people in this group, the focus should be on demonstrating the cost-effectiveness and operational advantages of AI-guarantee VL systems as a complementary tool for ensuring flower quality. The survey data also revealed regional differences in expectations for VL guarantee durations and willingness to pay for AI guarantee systems. Trust in AI for predicting VL varied significantly between Osaka and Fukuoka customers. Customers in Fukuoka exhibited higher trust levels than in Osaka, with a notable rate of finding AI highly reliable. The regional difference suggests that Fukuoka consumers may be more technologically optimistic, which could facilitate the adoption of AI-guarantee VL systems. However, broader surveys covering diverse regions and larger populations are essential to enhance the reliability and applicability of the data.

Conclusion

This study provided valuable insights into the perceptions of AI in the flower industry. The survey also indicated the differences between general consumers and industry professionals in their openness to adopting AI technologies. General consumers remained skeptical about AI applications, particularly due to concerns about reliability, but they were willing to pay a high price for flowers with an AI-guaranteed system. While flower industry-related people were more open to its potential applications, recognizing its ability to improve flower quality and optimize operations, they were willing to pay a minimal increase in price for flowers with a VL guarantee. The survey also indicated regional differences in the perception of the AI-guarantee VL system in cut flowers. The results revealed that understanding these differences in customer perceptions is important in developing VL guarantee systems based on AI techniques for the flower industry. In addition, publicity strategies should be developed by the flower markets to enhance the reliability of VL guaranteed through AI prediction models.