Advancing Sustainable Agriculture Through Digital Technology: The Role of the ‘Agricultural Guide’ App in Improving Olive Farming Practices in Saudi Arabia

Abstract

The Agricultural Guide application is a crucial component of the digital extension system in Saudi Arabia, providing modern and evidence-based information on sustainable agricultural practices to the farming community. The adoption of digital extension tools has been widely recognized as a key driver in enhancing crop productivity. This study aimed to assess the impact of the Agricultural Guide application on the adoption of sustainable olive farming practices in the Kingdom of Saudi Arabia. The impact was evaluated by analyzing the farming practices of the users and non-users of the application, identifying key determinants of application usage through machine learning techniques. The study also analyzed barriers to its adoption. A structured questionnaire was employed to collect data from 229 olive farmers in the Al-Jouf region. The findings reveal that the majority of respondents were non-users of the application. Significant differences were observed between users and non-users regarding the adoption of sustainable agricultural practices, including irrigation management, soil improvement, pest control, and harvesting techniques. Furthermore, farmers’ productivity, income levels, and digital information sources were significantly influenced by their usage of the application. A random forest analysis, with a predictive accuracy of 94.12%, identified key determinants of the application usage, including digital information sources, soil improvement practices, irrigation management, and education level. The study highlights the need for targeted educational programs under the supervision of the Agricultural Extension Department to enhance farmers’ awareness and knowledge of the Agricultural Guide application. Expanding its adoption within the farming community has the potential to significantly promote sustainable agricultural practices and improve overall agricultural productivity in Saudi Arabia.

1. Introduction

Over two billion smallholder farmers worldwide face persistent challenges related to poverty, largely driven by low agricultural productivity, climate change, environmental risks, and deforestation [1]. To address these issues, digital technologies have been developed to disseminate sustainable agricultural practices aimed at enhancing crop production and mitigating the adverse effects of climate change. Among these technological advancements, smartphones have emerged as one of the most rapidly evolving digital tools, with an estimated 4.3 billion users globally [2]. The proliferation of mobile applications has revolutionized information and communication systems, playing a pivotal role in agricultural development.

Empirical evidence indicates that digital technologies have significantly improved agricultural yields by facilitating the delivery of agricultural extension services [3]. These technologies have enhanced connectivity and expanded farmers’ access to vital agricultural information, particularly in rural areas [4]. Mobile technologies provide real-time access to critical farming information, including market trends and contemporary agricultural practices [5,6]. Furthermore, smartphone applications offer advanced functionalities such as pest detection and crop disease diagnosis, which contribute to improved farm management and productivity [7,8]. These innovations have not only facilitated the efficient use of resources but have also raised farmers’ awareness of and willingness to adopt sustainable agricultural practices [9,10]. The integration of such technologies into farming systems underscores their potential to enhance agricultural resilience and long-term sustainability.

In the 1980s, agricultural applications, such as precision agriculture technologies, sensors, and GPS devices, were used for field data collection and crop monitoring [11,12]. In the twenty-first century, the internet and mobile phone applications have supported agricultural operations, such as weather forecasting and crop management. For instance, artificial intelligence and internet-based applications are utilized to diagnose plant diseases and analyze data with high accuracy to address agronomic challenges [13,14]. Currently, integrated applications, such as drones and remote sensors, are being used to manage irrigation and harvesting operations.

Historically, the agricultural extension system in Saudi Arabia has relied on extension agents and utilized a one-way communication model to disseminate agricultural information to the farming community. This traditional approach, coupled with a limited awareness of digital extension systems, restricted farmers’ access to information on innovative and sustainable agricultural practices [15,16]. However, after sustained efforts, Saudi Arabia transitioned to a digital extension system, which has been widely adopted by extension agents [17]. The adoption of the “Agricultural Guide” application has played a crucial role in addressing agricultural challenges by providing farmers with timely and specialized information on sustainable farming practices. The Ministry of Environment, Water, and Agriculture (MEWA) in Saudi Arabia introduced the “Agricultural Guide” application to offer large-scale agricultural services across the country. In 2022, the application facilitated over 2.92 million agricultural consultations across 64 different fields. Additionally, approximately 233,000 farmers, including olive growers, registered on the platform. The application also facilitated 388 agricultural guidance sessions and conducted 965 soil and water analyses. Furthermore, 84 agricultural experts from the Arab Organization for Agricultural Development and the Food and Agriculture Organization of the United Nations (FAO) participated in interactive activities aimed at enhancing agricultural knowledge and practices [18]. This digital transformation in agricultural extension services has significantly improved the accessibility and dissemination of critical agricultural knowledge, fostering the adoption of sustainable and innovative farming techniques in Saudi Arabia

Saudi Arabia approximately accounts for 2% of the total worldwide olive output [19]. Olive cultivation plays a vital role in the national economy, with olive oil ranking as the fifth-largest foreign currency-generating export, constituting 5.5% of the Kingdom’s total exports [20]. Notably, olive oil produced in the Al-Jouf region is distinguished by its unique natural and nutritional properties [21]. With evolving dietary preferences, the demand for olive oil has increased by 25% in recent years [22]. Currently, domestic production satisfies 85.2% of the national consumption [22]. Despite ongoing efforts to enhance olive production, farmers in Saudi Arabia encounter multiple challenges. A key issue is the lack of access to advisory services and training on sustainable agricultural practices [21]. To address these challenges, the Ministry of Environment, Water, and Agriculture (MEWA) has developed an electronic extension system, the “Agricultural Guide”, to provide farmers with essential agricultural information and guidance. Initially launched as a trial version in 2019, the improved version, incorporating advanced smart services, was officially introduced in 2021 to promote sustainable olive cultivation practices [23].

Currently, approximately 206,000 farmers across Saudi Arabia utilize the Agricultural Guide application, representing 31.16% of the total farming population [23,24]. The application provides a range of services, including workshops, training programs, advisory support, seminars, and multimedia educational resources [25]. In line with the Saudi government’s commitment to digital transformation, efforts have been made to promote the adoption of this application [26]. However, despite these initiatives, no scientific studies have yet examined its effectiveness. This study aims to assess the impact of the Agricultural Guide application on the adoption of sustainable olive farming practices in Saudi Arabia. Moreover, it seeks to identify the key factors influencing the adoption of the application and the most significant barriers hindering its widespread use. The research is grounded in the Unified Theory of Acceptance and Use of Technology 2 (UTAUT2), which serves as a theoretical framework for analyzing factors influencing the adoption of the application. UTAUT2 is a well-established model for predicting the adoption of information and communication technologies (ICTs), incorporating factors such as hedonic motivation, price value, habit, performance expectancy, effort expectancy, social influence, and facilitating conditions. These factors collectively offer a comprehensive understanding of the psychological and social determinants affecting ICT adoption [27]. The application of this theory provides a robust foundation for investigating the motivations and barriers associated with the use of the Agricultural Guide, offering valuable insights into the conditions that facilitate or impede the adoption of digital agricultural solutions [28]. The findings will contribute to bridging the knowledge gap regarding the role of digital technologies in agricultural extension services within the study area.

2. Materials and Methods

2.1. Description of the Study Area

This study was conducted in the Al-Jouf region, located in the northwestern part of the Kingdom of Saudi Arabia. Al-Jouf is one of the most agriculturally significant regions in the country, covering an area of approximately 100,000 square kilometers and supporting a population of around 475,000 inhabitants [28]. Administratively, the region is divided into three governorates: Al-Qurayyat, Dumat Al-Jandal, and Tabarjal [26], with Sakaka serving as the regional capital. The climate of Al-Jouf is classified as a hot desert climate, characterized by high temperatures during the summer and dry, cold conditions in the winter [29]. Annual precipitation levels range between 50 and 60 mm [30]. The predominant soil texture in the study area is sandy loam [31], which supports the cultivation of various crops. Notably, olive and citrus production are the most widespread agricultural activities in the region (Figure 1). Al-Jouf is particularly recognized as the primary olive-producing area in Saudi Arabia, accounting for approximately 67% of the country’s total olive production [32].

2.2. Study Population

The olive farmers in the Al-Jouf region were considered as the study population. The total population of 6554 farmers was identified through the Agricultural Census Statistics [32]. A simple random sampling method was employed to ensure an equal representation of farmers across the defined population, including users and non-users of the application. The required sample size of 364 farmers was determined based on Krejcie and Morgan [33]. The following equation was used to calculate the sample size:

$$S={\displaystyle \frac{x^{2}NP(1-P)}{D^2\left(N-1\right)+x^{2}P(1-P)}}$$

(1)

where S = Required sample size, N = Population size, x² = Chi-square value at 1 degree of freedom and a significance level of 0.05 = 3.841, P = Proportion of the phenomenon’s probability = 0.5, D = Margin of error = 0.05. Upon applying the formula, a sample size of 364 farmers was obtained.

2.3. Data Collection

Data were collected using a structured, paper-based questionnaire administered in the participants’ native language. Prior to participation, all respondents provided written informed consent. Among the 364 farmers surveyed, 229 completed the questionnaire, yielding a response rate of 63%. The data collection process spanned four months, from March to July 2024. Ethical approval for the study was granted by the Research Ethics Committee of the Deanship of Graduate Studies at King Saud University (Approval Number: KSU-HE-23-287). This approval was obtained to ensure compliance with ethical standards, particularly concerning data confidentiality and the voluntary participation of respondents.

2.4. Survey Instrument

The study questionnaire was validated for construct reliability and content accuracy by the experts in the fields of agricultural extension and plant production at the College of Food and Agricultural Sciences, King Saud University. The questionnaire comprised multiple sections designed to collect data on various aspects, including the socio-economic characteristics of farmers, the adoption of sustainable olive farming practices, and barriers to utilizing the Agricultural Guide application. The socio-economic characteristics assessed included age, level of education, farming experience, landholding size, number of trees on the farm, olive production, income levels, dependence on agriculture as the primary income source, employment status, cultivation techniques, irrigation and harvesting methods, marketing strategies, engagement in agricultural activities, daily duration of mobile and internet usage, information sources, and the extent of usage of the Agricultural Guide application.

The adoption of sustainable olive farming practices facilitated by the Agricultural Guide application was evaluated using 23 items on a five-point Likert scale, ranging from 1 (never) to 5 (always). Prior to assessing adoption levels, all participants were asked to evaluate various sustainable farming practices. For individuals who utilized the Agricultural Guide application, an additional column was incorporated to indicate whether each practice was influenced by the application (marked as “Yes” or “No”). Only practices marked as “Yes” were included in the impact analysis, while those marked as “No” were considered in a broader context unrelated to the application’s influence. Furthermore, barriers to utilizing the Agricultural Guide application were assessed using 16 items on a five-point Likert scale, ranging from 1 (strongly disagree) to 5 (strongly agree).

2.5. Data Analysis

The data were analyzed using R software (v4.3.1). Descriptive statistical analyses, including frequency distributions, percentages, means, and standard deviations, were conducted to summarize the data. Inferential statistical analyses comprised Principal Component Analysis (PCA) and t-tests. PCA was employed to identify the key questions within each scale construct that accounted for most of the variance in the respective aspects. Two primary constructs were assessed using a 5-point Likert scale: (1) the adoption of sustainable agricultural practices facilitated by the Agricultural Guide application and (2) barriers to its utilization. The Kaiser–Meyer–Olkin (KMO) Measure of Sampling Adequacy was applied to assess the suitability of the survey response data for factor analysis. t-tests were conducted to examine differences in the adoption of sustainable practices and barriers to using the Agricultural Guide application based on socio-economic characteristics.

Internal consistency was evaluated to assess the reliability of the measurement instrument and its individual items. This approach involves measuring the correlation among test items designed to capture the same construct. A commonly accepted threshold for adequate internal consistency is a Cronbach’s alpha value of 0.7 (36). In the present study, the Likert scale demonstrated a high level of internal consistency, rendering it appropriate for research purposes. Specifically, the internal consistency coefficient (Cronbach’s alpha) for items measuring both the adoption of sustainable agricultural practices via the Agricultural Guide application and barriers to its use was 0.844, indicating a relatively high degree of reliability within the data.

The random forest (RF) algorithm is a well-established machine learning technique that belongs to the family of ensemble models. It employs a bagging strategy, wherein multiple decision trees are constructed using bootstrap samples drawn from the original training dataset. Each tree is trained independently, and the final prediction is obtained by averaging the outputs of all trees. Unlike traditional decision trees, RF does not consider all available features at every split. Instead, it selects a random subset of features for each split, thereby reducing the correlation among trees and preventing the formation of similar tree structures. This approach enhances the overall predictive accuracy of the model [34,35,36,37,38].

Beyond its predictive capabilities, random forest provides an effective mechanism for feature importance estimation. It identifies the most influential variables and quantifies their individual and interactive effects, making it a valuable tool for analyzing complex, high-dimensional datasets. The integration of machine learning techniques, particularly the RF algorithm, enables a comprehensive understanding of the factors influencing farmers’ decisions regarding the adoption of specific applications. Figure 2 presents a decision tree generated during the analysis, illustrating the model’s operational mechanism by partitioning the data. The figure represents one instance from the ensemble of 500 randomly generated trees in this study.

The random forest approach is a machine learning approach. It can calculate the importance of variables using measures such as average drops in accuracy, providing insights into the relative influence of different factors. It can generate robust classifications [34]. The random forest approach was used to analyze the impact of smartphone applications on behavior. This approach proved effective by providing classifications and calculating the most critical changes accurately [39].

A stratified sample of data were taken from the datasets representing two groups of users and non-users due to their different numbers. Then, 70% of the data were allocated for training the model and 30% for testing the model after training. Five hundred decision trees were used after the model showed stability at this number, with randomly selected features in each division. Then, the necessary statistical criteria were applied to evaluate the performance and ensure the stability of the model [35,40].

2.6. Mathematical Formulation of the Model

Final prediction in the random forest model [34]:

$${\widehat{y}}_i\left(x\right)={\displaystyle \frac{1}{T}}{\sum }_{t=1}^T{f}_t\left(x_i\right)$$

(2)

This equation shows the final output of the model prediction ${\widehat{y}}_i\left(x\right)$, which is calculated by adding all the individual predictions for each tree in the model $\sum _{t=1}^T$ by adding the votes resulting from each tree T, where $f_t\left(x_i\right)$ represents the individual prediction of the tree.

Objective function for error minimization [41]:

$$Obj\left(\theta \right)={\sum }_{i=1}^{n}L\left(y_i,{\widehat{y}}_i\right)+{\sum }_{t=1}^T\Omega \left(f_t\right)$$

(3)

This equation is divided into two parts: the first part calculates the difference between the true values $\left(y_i,{\widehat{y}}_i\right)$ and the predicted values $\left(y_i,{\widehat{y}}_i\right)$ using the loss function, while the second part expresses the regularization function, which works to reduce the tree repetition and overfitting and make the model more efficient and stable.

Mean decrease in accuracy equation [34]:

$$\Delta A={\displaystyle \frac{1}{T}}{\sum }_{t=1}^T\left(A_{\mathrm{original}}-A_{\mathrm{permuted}}\right)$$

(4)

This equation measures the degree of importance of each variable based on its impact on the model’s performance by measuring the original prediction accuracy $A_{\mathrm{original}}$ and the model’s accuracy after rearranging the variable values $A_{\mathrm{permuted}}$. The change was performed randomly. If the decrease in accuracy is large, it indicates that the variable is important in the model.

2.7. Model Evaluation

The model was evaluated using the following performance metrics:

Accuracy [42]:

$$\mathit{Accuracy}={\displaystyle \frac{\mathit{Number} \mathit{of} \mathit{Correct} \mathit{Predictions}}{\mathit{Total} \mathit{Number} \mathit{of} \mathit{Samples}}}$$

(5)

Sensitivity [43]:

$$\mathit{Sensitivity}={\displaystyle \frac{\mathit{True} \mathit{Positives}}{\mathit{True} \mathit{Positives}+\mathit{False} \mathit{Negatives}}}$$

(6)

Specificity [44]:

$$\mathit{Specificity}={\displaystyle \frac{\mathit{True} \mathit{Negatives}}{\mathit{True} \mathit{Negatives}+\mathit{False} \mathit{Positives}}}$$

(7)

The main packages used in the analysis were random forest [40] for model construction and significance, caret [45], for data classification and analysis, pROC and AUC [46] for ROC curve analysis, and ggplot2 for data visualization [47]. They were used to generate the random forest and analyze the data.

3. Result

3.1. Socio-Economic Characteristics of the Respondents

Table 1. Socio-economic characteristics of the respondents.

Variable	Frequency (n = 229)	Percent	Variable	Frequency (n = 229)	Percent
Age			Education
Less than 40 years	41	17.90	Literate	5	2.2
40–56 years	109	47.60	Primary	18	7.8
More than 56 years	79	34.50	Middle	20	8.7
Mean	51.5		Secondary	63	27.4
SD	11.32		College	101	43.9
			Higher education	22	9.6
Farming experience			Land area
Less than 5 years	11	4.8	Less than 20 dunums	13	5.7
5–10 years	41	17.8	21–40 dunams	23	10
10–15 years	50	21.7	From 41 to 60 Dunam	50	21.7
More than 15 years	124	53.9	61–80 dunams	38	16.5
			More than 80 Dunam	104	45.2
Number of olive trees on the farm			Olive production per ton
Less than 1050 trees	23	10.3	Less than 30 tons	21	9.7
1050–1820 trees	117	52.2	30–52 tons	100	46.1
More than 1820 trees	84	37.5	More than 52 tons	96	44.2
Mean	1643.08		Mean	48.37
SD	465.61		SD	13.81
Income (thousand SAR)			Relying on agriculture as the sole source of income
Less than 130 K SAR	23	10	No	165	71.7
131–227 K SAR	143	62.4	Yes	64	27.8
More than 227 K SAR	63	27.5
Mean	196.04
SD	52.84
Employment			Planting method
Agricultural farms	97	42.2	Traditional	143	62.2
Government employees	91	39.6	Mixed	13	5.7
Merchant	9	3.9	Dense	73	31.7
Craftworker	4	1.7	Harvest
Retired	19	8.3	Manual	190	82.6
Other	9	3.9	Mechanism	36	15.7
Marketing			Irrigation methods
In Al-Jouf region	209	91.3	Immersion	18	7.8
Outside Al-Jouf region	128	55.9	Modern irrigation	211	91.7
Outside Saudi Arabia	39	17

presents the demographic and agricultural characteristics of the respondents. The data indicate that over 40% of the participants were college graduates within the age range of 40 to 56 years. More than half of the respondents had over 15 years of farming experience. A significant proportion (over 40%) owned agricultural farms exceeding 80 Dunams in size. Furthermore, more than 50% of the respondents cultivated between 1050 and 1820 olive trees, while over 40% reported harvesting in excess of 52 tons of olive fruit annually. Despite their agricultural engagement, fewer than 30% of the respondents relied solely on farming as their primary source of income. More than 60% reported an annual income ranging between 131,000 and 227,000 SAR. Over 40% of the respondents were employed in the agricultural sector, and a substantial majority (90%) marketed their produce in the Al-Jouf region. Regarding agricultural practices, more than 60% of the respondents employed traditional olive farming techniques, while over 75% relied on manual harvesting methods. However, modern irrigation systems were widely adopted, with over 90% of respondents utilizing these technologies.

3.2. Farmers’ Sources of Information

Figure 3 illustrates that the majority of farmers utilized cell phones and the internet for more than four hours per day. In contrast, fewer than 10% of farmers engaged with these technologies for less than an hour daily. Meanwhile, approximately 10% to 16% of farmers used mobile phones and the internet for one to four hours per day.

Table 2. Sources of information for farmers: digital vs. non-digital.

Sources of Information		Frequency	Percentage	Percentage of Source Usage *
Traditional information sources	Family members	75	7.71	32.75
	Contact other farmers	197	20.25	86.03
	Rural leader	64	6.58	27.95
	Local agricultural materials dealer	141	14.49	61.57
	Agricultural cooperative	109	11.20	47.60
	Local plant protection station	26	2.67	11.35
	Agricultural technology books	18	1.85	7.86
	Agricultural magazine	11	1.13	4.80
	Agricultural newspapers	8	0.82	3.49
Digital information sources	Social media groups (WhatsApp, Snapchat, Twitter)	172	17.68	75.11
	Agricultural website	107	11.00	46.72
	Mobile text message	45	4.62	19.65

* Percentage represents the proportion of participants who indicated “Yes” for each information source.

illustrates the utilization of both digital and traditional information sources for acquiring agricultural knowledge. Among non-digital sources, over 85% of respondents obtained information from fellow farmers. In contrast, more than 75% of respondents relied on social media platforms, such as WhatsApp, Snapchat, and Twitter, for agricultural insights.

3.3. Use of Agricultural Guide App

Figure 4 presents that the majority of the farmers (61%) in the study area did not use the Agricultural Guide application. Further, around 39% of the farmers who reported the use of the application.

Figure 5 shows the extent of use of the Agricultural Guide application. Approximately two-fifths of the farmers used the Agricultural Guide application. Additionally, one-fourth used it once a month, while 3.4% to 17% used it for 1–2 h to once a week.

3.4. Adoption of Sustainable Olive Farming Practices

Principal Component Analysis (PCA) was conducted to better understand the adoption of sustainable olive farming practices (

Table 3. Results of principal component analysis: extracted factors and loadings.

Statements	Factor
	1	2	3	4
Identifying suitable seedlings		0.716
Seasonal tillage and soil turning		0.696
Soil sterilization		0.576
Use of modern irrigation systems for watering trees.	0.538
Stopping or reducing irrigation during the flowering period.	0.519
Watering early in the morning or evening.	0.593
Watering away from the tree trunk.	0.517
Use the recommended agricultural pesticides according to the instructions.			0.595
Using natural and organic pest control to eliminate crop pests and diseases.			0.626
Using sticky, plastic, and pheromone traps to reduce insect damage.			0.600
Collecting fallen fruits separately and not mixing them with the rest of the fruits.				0.717
Using plastic combs while picking fruits.				0.742
Extraction Method: Principal Axis Factoring. Rotation Method: Varimax with Kaiser Normalization.

Factor 1: percentage of explained variance = 22.6%; Factor 2: percentage of explained variance = 13.4%: Factor 3: percentage of explained variance = 9.12%; Factor 4: percentage of explained variance = 8.98%.

). Four factors were extracted that together explained 54.1% of the total variance. Based on their factor loading values, Factor 1 comprises four items: (1) use of modern irrigation systems for watering trees, (2) stopping or reducing irrigation during the flowering period, (3) watering early in the morning or evening, and (4) watering away from the tree trunk. Factor 2 includes three items: (1) identifying suitable seedlings, (2) seasonal tillage and soil turning, and (3) soil sterilization. Factor 3 includes three items: (1) using the recommended agricultural pesticides according to the instructions, (2) using natural and organic pest control to eliminate crop pests and diseases, and (3) using sticky, plastic, and pheromone traps to reduce insect damage. Factor 4 includes two items: (1) collecting fallen fruits separately and not mixing them with the rest of the fruits and (2) using plastic combs while picking fruits. These factors highlight key agricultural practices that influence olive crop management and sustainability.

3.5. Factors Affecting the Adoption of Sustainable Agricultural Practices

Table 4. Adoption of sustainable agricultural practices among farmers.

Categories of Practices	Adoption Level
	Both Groups		User		Non-User
	Mean (%)	SD	Mean (%)	SD	Mean (%)	SD
Irrigation management Practices	66.72	3.43	75.27	2.56	57.12	3.16
Soil improvement Practices	62.36	2.84	76.37	2.01	53.23	2.43
Pest control Practices	59.39	2.54	70.44	2.24	51.99	2.10
Harvesting Practices	66.20	1.83	72.55	1.47	55.03	1.72

reveals the adoption rates of various agricultural practices among users and non-users of the Agriculture Guide application. Among all practices, irrigation management showed the highest overall adoption rate (66.72%), followed closely by harvesting (66.20%). Soil improvement and pest control demonstrated comparatively lower adoption rates, at 62.36% and 59.39%, respectively. Notably, adoption rates were significantly higher among users of the application. Soil improvement records the highest adoption rate among users (76.37%), followed by irrigation management (75.28%). In contrast, among non-users, the adoption rates for soil improvement and irrigation management were 53.23% and 57.12%, respectively.

Table 5. t-test comparison between application users and non-users across selected independent variables.

Use of the Agricultural Guide Application		N	Mean	S. D	t	p	Cohen’s d
Olive Production (tons)	No-user	139	45.46	13.60	−4.01	0.000	0.54
	User	90	52.59	12.36
Income (thousand SAR)	No-user	139	182.83	49.46	−4.93	0.000	0.66
	User	90	216.44	51.64
Traditional information sources	No-user	139	30.71	18.76	0.17	0.861
	User	90	29.71	18.60
Digital information sources	No-user	139	29.70	25.04	−13.12	0.000	1.77
	User	90	73.88	24.61

presents the results of an independent t-test conducted to examine differences in annual olive production, farm revenue, and reliance on digital and traditional information sources between users and non-users of the Agricultural Guide application. The findings indicate statistically significant differences in olive production (t = −4.01, p < 0.001), farm revenue (t = −4.93, p < 0.001), and the use of digital information sources (t = −13.12, p < 0.001). Specifically, application users reported significantly higher olive production compared to non-users, with the difference in means representing a medium effect size (Cohen’s d = 0.54). Similarly, farmers who utilized the application generated higher farm revenue than their non-user counterparts, with a medium effect size (Cohen’s d = 0.66). Additionally, users demonstrated a significantly greater reliance on digital information sources, with the difference in means indicating a large effect size (Cohen’s d = 1.77).

Table 6. The t-test comparison for differences in the farmers’ adoption of sustainable agricultural practices according to users and non-users of the application.

Use of the Agricultural Guide Application		N	Mean	S. D	t	p	Cohen’s d
Adoption of sustainable practices	Non-user	139	2.66	0.55	−12.84	0.000	1.65
	User	90	3.50	0.44
Irrigation management practices	Non-user	139	2.96	0.80	−9.04	0.000	1.23
	User	90	3.91	0.72
Soil improvement practices	Non-user	139	2.61	0.75	−11.27	0.000	1.63
	User	90	3.84	0.73
Pest control practices	Non-user	139	2.97	0.86	−8.71	0.000	1.17
	User	90	3.83	0.76
Harvesting practices	Non-user	139	2.64	0.85	−7.67	0.000	1.05
	User	90	3.91	0.72

presents the results of an independent samples t-test conducted to examine differences in the adoption of sustainable agricultural practices between users and non-users of the application across these domains: irrigation management, soil improvement, pest control, and harvesting practices. The analysis revealed statistically significant differences in the adoption of sustainable agricultural practices based on the application usage. Specifically, significant differences were observed in overall application usage (t = −12.84, p < 0.001), irrigation management (t = −9.04, p < 0.001), soil improvement (t = −11.27, p < 0.001), pest control (t = −8.71, p < 0.001), and harvesting practices (t = −7.67, p < 0.001). Users of the application demonstrated a significantly higher adoption of sustainable agricultural practices compared to non-users. The effect sizes, measured using Cohen’s d, indicated a large effect across all domains. Specifically, the adoption of sustainable agricultural practices among users showed a substantial effect (Cohen’s d = 1.65). Similarly, application users exhibited higher adoptions of irrigation management practices (Cohen’s d = 1.23), soil improvement practices (Cohen’s d = 1.63), pest control practices (Cohen’s d = 1.17), and harvesting practices (Cohen’s d = 1.05), all demonstrating large effect sizes.

3.6. Factors Influencing the Use of the Agricultural Guide Application

Random forest analysis was used to identify the factors affecting the use of the application. This analysis aims to predict the most critical factors influencing the use of the application. The results of the model are as follows.

3.6.1. Overall Model Performance

Figure 6 presents the random forest model is suitable for predicting the effect of variables on users and non-users. The accuracy of the model was 94%. To further enhance the effectiveness of the model evaluation, the sensitivity and specificity were calculated as 96.3% and 92.68%, respectively. The out-of-bag (OOB) error rate was 10%, which shows the model’s good performance.

3.6.2. Model Optimization Analysis

Figure 7 showed the performance of the model was improved by adjusting the number of trees and the parameters used in each split, and the results indicate that using 500 trees achieved stable performance while reducing the error rate, which contributed to improving the accuracy of the test data. This analysis emphasizes the importance of choosing an appropriate number of trees to accurately classify application users and non-users. Figure 7 shows the stability of the model through the number of trees and splits.

Confusion matrix estimation in random forest analysis: Figure 8 presents the confusion matrix revealed that the model correctly classified 38 samples as non-user and 26 samples as user, while misclassifying three samples as “non-user” and one sample as “user”. Furthermore, the F1-Score analysis showed a balanced performance, scoring 0.97 for the “non-user” class and 0.93 for the user class. These results confirm the robustness of the model and its ability to effectively address the analytical objectives of the study.

3.6.3. Analysis of Factors Influencing Application Usage

Figure 9 presents the most influential factors affecting the utilization of the application. Among these, digital information sources emerged as the most significant determinant, with a Mean Decrease Accuracy (MDA) value of 24.78. This finding suggests that farmers who depend on information and communication technology (ICT) for agricultural knowledge exhibit a higher propensity to adopt the application. This correlation may be attributed to their familiarity with digital tools, which facilitates ease of use. Additionally, electronic marketing campaigns conducted via platforms such as X (formerly Twitter) and Snapchat by the Agricultural Extension Department may play a role in promoting the application’s adoption. Soil improvement and irrigation management practices ranked as the second and third most influential factors, with MDA values of 20.24 and 16.98, respectively. These results highlight the relevance of the application in enhancing agricultural practices and indicate that increased engagement with the application correlates with a higher adoption rate of sustainable farming methods. Education also demonstrated a considerable impact on application usage, ranking fourth with an MDA value of 13.37. This finding suggests that farmers with higher educational attainment are more inclined to utilize the application, thereby facilitating the adoption of advanced agricultural techniques. Other factors, including pest control practices, daily mobile phone usage, olive production, harvesting practices, marketing outside the Al-Jouf region, farmer age, and farmer income, exerted a moderate influence on the application usage. These results underscore the multifaceted nature of technology adoption in agriculture and emphasize the importance of integrating digital solutions to enhance farming efficiency and sustainability.

Figure 10 illustrates a significant correlation between olive production and annual farm income, indicating that an increase in production corresponds to a proportional rise in farmers’ earnings. Additionally, a significant correlation was observed between mobile phone usage and internet consumption. This suggests that increased mobile phone usage is associated with higher internet utilization, which, in turn, leads to greater engagement with the Agricultural Guide application. Moreover, a negative correlation was observed between age and the use of the Agricultural Guide application, implying that older individuals are less likely to utilize the application. Conversely, a positive correlation is evident between educational attainment and application usage, indicating that individuals with higher levels of education are more inclined to use the application.

3.7. Barriers to Using the Application

Table 7. Barriers to using the application.

Statements	N	Strongly Disagree	Disagree	Neutral	Agree	Strongly Agree	Mean	S.D	Rank
		%	%	%	%	%
The network and internet services in my area are insufficient to access the app’s services.	137	22.6	35.8	23.4	18.2	0.0	2.37	1.03	16
There is not enough information about my agricultural activity in the app.	136	5.1	5.1	22.8	36.0	30.9	3.82	1.09	2
Eye strain and fatigue due to using mobile devices for agricultural advisory.	136	6.6	14.7	16.2	27.2	35.3	3.70	1.27	5
I prefer to receive agricultural knowledge and information directly from the advisor.	137	10.9	20.4	48.2	17.5	2.9	2.81	0.95	14
The electronic information does not match my agricultural activity.	138	2.9	42.8	27.5	23.2	3.6	2.82	0.95	13
The electronic information does not fit my region.	137	19.7	21.9	38.7	15.3	4.4	2.63	1.10	15
I prefer not to download many applications due to low phone memory.	137	0.7	13.9	21.2	37.2	27.0	3.76	1.03	4
I am bothered by the number of apps on my phone.	137	2.2	10.2	23.4	29.2	35.0	3.85	1.08	1
I don’t trust the apps and feel that they invade my privacy.	137	0.0	17.5	20.4	28.5	33.6	3.78	1.10	3

Summated mean = 52.22; SD = 8.9; range = 35; low = 33; and high = 68; item mean = 3.25.

presents the challenges associated with the use of the Agricultural Guide application among farmers. The most prominent challenge identified was the perceived inconvenience of having an excessive number of applications installed on mobile devices, with a mean score of 3.85 and a standard deviation of 1.08. This was followed by the insufficiency of relevant agricultural information within the application, which had a mean score of 3.82 and a standard deviation of 1.09. Concerns regarding data privacy and trust in mobile applications also emerged as a significant barrier, with a mean score of 3.78 and a standard deviation of 1.10. Additionally, the issue of device performance deterioration due to excessive applications was reported, with a mean score of 3.76 and a standard deviation of 1.03. Conversely, the least significant challenges included the incompatibility of digital agricultural information with farmers’ specific activities (mean = 2.81, SD = 0.95), a preference for face-to-face consultations with agricultural advisors (mean = 2.81, SD = 0.95), the unsuitability of digital information for certain geographical regions (mean = 2.63, SD = 1.10), and inadequate internet services hindering access to the application (mean = 2.37, SD = 1.03).

4. Discussion

This study highlights the pivotal role of the Agricultural Guide application in promoting sustainable agricultural practices among olive farmers in the Al-Jouf region. The application not only enhances agricultural resource management—particularly in soil and water conservation—but also increases olive production, thereby contributing to farm sustainability. Additionally, the timely provision of information via information and communication technologies has improved farmers’ economic outcomes. These findings are consistent with previous research [6,48,49]

4.1. Comparison of Agricultural Information Sources and Their Role in Sustainable Practices

Agricultural information sources are critical to enhancing production and supporting farmers in addressing challenges and making informed decisions. The present study demonstrates that these sources substantially influence the adoption of digital technologies. Notably, social media emerged as a predominant channel, with many farmers relying on these platforms. This observation aligns with Alotibi and Dabiah’s findings [50], which identified social media as a key agricultural information source in Saudi Arabia, and with those of Al-Sakran et al. [51], who reported its extensive use among Saudi farmers. Additionally, Aldosari et al. [52] argued that a majority of Saudi farmers consider the internet a valuable resource for agricultural information. The utilization of digital platforms such as social media not only facilitates interaction among farmers but also contributes to narrowing the knowledge gap [6]. In this context, the adoption of the Agricultural Guide application could further enhance farmers’ technical knowledge and crop production. Observations from the study indicate a willingness among farmers to learn about and integrate this application into their agricultural practices.

Conversely, traditional sources, particularly peer-to-peer exchanges, continue to play a significant role in agricultural extension. Approximately 80% of farmers reported relying on information from their peers, which is consistent with previous studies [53,54]. Therefore, introducing innovative extension models like the Agricultural Guide application may promote the adoption of improved agronomic practices and mitigate the limitations associated with outdated technologies [55]. For instance, in China, platforms such as WeChat have been instrumental in providing innovative agricultural information and addressing technical issues related to irrigation, pest control, and soil testing [56]. Similarly, Tong et al. [57] recommends technology demonstrations as a means to enhance farmers’ acceptance of innovative agricultural technologies. Implementing similar initiatives in the current study area could further promote the use of the Agricultural Guide application.

4.2. Role of the Agricultural Guide in Promoting the Adoption of Sustainable Olive Farming Practices

The results indicate that the adoption of agricultural practices, including irrigation management and soil improvement, was significantly higher among users of the Agricultural Guide than among non-users. This finding underscores the crucial role of the application in disseminating effective agricultural guidance and enhancing productivity. These results are consistent with Li et al. [58] indicated that agricultural applications improve fertilization practices, and Palmer and Darabian [59] argued that users of innovative agricultural extension applications improved sustainable control practices for their crops. This suggests that digital extension services can effectively upgrade farming techniques [60]. Moreover, digital applications have proven to be more effective than traditional extension methods [61]. For instance, the Agricultural Guide application delivers high-quality advisory services on agronomic practices and pest control through contributions from international experts, thereby illustrating the positive impact of digital technology on agricultural efficiency. Similarly, the AgroTIC app in Colombia has effectively improved farming practices by utilizing artificial intelligence to diagnose pests and connect farmers with experts and markets, thereby supporting agricultural sustainability and food security [62].

4.3. Factors Influencing the Use of the Agricultural Guide Application

Usage of the application is influenced by a range of social, economic, technological, and behavioral factors, reflecting the complex interplay among these elements that ultimately drives adoption [63]. According to the random forest analysis, the adoption of agricultural practices related to water management and soil improvement significantly affects application usage. This may be attributed to the greater openness of farmers who already implement such practices to modern technologies, including digital agricultural extension tools [64,65]. The random forest model demonstrated a high predictive accuracy (94%), which aligns with previous studies employing similar methods. For instance, one study achieved an accuracy of 87% when examining agricultural extension researchers’ attitudes toward internet usage, while another assessing graduate students’ attitudes toward mobile technology also reported high accuracy [66,67]. These findings collectively underscore the robustness of the model in evaluating agricultural practices.

Furthermore, the random forest analysis identified the use of digital information sources as the most critical factor influencing application usage. Heightened technological awareness and improved access to digital resources simplify both awareness of and engagement with the application [49,68]. This outcome supports the argument that education exerts a substantial impact on technology adoption, consistent with the Unified Theory of Acceptance and Use of Technology 2 (UTAUT2) [69,70]. Moreover, the duration of daily use of smart devices was found to influence application adoption, presumably because frequent interaction with mobile devices fosters habitual behavior, thus facilitating adoption [71].

Marketing activities outside the Al-Jouf region also shaped application usage. Farmers who seek market information and crop prices tend to consult reliable information sources, such as the Agricultural Guide application, which offers daily market data [24]. This reliable information can improve agricultural income, thereby fostering application adoption [72]. In addition, the volume of olive production emerged as a significant factor, suggesting that the economic capacity of farmers can promote the adoption of modern technologies for accessing agricultural information [73,74].

Lastly, farmers’ age was found to be influential in determining application usage patterns. Younger farmers, being more technologically adept, are more likely to adopt digital applications [69,70]. In contrast, older farmers often experience discomfort with technology, stemming from concerns related to security and privacy, as highlighted by Kenny and Regan [75]. These findings underscore the importance of designing targeted strategies to address generational differences and encourage wider adoption of digital agricultural tools among older farmers.

4.4. Barriers to the Use of the Agricultural Guide Application: Recommendations

The findings indicate that most farmers do not utilize the application, opting instead to rely on alternative sources of agricultural information. The primary obstacle identified was the excessive number of applications installed on farmers’ mobile devices. As Leso and Cortimiglia [76] suggest, a proliferation of applications and complex user interfaces can significantly hinder adoption. This issue may be exacerbated by a limited awareness of the application’s potential benefits, leading to inadequate user retention. Furthermore, the absence of dedicated content reduces the application’s perceived value, aligning with the Unified Theory of Acceptance and Use of Technology 2 (UTAUT2), which underscores the necessity of providing localized agricultural information tailored to end users [73].

Another critical barrier is privacy and trust, as many individuals hesitate to engage with digital technologies due to concerns about data security [75]. These findings are consistent with theoretical frameworks that emphasize the importance of performance expectations and trust in driving technology acceptance. Notably, the study reveals that only 39% of farmers currently use the application, a relatively low percentage given the considerable governmental support it has received. Addressing these issues calls for improved interface design, stronger privacy safeguards, and meaningful engagement with farmers to identify and integrate their informational and educational needs.

5. Conclusions

This study aimed to evaluate the effect of the Agricultural Guide application on the adoption of sustainable olive farming practices in the Al-Jouf region of Saudi Arabia. The findings indicate that farmers who used the application exhibited a higher rate of adoption of sustainable agricultural practices than those who did not. In addition, the study revealed that socio-economic factors, such as level of education, income, and frequency of mobile phone use, significantly influenced the degree to which farmers utilized the application. The results further suggest that digital information sources can effectively increase awareness of modern agricultural practices. Despite these benefits, several barriers to using the Agricultural Guide application were identified. The most frequently cited challenge was farmers’ discomfort with the proliferation of applications on their mobile devices. These findings highlight the need for policymakers to organize extension programs and workshops to promote awareness of, and proficiency in, agricultural innovations like the Agricultural Guide application. Moreover, it is recommended that the Ministry of Environment, Water, and Agriculture address the obstacles that hinder widespread application usage. Finally, the contents of the Agricultural Guide application should be regularly updated to better align with farmers’ evolving needs.

6. Study Limitations

One main limitation of the study is that it focused only on the farmers who were involved in olive production in the Al-Jouf region. Therefore, it restricts our generalizability to extend the scope of the study to other farmers involved in other agricultural activities and how the Agricultural Guide influences their agricultural practices. Second, the measurement of practices was based on the farmers’ self-assessment, which may lead to a kind of bias. Therefore, it is recommended that the same research should be conducted to promote the adoption of the Agricultural Guide in other regions.