Integrating AI-Enabled Platforms in Agricultural Marketing in Jordan’s Agriculture Sector to Connect Agriculture with Technology

ABSTRACT:

Introduction: This study examined how Artificial Intelligence (AI) platforms are applied in agricultural marketing and the issues faced by stakeholders and their input on marketing performance.

Methodology: The mixed-methods design was applied, where in the first stage, in-depth qualitative interviews were conducted with farmers, agribusiness managers, and marketers (n=10), and the themes that were identified include AI integration, barriers to the adoption of AI, and the effect of AI on the performance of agricultural marketing. Second, the Analytic Hierarchy Process (AHP) was utilised with five agricultural economics experts to rank the most significant challenges, noting limited technical assistance and unreliable internet connectivity as the most important hindrances. Third, a quantitative survey of 290 respondents evaluated the association between AI adoption, the utilisation of AI-powered tools, and marketing performance, and estimated these associations using regression models.

Results: Results indicated that ML-enabled market and trend forecasting (β = -0.104, p = 0.072), big-data analytics for price and demand forecasting (β = 0.155, p = 0.029), and digital connectivity and market integration (β = 0.384, p < 0.001) significantly affected agricultural marketing in Jordan.

Conclusion: The research adds to the literature by connecting qualitative findings, expert rankings, and quantitative verification, providing useful policy implications.

Keywords: AI platforms, agricultural marketing, challenges, AHP, Jordan, technology.

1. INTRODUCTION

Agriculture has been recognised as the most significant sector for a long time, to sustain international food security, rural development, and livelihoods (Kannan, 2022). In recent years, the swift advancement of digital and Artificial Intelligence (AI)-enabled technologies has now offered novel ways to bridge conventional agricultural practices with contemporary AI-enabled marketing platforms, as reflected by (Akintoye et al., 2023; and Soni et al., 2025) as well. As indicated by (Fitriyani & Nasir, 2025), to improve agricultural marketing transparency, efficiency, and competitiveness, AI-based systems, including predictive analytics, smart marketplaces, and intelligent supply-chain tools, are already being deployed in various parts of the world. In this global trend, Jordan has a new case. Despite being a relatively small economy with little natural resources, particularly water, the country’s agricultural sector is significant for food security, employment, and exports, as (Abaddi, 2025) suggests as well. According to (Qazinform, 2025), in 2024 the sector registered growth of 6.9, an upsurge of 5.8 in 2023, and exports of agricultural products also rose to JD 1.5 billion, (Jordan News Agency, 2025), an increase of 39% to 2023. The sector’s GDP contribution also rose to 7% in 2024, driven by a significant increase in trade in the last quarter of the year, as revealed by Trade. (Gov, 2024). These economic benefits indicate the sector’s potential and strength, despite farmers’ difficulties in ensuring that agricultural products are widely marketed.

Further, as reflected by (Al-Taha’at et al., 2025), the marketing issues facing the agricultural sector in Jordan are complex. For example, farmers lack effective, timely market data, including price dynamics, consumer preferences, and demand predictability. According to (Sott et al., 2020; and Bhat & Huang, 2021), these issues create a gap between demand and supply, excessive production of agricultural products that expire and go to waste, and unfulfilled consumer preferences. On the other hand, the stated marketing information gap leaves farmers susceptible to exploitation by intermediaries, who frequently capture disproportionate profit shares, and to fluctuations in demand, as highlighted by (Araújo et al., 2021; and Birner et al., 2021). Moreover, as noted by (Anbar et al., 2020; and Devlet, 2021), government initiatives, such as the Sustainable Agriculture Plan and the Economic Modernisation Vision in Jordan, emphasise modernisation, yet the implementation of digital platforms and AI-driven solutions in agricultural marketing remains limited. The challenges signify numerous research gaps. For instance, while AI is increasingly analysed in agricultural production, few studies, such as (Beithou et al., 2022; and Mohammad et al., 2025), have examined its potential role in agricultural marketing, particularly in developing economies in the Middle East, such as Jordan. Moreover, there is a lack of empirical evidence on how AI-enabled platforms work. Moreover, there is little empirical evidence on how AI-enabled platforms can redesign agricultural marketing practices to address market information asymmetries, improve demand forecasting, predict consumer preferences, and enhance competitiveness in domestic and export markets. In addition, limited research, such as (Ali & Salhab, 2025; and Fitriyani & Nasir, 2025), explored the application of AI-enabled marketing, which is critical for optimising marketing strategies and processes in Jordan’s agricultural sector. Bridging these gaps is important to advance both academic understanding and practical policy interventions in Jordan’s agricultural sector. Furthermore, previous studies did not utilise AHP analysis, which could have indicated the prioritisation of challenges in implementing AI.

The current study makes three primary contributions. First, the theoretical contribution is the expansion of pre-existing models of digital agriculture by introducing the role of an AI-based marketing platform within the marketing process and by emphasising local contextual conditions, including infrastructure and regulatory systems. Second, the empirical study contributes novel evidence on farmers’ perceptions, adoption, and marketing outcomes of AI-based products in Jordan, thereby filling a significant literature gap. Third is the policy and industry contribution to making recommendations to policymakers, industry regulators, and agribusiness stakeholders for practical implementation to ensure AI adoption, whether through infrastructure investments or capacity-building programmes.

2. LITERATURE REVIEW

2.1. Theoretical Framework

The research is based on the Technology Acceptance Model (TAM) and the Resource-Based View (RBV), which, in association, make a multi-level theoretical framework for analysing AI-based platform integration in agricultural marketing. The existing literature (Dissanayake et al., 2022; Mohr & Kuhl, 2021) on digital agriculture mostly focuses on the functionality of technology and performance outcomes, whereas few studies directly combine the behavioural adoption theory with the strategic competitiveness theory in developing-country agricultural settings. Therefore, by situating this study within the relevant literature, this study uses TAM and RBV jointly to focus on individual- and sector-level competitive advantage. As suggested by (Ma & Zhang, 2022), TAM describes technology adoption in terms of perceived usefulness and perceived ease of use. Farmers and other stakeholders use AI-enabled platforms in the agricultural marketing sphere when they see concrete advantages, such as better price forecasting, market access, and demand forecasting, and when the system is user-friendly and can be integrated with stakeholders’ digital functionality, as indicated by (Fitriyani & Nasir, 2025) as well. In scenarios such as Jordan, where digital literacy and infrastructural differences remain evident, ease of use is a decisive factor in behavioural intention.

In addition to TAM, RBV assumes that sustainable competitive advantage is attained by firms with valuable, rare, inimitable, and non-substitutable resource (Permono & Kurniati, 2024; Petcu et al., 2024). The vision for AI-based marketing sites is to develop strategic digital tools that enhance market intelligence, operational efficiency, and value chain coordination. In contrast to previous literature, such as (Beithou et al., 2022), which assumes that technology is an operational instrument, the current study frames AI platforms as strategic resources that determine the sector’s competitiveness. The framework connects adoption behaviour to strategic impact, filling the gap between TAM and RBV, solidifying the current literature, and situating AI-oriented agricultural marketing in the context of Jordan.

2.2. Machine Learning enabled Market and Trends Forecast Tools’ Effect on Agricultural Marketing Performance

Forecasting tools that utilise Machine Learning (ML) are becoming widely acknowledged at increasing agricultural marketing because they convert intricate datasets into market insights that can be acted upon. According to (Adegbola, 2025; Alice & Ebuka, 2024) ML allows to identify growth trends, possible supply chain inefficiencies, and seasonal changes and makes more proactive and data-driven choices. Nevertheless, these studies focus on predictive accuracy and strategic responsiveness, but are mostly conceptual, and do not take into account contextual limitations to implementation. On the contrary, the article by (Dissanayake et al., 2022) presents empirical evidence through the AgriAutoMark platform and proves that conversion rates and repeat purchases are significantly improved. Nonetheless, their conclusions are founded on their small-scale application, which makes their results questionable concerning applicability and scaling to heterogeneous agricultural systems like Jordan.

This difference between theoretically positive results and context-based empirical constraints points to the significant discrepancy in the literature. The current study is prioritized on the performance of technological aspects at the expense of behavioural and structural adoption variables. Perceived usefulness and ease of use are key constructs in the eyes of Technology Acceptance Model (TAM); nevertheless, these constructs are seldom empirically investigated among farmers in developing economies. Furthermore, the relationship between the technological capability and user preparedness is not well evidenced. As a result, the necessity of conducting context-specific empirical research, which combines the performance result with adoption behaviour to assess the feasibility of using ML-enabled marketing tools in practice, is evident. These arguments lead to the formulation of the following hypothesis;

H1: There is a significant positive effect of Machine Learning-enabled market and trends forecast tools on Agricultural Marketing Performance in Jordan’s agriculture sector.

2.3. Big Data Analytics (BDA) for Price and Demand Forecasting and Agricultural Marketing Performance

Big Data Analytics (BDA) has been emerging as a disruptive instrument in agriculture marketing especially in improving price and demand forecasting by incorporating extensive market data. (Giannakopoulos et al., 2024) reveal that agroeconomic indexes (AEIs) contain several indexes branded traffic and search visibility that are strongly related to digital marketing performance indicators through advanced analysis tools such as regression models, fuzzy cognitive mapping, and artificial neural networks. Likewise, (Ma & Zhang, 2022) have found significant beneficial changes in turnover and profitability of BDA-powered smart marketing models in the fruit market in China. These results so far indicate that BDA is capable of aligning marketing strategies with macroeconomic indicators and boosting the performance of firms.

Although (Giannakopoulos et al., 2024) use a small sample of big business, their results might not be indicative of the realities of small and medium-sized agricultural businesses, which predominate in the economies of developing countries. (Ma & Zhang, 2022) on the other hand, provide optimistic results derived out of estimated models instead of strong empirical validation, casting doubts on reliability and generalisability. This deviation refers to a disjunction of the analytical and the practically applicable.

According to the Resource-Based View (RBV), BDA can be conceptualized as a strategic competence that have the ability to create competitive advantage by way of better forecasting and decision-making. Though, the success of such capabilities is conditional upon the absorptive capacity of firms that has not been investigated in empirical studies. Moreover, the current literature does not pay much attention to the interpretation, trusting, and use of data-driven insights by farmers. Thus, context-specific studies to unify analytical performance and organisational preparedness with user-level adoption are required to determine the actual effect of BDA within the agricultural marketing systems. Based on these literature arguments, the second hypothesis, H2 of the study, is postulated.

H2: The application of Big Data Analytics (BDA) for price and demand forecasting significantly improves the Agricultural Marketing Performance of the agricultural sector in Jordan

2.4. Digital Connectivity and Market Integration Effect on Agricultural Marketing Performance

The use of Advanced Networking Technologies (ANT) is also seen as improving agricultural marketing by increasing digital connectivity and real-time information sharing, as well as better integrating farmers and the market. (Soni et al., 2025) show that AI-based mobile applications can enhance the involvement and productivity of smallholder farmers to a significant degree as they can gain access to market information and financial services. These results indicate that ANT is able to minimize information asymmetry and enhance market participation at an individual level. On a larger level, (Eshbayev et al., 2024) theorize the concept of agricultural markets as networked systems by stating that digital integration would improve supply-chain coordination and regional competitiveness by aligning stakeholders.

Notwithstanding these positive findings, there are significant inconsistencies in the literature. Though (Soni et al., 2025) provide empirical evidence related to better performance at the farm level, they base their study on a small-scale pilot version, which restricts generalisability. On the other hand, (Eshbayev et al., 2024) take a rather system-level approach, but rely on conceptual modelling extensively, without empirically proving practical applicability. This difference is indicative of a gap between empirical results on a micro level and theoretical constructs on a macro level.

Perceived usefulness in terms of better access to the market and efficiency and perceived ease of use in terms of accessible and easy to use platforms are among the factors behind ANT adoption (Technology Acceptance Model (TAM)). Nevertheless, previous researchers such as (Eshbayev et al., 2024; Soni et al., 2025) pay inadequate attention to the impact of barriers such as digital illiteracy, the infrastructural constraints, and the uneven technological diffusion on adoption. This makes it necessary to conduct integrated empirical studies that would fill the gap in research that focus across individual user behaviour on one hand and systemic market transformation on the other to further comprehend the role of ANT in agricultural marketing. Grounded on these literature arguments, the following hypothesis H3 of the study is formulated;

H3: Digital connectivity and market integration significantly and positively influence Agricultural Marketing Performance in Jordan’s agriculture sector

2.5. Challenges in Adopting AI-Enabled Marketing Platforms in Agriculture

A growing body of research points out that the use of artificial intelligence (AI) in agriculture is limited by a number of technological challenges. The first of them is the lack of digital skills, particularly regarding the smallholder farmers of the developing world who struggle to use refined platforms and applications, which is also supported by (Tiwari et al., 2021) in their findings. As reflected by (Dhillon & Moncur, 2023), the training programmes exist, and they are usually provided in a form that is not suitable in the local context, which strengthens inequality in the adoption of technology. On the other hand, (Ma & Zhang, 2022) are of the opinion that the second barrier is the limited availability of smartphones and AI tools, as the cost and maintenance of infrastructure remain limiting the use of smartphones and AI tools in rural settings. Empirical evidence provided by (Mishra et al., 2023) indicates in this view that rural users incur a relatively higher cost of connecting than their urban peers and can afford to use digital resources in the long run. It is also crucial that financing and subsidies are not provided, as farmers tend to lack sufficient access to financing or government subsidies to meet the initial expense of acquiring smartphones, subscriptions, and AI programmes mentioned in the discussion of (Ali et al., 2024). In the cases where the tools exist, the lack of certainty in ROI prevents their usage, as the evidence of the increase in profitability has not yet been provided (Giannakopoulos et al., 2024). Together, these obstacles make it a cycle of exclusion in the sense that the farmers are unwilling to invest in the technologies that seem to be uncertain, expensive, and not supported by finance sources.

The policy and institutional obstacles are also a major impediment to AI adoption in agriculture, as per (Permono & Kurniati, 2024). Of the greatest importance is the fact that the regulatory environment is poor, with no clearly defined policies on data ownership, digital compliance, and accountability, which restricts the trust in digital platforms and prevents their use on a large scale. In this perspective, (Fitriyani & Nasir, 2025) also believe that in the absence of these specially designed aid programmes, higher, more affluent stakeholders are favoured in adoption. All these institutional shortcomings make the integration of AI more gradual, which continues to perpetuate structural inequalities in agricultural markets.

According to (Alice & Ebuka, 2024), the extent to which farmers embrace AI technologies is more or less based on socio-cultural factors. Among the significant barriers, one should mention the lack of confidence in digital platforms, as farmers are inclined to trust more traditional intermediaries and traditional relations than new digital systems. The issues of credibility, disinformation, and transparency further limit acceptance. Another situation that keeps recurring is resistance to change, where the generation and mindset gap is an obstacle to the diffusion of innovation in rural areas, as suggested by (Ma & Zhang, 2022). Finally, there is a barrier to access due to language, with most of the sites being written in international or national languages and thus not accessible to the rural areas, as the majority of the population speaks local languages, as recommended by (Ahmad et al., 2024; and Alice & Ebuka, 2024) as well. These socio-cultural issues highlight the importance of trust-building, local, and inclusion-based designs and communications to digital agriculture.

Infrastructure bottlenecks remain the key obstacles to successful AI implementation in agriculture. In this context, (Tiwari et al., 2021) argued that poor internet connectivity continues to limit farmers’ access to digital channels, especially in rural regions where network coverage is low and unreliable. There are rare chances to have real-time market data, weather forecasts, and digital advisory services without frequent connectivity. Limited technical assistance is another primary limitation, as the majority of farmers lack access to on-the-ground troubleshooting or training on how to use AI-enabled instruments. Such a lack of continuous support impedes long-term use of digital technologies. Moreover, unstable power sources create barriers to adoption, with frequent blackouts preventing device charging and making it more difficult to ensure the proper functionality of platforms in developing countries, as (Alice & Ebuka, 2024) also revealed. These structural obstacles create the digital divide, and the focus is on investments in connectivity, energy accessibility, and support systems to enhance transformations in agriculture through AI.

2.6. Research Gaps

Despite the growing scholarly interest in the use of AI in agricultural marketing, where much remains to be done, especially in the Jordanian setting, at first most existing research is theoretical or simulation-based. (Adegbola, 2025; and Khalili, 2024) do point to predictive intelligence and agent-based simulation, though little empirical data depict the actual performance of these technologies in low-resource settings like Jordan, where small-scale farmers are the rule and infrastructural challenges persist. Second, research findings developed cannot be generalised. AgriAutoMark was forwarded by (Costa et al., 2023) on only 12 agri-businesses; in contrast, (Giannakopoulos et al., 2024) tested five large firms, abandoning the fact that in Jordan, over 90% of farms are small-scale and that they are experiencing problems with the lack of marketing channels, low digital competencies, and financial capacities, as unveiled in the findings of (Anbar et al., 2020). Third, despite the fact that the technological potential of ML allowing market and trends forecast, BDA price and demand forecasting, and digital connectivity and market integration are being recognised, there is a scarcity of research on the barriers, such as low digital literacy, lack of access to smartphones and AI tools, the privacy issue of data, and mistrust in AI recommendations. These challenges are particularly important in Jordan, where adoption relies significantly on affordability, usability, and trust (Araújo et al., 2021; Birner et al., 2021). Lastly, while research examines these challenges, there is no systematic prioritisation of the most important barriers to address. The lack of systematic evaluation procedures, e.g., the Analytic Hierarchy Process (AHP), means policymakers and stakeholders do not receive evidence-based advice on which technical, economic, infrastructural, institutional, or socio-cultural problems should be prioritised first to optimise adoption and effectiveness.

3. METHODOLOGY

The study is based on the mixed method research design. It involves three phases in which the research purpose is addressed. Firstly, the study conducts qualitative interviews with the farmers, agri-business managers, and agri-marketers. It involves a sample of 10 respondents in which 4 were agri-business managers, 4 were agri-marketers and 2 were farmers. The research has used purposive sampling to identify the respondents who have first-hand experience in agricultural marketing and the adoption of AI platforms. The reason why ten participants were selected was that they are some of the most important stakeholders in the agricultural value chain; they are agri-business managers, Agri-marketers, and farmers who will provide distinct information about the adoption of AI and its marketing performance. In qualitative studies, emphasis is laid on rich and in-depth information as opposed to huge sample sizes. The chosen respondents have a professional background that is pertinent to the research aims, which allows studying the issues of AI integration and its effects in a meaningful way. Further, ten participants are sufficient in qualitative research since at this stage the researcher can achieve thematic saturation in that no new thematic information can be gained through further interviews. The sample size is consistent with the qualitative study of (Che et al., 2020; and Kendall et al., 2022). (Braun & Clarke, 2021) have also explained that in qualitative research, smaller purposive samples are adequate when participants have specialised knowledge of the phenomenon being studied. These studies have suggested that thematic saturation can be attained with 6-12 interviews in focused qualitative design. Furthermore, (Creswell & Poth, 2016) illustrated that qualitative research focuses on exploring experiences, meanings, and perspectives in depth, but its findings are context-specific and not broadly generalisable. The small sample is also explained by the narrow scope of the study and the homogeneity of the sample, with all participants having a relevant background knowledge in agricultural marketing and AI adoption. This specificity minimises the size of the necessary samples as significant patterns can be revealed faster. Moreover, semi-structured, in-depth interviews yield rich, detailed information, which can be analysed in depth using thematic methods, that is, more in keeping with the qualitative research traditions.

Additional rigor of the methodology is reflected in the elaborate participant selection and validation. As an example, agri-business managers were chosen according to their role in strategic marketing decisions, whereas agri-marketers were directly familiar with digital tools and AI platforms, and farmers provided their practical experience of usability and adoption barriers. Such triangulation of views increases the validity of results.

These respondents were asked questions regarding the integration of AI platforms, challenges faced in integrating AI platforms and impact of AI on agricultural marketing performance. The interview questions are attached in Appendix A. These themes were analysed using thematic analysis using the approach of (Braun & Clarke, 2023). The demographic profile of respondents is presented in Table 1.

Table 1. Demographic profile.

In second phase, AHP analysis was conducted to highlight most prioritised challenges. The AHP process involves weighing the different criteria developed based on the interviews. The questionnaire was formed in the form of pairwise comparison and distributed to the five agricultural economics experts involved in designing market models, pricing, demand forecasts, value chain analysis, and business cases for farmers. Pairwise comparisons are applied to consider the decision factors by considering the relationship between criteria and sub-criteria. The first stage to define the priorities of elements in the decision problem is making the pairwise comparison on the provided criteria to obtain a useful scale while comparing two components. Pairwise comparison is provided in matrix form. The AHP questionnaire (Appendix B) is developed on a 1 to 9 scale to assess the degree of importance of elements compared to other elements. The scoring is considered as in Table 2.

Table 2. AHP questionnaire sample.

Note: Intensities of 2,4,6, and 8 are used for expressing the immediate values

Scale values from 1 to 9 are assigned to compare the pairs at each level of the hierarchy. The validity will be evaluated using the consistency test. If the consistency test results are smaller than 0.1, then the results are obtained to be consistent. The formula for measuring the consistency ratio is as follows.

Consistency Index (CI) = λ_max – n / n – 1

The consistency ratio is measured as:

Consistency Ratio = CI/RI

The third phase involved identifying the impact of AI on the agricultural marketing performance using the quantitative analysis. The variables identified from literature, validated form qualitative analysis and the challenge most prioritised in the AHP has been used as the independent variable to evaluate their impact on agricultural marketing performance. The questionnaire survey (Appendix C) was distributed to farmers, agri-business managers, and agri-marketers. The survey was distributed to 600 respondents out of which 300 responses were received completed. After cleaning for the missing data and outliers, 290 responses were left. The regression analysis using SPSS was used to analyse the impact of AI on the agricultural marketing.

Several processes were implemented in order to deal with reliability and validity. Qualitative validity was enhanced by having theme verification and consistent cross-checking of coded interview transcripts to have consistency in interpretation. Cronbach alpha was used to test quantitative reliability of the survey constructs to ensure an internal consistency among measurement items. Moreover, construct validity was also achieved through the adaptation of survey tools and measurement items based on the well-established and peer-reviewed literature. In order to reduce possible bias of the researcher, the responses on interviews were coded in a systematic manner based on a pre-determined thematic analysis process. The coding and interpretations were checked and counterchecked with the original transcripts several times, and the results were presented in such a way that they represented the perceptions of the participants correctly and provided the transparency of the analysis. Data triangulation across qualitative, AHP and quantitative stages also enhanced methodological rigor and increased credibility and consistency of findings. Transparency was ensured by means of clear audit trails, standardized protocols and iterative validation. Also, measures like reducing interviewer bias, anonymity of respondents and use of uniform measurement procedures were among the factors which made the results robust and credible.

4. RESULTS

4.1. Interview Analysis

Table 3 summarises the results obtained from the interview analysis in a tabular form.

Table 3. Thematic table.

4.1.1. Integration of AI Platforms

The interviews revealed the adoption of AI platforms in farm marketing is gradually taking shape but not equally well with stakeholders. Farmers, agri-business firms, and agri-marketers all provided opinions that reflect both benefits and constraints of the technologies. Among the prevailing themes was the use of AI platforms to access markets better and usher in pricing transparency. Farmers explained that having the ability to get revised prices in real-time makes them confident in their interactions with traders because they are not as reliant on middlemen. One farmer explained that;

“Now I can check the market price before selling,” one farmer explained. It makes me more confident to negotiate better.”

Agri-marketers too concurred with this view, noting that AI-based price forecasts are becoming increasingly useful in bridging sellers and buyers, though they noted that the forecasts are not always aligned with ground-level market forces.

The second trend that came through was the application of AI to facilitate decision-support as well as boost productivity. Agri-businesses noted how predictive analytics allow for the forecasting of demand, optimising supply chains, and direction of input use. One agri-business executive commented,

“AI helps us predict what crops will be needed next season so we can guide farmers on that.”

Farmers similarly reported being guided on planting calendars and on crop care, although a few questioned the validity of such guidance when weather is unpredictable. Integration normally takes the form of a hybrid system which is a combination of digital and traditional practises. Most farmers said that they used AI platforms as the initial source of information but finalised the judgement using conventional networks. One explained,

“I check the prices on the app; however, I always check it with the local trader.”

Another observation by Agri-marketers was that adoption was often gradual with farmers beginning with basic features like price updates and progressing to the more sophisticated features like predictive analysis or supply chain optimisation.

Lastly, institutional and ecosystem support was also found to be a significant factor that assisted in the promotion of integration. Agri-marketers and agri-businesses emphasised that pilot projects, trainings and subsidies are essential towards persuading farmers to use AI tools. One agri-marketer states that,

“Farmers will not even attempt these platforms without training sessions and subsidies.”

It is concluded that although AI-based platforms are already being implemented into agricultural marketing to make it more transparent, decision-making, and connected, they will only be successful in the long term when they are able to generate trust, be contextualised, and supported by institutions.

4.1.2. Challenges Faced by Farmers

Themes in this study were formulated through coding and categorising responses of the farmers, agri-businesses and agri-marketers to interviews. All the general problems were put together into five umbrella themes, which included technological, economic, institutional/policy, socio-cultural, and infrastructural issues. The themes depict certain sub-challenges that are typically featured by participants, and suggestive accounts of hypothetical responses.

The biggest technological obstacle was the farmers being not digitally literate. A farmer stated,

“I struggle to know how to use these applications. It is too fast, not experiential even to us when we are trained.”

It is also clear that the training mechanisms exist but are not feasibly adjusted to the learning tendencies of rural farmers.

Another problem was restricted access to smartphone tools or AI, especially with smallholder farmers. One of the respondents described it as follows:

“I have a basic phone; it is not cheap to purchase a smart phone and use these platforms.”

This explains the direct hindrance of participation in AI-mediated market by the digital divide. Issues of data protection and security were also raised with a lot of concern. One of the agri-business managers commented,

“We do not have faith to post sensitive information because we are not aware of the location of where the information is kept, and how it can be accessed by other people.”

The issue of misuse discourages the stakeholders to utilise digital platforms in the best way possible. Lastly, the respondents doubted the validity of AI recommendations. A marketer expressed,

“Sometimes AI will give unrealistic prices that do not consider the local seasonality and demand.”

This kind of scepticism implies that there is a necessity to align AI output to situational realities to achieve credibility and confidence.

Moreover, economic problems were also identified as the major one that affected the implementation of AI in agriculture during the interviews. The most widespread reason was the prohibitively high cost of using the internet and platform. A farmer commented,

Even on the occasions when I would like to use the app, the price of the internet swallows up my already limited income.

This implies that one of the underlying barriers to the adoption of technology is affordability. Moreover, absence of subsidies or funding was also an issue particularly to agri-business. A farmer business operator described,

“We are not able to pay subscription fees unless there is an option of credit or government subsidy.”

This implies that adoption remains heavily reliant on third party funding schemes. It was also not adopted due to lack of certainty on the return on investment (ROI) as one of the marketers explained;

“It is selling digital tools, yet farmers are not seeing evidence that these sites are earning them more money.”

In the absence of these economic rewards on the surface, most would not be willing to commit resources behind pick-up. It points out that the absence of ROI is one of the major challenges.

Moreover, stakeholders singled out the weak regulatory environment with institutional and policy issues as the important one. A manager in an agri-business said,

“Digital agriculture has no visible laws or policies that govern it, and this keeps us in the dark about compliance.”

Confidence is lost through lack of governance. Lack of government patronage was demoralising as well. A farmer had to say,

“We are reading about digital farming, but the government has not provided us with any physical assistance to implement these tools.”

It is an indicator of a laggard adoption policy. Poor coordination between institutions was also observed. An agri-marketer stated,

“People are operating in silos. No collaborative intention to scale digital agriculture.”

This division decreases effectiveness in marketing and backing platforms. Therefore, the lack of coordination is one of the issues of significance in this respect.

Moreover, socio-cultural issues are also the major issues. The general problem was low confidence in online platforms. A lack of trust was often complained of by farmers, one of them explaining,

“I believe in the middleman I have known and not a phone application that determines the price.”

This is a cultural resistance in terms of familiarity and personal relations. Another barrier was resistance to change particularly among the older farmers. The following insight was given by one of the agri-business stakeholders:

“The reason is that many farmers are still adhering to the traditional practises and they perceive digital tools as an unwarranted risk.”

This brings out differences in generation acceptance. Lastly, it was not easily accessible due to language barriers. One farmer described,

“The site is not in our local language hence we cannot comprehend the words.”

This isolates huge masses of rural consumers, citing localisation. Therefore, it found the absence of trust, change resistance and language barriers as the major socio-cultural challenges.

Lastly, the infrastructural issues have been identified based on the responses of the interview. Lack of internet connectivity was reinforced severally. As relates one farmer,

“The internet in my village is too slow, the application is even not loading.”

The inadequate connectivity affects the platform reliability. There was also limited support. As pertains to an agri-business owner,

“When we are in trouble, we cannot get the immediate assistance, thus slackening our processes.”

This is a reference to the difference between adoption and continued use. Lastly, untrustworthy power supply was also a limiting factor. A farmer described,

“I can even not charge my phone sometimes because of the constant power cuts.”

This explains why simple infrastructural weaknesses slow down digital transformation in agriculture. It reveals that affordability, competency, and trust are important to the farmers, financing, data security, and policy vacuum are important to agri-businesses, and ROI, contextual reliability, and communication gap are important to agri-marketers. The problem is multidimensional and interrelated and needs to be addressed on the training, finance, infrastructure, and policy level to facilitate the successful use of AI in agricultural marketing. The analysis below shows the AHP of these challenges in a hierarchy.

4.1.3. Impact of AI on Agricultural Marketing Performance

The interviews with farmers, agri-business, and agri-marketers analysis highlighted how AI is changing the performance of agricultural marketing, but also revealed some long-term challenges and emerging themes. Three areas of influence were noted in the interviews, including predictive intelligence enabled by machine learning, competitive enabled by big data analytics, and connected enabled by advanced networking technologies.

Farmers emphasised how AI-based forecasting tools improved decision-making by predicting the changes in the demand and prices. According to one farmer,

“I would sell my produce without having the knowledge of future price fluctuation. I can now plan storage or sales in a better way to eliminate losses.”

This is echoed in the argument that machine learning platforms can be used to increase perceived usefulness by being more accurate in prediction. Nonetheless, the problem of digital literacy arose. Another farmer said,

“The application is helpful, yet I need to be trained. I sometimes cannot use all the features.”

This implies the ease-of-use factor of TAM and the importance of simple platform design. Agri-business managers highlighted the importance of big data analytics in the competitiveness of marketing. One of the respondents expounded,

“Our company can keep track of consumer trends and market indicators using data dashboards. It renders us more aggressive and faster than the competition.”

There were others who were concerned with absorptive capacity. One manager acknowledged,

“We have the data, and our employees lack the experience with the interpretation of advanced models.”

This gap shows that big data is an asset with high potential but organisations must invest in human capital so that they can maximise it. Agri-marketers placed a lot of emphasis on the recent networking technologies and particularly the mobile applications. One of the interviewees said,

“I will be able to connect farmers and the right buyers in real-time with updated weather and prices.”

But there were still questions of cost and inadequate infrastructure. One of the marketers states that,

“The technology is most effectively used in areas that have good internet connection, but those in the rural areas are still lagging.”

Another fourth factor is found that is trust in AI systems. Other respondents highlighted that adoption required transparency and trustworthiness.

One farmer summed up,

“I will apply AI tools when I am sure that the information presented is truthful and objective.”

This implies that besides utility and ease of operation, trust-building is also necessary in integrating AI in agricultural marketing. The fourth factor that emerged as a necessary condition of AI systems adoption was trust. The agri-businesses and farmers stressed that unless transparent and reliable, easy-to-use and accurate tools would be resisted. Trust is a guarantee of the reliability of AI-generated information and is a critical supplement to perceived usefulness and ease of use in agricultural marketing.

Overall, the discussion shows that AI has a tremendous positive impact on agricultural marketing in the aspects of predictive intelligence, data-based competitiveness, and improved connectivity. There are still barriers such as digital literacy, infrastructure and absorptive capacity. Interestingly, the confidence in AI systems emerged as one of the most important ones, as transparency and reliability are the key to sustainable adoption.

4.2. AHP Analysis

4.2.1. Identification of Challenges and Hierarchy Development

The challenge identified from the interviews has been characterised into the hierarchy as given in Fig. (1). This provides the key challenges to adopt AI-enabled agricultural marketing platforms in Jordan which are further evaluated using the AHP analysis.