Introduction

Automation in the wine industry has led to a revolution in reducing costs and improving quality [18], allowing producers to focus on innovation and flavor improvement [68], as well as enhancing detailed production record keeping [1], which is essential for early detection of problems and ensuring consumer satisfaction.

Automation can be integrated throughout the entire value chain of the wine industry, starting with the implementation of advanced machinery in grape harvesting, which facilitates winemakers to accelerate this essential process [7], continuing with grape pressing, optimizing the extraction of juice from the skins in an efficient manner [65] and ending with the bottling, labeling and transportation phase, allowing producers to increase their production capacity in reduced timeframes [4] and ensuring faster and more accurate execution [58].

Nonetheless, the adoption of automation in the wine sector is not without challenges; first, the wine industry is characterized by its deep roots in traditions and processes that have remained unchanged for years, making the automation of these a task of considerable complexity [61]; second, the acquisition of automated equipment represents a significant investment, forcing wineries to carefully evaluate the profitability of such investment and the compatibility of new technology with existing processes [54],third, a technical skills gap may exist in the workforce, as winery employees lack experience with advanced technologies, which can complicate the effective implementation of automated solutions [68] and, fourth, although automation promises quality improvements, it is imperative that automated systems be meticulously designed and evaluated to ensure compliance with established quality and safety standards [21].

The main objective of this study is to examine the structure of knowledge associated with automation in the wine industry since, to the best of the authors' knowledge, there are no previous literature reviews that have addressed this topic, making it possible to discover for the first time the key characteristics of this scientific production, such as the leading institutions, prolific authors and countries where production is concentrated, as well as to examine the main topics that currently occupy researchers in this field. To achieve this end, a bibliometric analysis is initially conducted, proceeding with the quantitative examination of the scientific production, and, subsequently, a systematic analysis is carried out to review in-depth the 39 records identified from 1996 to 2022, thus allowing to pinpoint the research fronts on the subject and, based on them, to propose future research lines. In fact, such a mixed methodological approach, based on bibliometric (quantitative approach) and systematic (qualitative approach) review, is carefully designed to provide a comprehensive view on contemporary research in the specified area, serving as a navigational tool through the complex landscape of prevailing trends, leading scholars and organizations, major publications in the domain, research fronts, and the future trajectory of publications.

Thus, the study aims to answer the following three Research Questions (RQ): (RQ1) What are the current trends and future prospects for automation in the wine industry? (RQ2) How is the scientific knowledge around this topic structured? (RQ3) What are the implications of these trends for academics, practitioners and policy makers in the wine industry? By addressing these RQs, the investigation aims to improve the understanding of automation in the wine industry, with the study being valuable for both wine academics and wine industry practitioners.

In order to achieve the objective set, the introductory section is followed by Materials and methods, which details the methodology employed, Results and discussion presents the findings of the study, and finally, Conclusions and future research agenda shows the main conclusions derived from the study, the research agenda, as well as the limitations and future research lines.

Materials and methods

This study conducted a bibliometric analysis using the Web of Science (WoS) database. Boolean and proximity operators, along with markers, were utilized to evaluate the quality and accuracy of the selected works. It should be noted that the WoS Core Collection was chosen for its rigorous inclusion criteria for articles.

The three indexes selected from the WoS Core Collection were the Science Citation Index Expanded (SCI-E), Social Sciences Citation Index (SSCI), and Emerging Sources Citation Index (ESCI). SCI-E is an extensive index of citations from scientific and technological journals since 1900, including 8,000 scientific journals and 12,000 conference and press journals [15]. SSCI includes references from approximately 3000 social science journals, featuring both press and conference journals [13]. ESCI is an index of citations which encompasses emerging science journals as well as some press and conference journals, in an effort to represent the variety of scientific publications, such as those from developing countries, and includes more than 5000 journals [14].

After evaluating the value of the WoS Core Collection, a search was implemented to identify articles related to the topic. Multiple efforts were made to evaluate the most relevant results and the least pertinent ones in order to eliminate any extraneous discoveries. Ultimately, it was determined that the most appropriate search equation among the available options was the following:

$${\text{TS}}\, = \,\left( {\left( {{\text{automation}}} \right)\;{\text{AND}}\;\left( {{\text{wine}} * } \right)} \right)$$

The search equation was divided into two categories: automation and the wine industry. The AND operator was utilized to limit the results to papers that provided insights from both groups, while the wildcard (*) was used to include different word forms in the accessible results. It is important to note that only the AND operator was used because we wanted to prioritize the optician of results linked only and exclusively to the pre-established research objective. Likewise, the wildcard symbol was not applied to the automation category to avoid obtaining results that were not directly and explicitly linked to automation. Boolean operators are a major element of bibliometric reviews as they enable users to formulate intricate and specific queries to gain more relevant information. They can also be employed to narrow or extend the search to particular topics, documents of certain years or types, allowing researchers to get the most meaningful results while reducing searching time. These parameters were applied to the title, abstract, and keywords of the papers, and the documents added by WOS until 2022.

On October 5, 2023, the application of a search algorithm resulted in the acceptance of 39 articles. In order to analyze the scientific output, the “Preferred Reporting Items for Systematic Reviews and Meta-Analyses” (PRISMA statement) was utilized due to its potential to enhance the reliability and reproducibility of reviews, its comprehensive nature, and its widespread application in bibliometric studies [43]. The PRISMA methodology is beneficial for augmenting the transparency and communication between authors and readers, thus verifying that the results are dependable and reproducible [41]. This is recommended for the purpose of improving the quality of research studies and allowing readers to comprehend the results more easily, which reduces the chances of bias and mistakes in data collection [72]. After accepting only scientific production in the form of articles and eliminating research that did not aim to analyze automation in the wine industry, the number of documents was reduced from 108 to 39, thus forming the corpus of articles to be examined (see Fig. 1). In particular, after the selection of articles only, the number of records was reduced from 108 to 77 and, subsequently, articles were checked record by record to ensure that they explicitly addressed automation in the wine industry, eliminating 38 records that did not address automation in the wine industry, thus resulting in a sample of 39 articles.

Fig. 1
figure 1

Source: own elaboration based on PRISMA guidelines

Flow diagram about the bibliometric review procedure developed.

Full size image

The analysis of the scientific production was conducted by selecting multiple classification variables. Initially, the records were separated according to the year of publication in various journals to determine the level of interest in the topic over time. Subsequently, the main journals for disseminating research results were specified. In addition, a network and overlay map was generated through VosViewer to find out the main keywords used in the studies analysed, as well as their study over time. Authors were identified and the institutions to which the authors belonged were studied. This analysis was completed with the study of the network of collaborations, as well as the analysis of scientific production by country, both of which were carried out using the Bibliometrix software. Finally, once the scientific production had been quantitatively examined using the VosViewer and Bibliometrix applications, the content of the articles examined was reviewed following the PRISMA guidelines.

Results and discussion

The present study analyzed a collection of articles on automation in the wine industry. The findings shown in Table 1 account for the relevant data of the papers, such as the number of sources and papers and the average age of the articles, as well as the number of citations received and references used. The keywords and authors of each paper are also included.

Table 1 General information on the scientific production analyzed
Full size table

As for the temporal evolution of the scientific production, Fig. 2 shows how it is distributed irregularly from 1996, the year in which the first academic articles on the subject were published, to 2022. Specifically, the distribution of scientific production follows the shape of sawtooths, experiencing three notable peaks with 4 publications relating to the years 2005, 2014 and 2019. However, despite the discontinuous evolution, it is necessary to highlight the efforts to address the subject throughout the period analyzed. This academic effort on the subject may be due to several factors. On the one hand, the study of automation solutions in this sector allows greater efficiency and productivity in an increasingly competitive industry, enabling greater precision in manufacturing processes, resulting in higher quality wines [44]. Automation can help reduce labor costs and save time, resulting in higher profits for producers, and can also help improve quality control and food safety, which are of great importance to the industry [22]. On the other hand, the growth of scientific production in the WoS Core Collection in the last decade, derived from the increased indexing of journals in this collection [60], may result in an increasing number of computable articles.

Fig. 2
figure 2

Source: own elaboration based on Bibliometrix®

Evolution over time of the scientific production analyzed.

Full size image

Figure 3 shows the main avenues for disseminating research results related to automation in the wine industry. The first journal is Analytica Chimica Acta (4), from the Elsevier publishing house, followed by Journal of Agricultural Engineering (3), Journal of Chromatography A (3), Journal of AOAC International (2), Water Science and Technology (2), Agriculture and Human Values, (1), Analytical Chemistry (1), Chemical Papers (1), Computers and Electronics in Agriculture (1) and Data (1). It should also be noted that the top three publishers in terms of scientific production indexed in one of their associated journals are Elsevier (14), Springer (4) and Wiley (3).

Fig. 3
figure 3

Source: own elaboration based on Bibliometrix®

List of the top ten most prolific journals in the field of automation in the wine industry.

Full size image

To determine the topics analyzed in the scientific production under study, an analysis of co-occurrence of keywords was carried out, as shown in Fig. 4. For practical reasons, keywords that appear at least 3 times in the records considered were included. As can be seen, there are 7 clusters of keywords around the subject matter.

Fig. 4
figure 4

Source: elaborated on the basis of WoS and VOSviewer

Network map on the co-occurrence of keywords.

Full size image

Cluster 1, in light blue, is headed by the word automation, agricultural robot, gas-chromatography and energy. In the context of viticulture, agricultural robots automate labor-intensive tasks such as pruning, harvesting and monitoring vineyard conditions, thereby increasing efficiency and reducing manual labor [51]. Also, gas chromatography is essential for analyzing wine composition, including aroma and flavor compounds, which are crucial for quality control and inclusion of the keyword energy can be linked to the growing concern for energy-efficient practices in wine production, where automation technologies can play an important role in reducing energy consumption and optimizing resource utilization [36].

Cluster 2, in red, is headed by the word grape, liquid, ethanol, acetic-acid bacteria and spectroscopy. This cluster covers the transformation of grapes into wine, focusing on the chemical composition of the liquid, such as ethanol production during fermentation. On the one hand, acetic acid bacteria are a concern in winemaking because of their role in wine spoilage, leading to unpleasant vinegar flavors [47]. On the other hand, spectroscopy is an essential analytical tool to evaluate wine composition, including the detection of ethanol and other chemical components, which is crucial for quality control and to ensure the consistency of the final product [11].

Cluster 3, in pink, has as main banners the words fermentation, design and amperometric glucose. This group emphasizes the design of fermentation technologies and control systems, such as amperometric sensors for glucose monitoring, as such monitoring is vital to control the fermentation process and ensure that sugars are properly converted into alcohol and other desired end products [62]. In this context, automation is about optimizing and controlling the fermentation process for consistent, high quality wine production [20].

Cluster 4, in yellow, is mainly represented by the keywords bottle storage, samples, low-cost and detectors. Automated bottle storage systems help manage large inventories, ensure optimal aging conditions and track wine provenance [22]. Samplers and detectors refer to automated sampling and detection systems for quality control, which ensure that each bottle meets the desired standards before reaching consumers [9], and the emphasis on the keyword low cost indicates that the aim is to make these technologies accessible and economically viable for wineries of different scales [23].

Cluster 5, in lilac, is headed by the keywords: computer vision, image processing, cork and neural networks. These technologies can be used for a variety of purposes, such as inspecting and sorting grapes, detecting defects in corks, and even analyzing wine color and clarity [59]. In this sense, image processing and neural networks improve the accuracy and efficiency of these tasks, contributing to overall quality control and process optimization [16].

Cluster 6, in light green, is led by the keywords: aroma and solid-phase microextraction. Solid-phase microextraction is a method used to extract volatile compounds from wine, which are fundamental to its aromatic profile [45]. Automation in this field implies the development of high-performance, accurate and consistent methods of aroma analysis, which would contribute to the standardization of the sensory characteristics of wine [50].

Cluster 7, in dark green, is composed of cinerea, berries and botrytis. Automation in this context may involve the development of precision agriculture tools for monitoring and managing vineyard health, using data analytics and machine learning to predict and control the spread of such diseases, thereby ensuring the health and quality of the grapes [26].

The analysis of the keywords is complemented by the study of their use over time, In this regard, Fig. 5 shows that the most commonly used keywords currently used are computer vision, data aggregation, life cycle assessment, precision viticulture, extreme learning machine and collaborative platform.

Fig. 5
figure 5

Source: elaborated on the basis of WoS and VOSviewer

Overlay map of the co-occurrence of keywords.

Full size image

First, computer vision embodies the integration of artificial intelligence and image analysis, enabling accurate monitoring of grape maturity, disease detection and yield estimation, its growing importance representing a move towards more efficient, data-driven vineyard management [68]. Second, data aggregation complements it by amalgamating diverse data sets—from weather conditions to vine health metrics—to facilitate informed decision making, given that, in an industry where subtle environmental variations can significantly influence grape quality, the ability to synthesize broad data sets into actionable information can enable optimization in the use of resources and maximization of yield quality [39]. Third, life cycle assessment emerges as another fundamental concept, underscoring the industry's growing concern for environmental sustainability, since, by assessing the environmental impact of wine production from vine to bottle, this methodology helps identify areas for improvement, whether in energy use, water management or carbon footprint reduction [19]. Fourth, precision viticulture stands out as a testament to the evolving nature of viticulture, emphasizing the use of technology to adapt vineyard practices to the specific conditions of each plot [8]. This method uses advanced technologies such as GPS, remote sensing and Internet of Things (IoT) devices to achieve optimal grape quality and sustainable farming practices, allowing to improve not only wine quality, but also ensures more environmentally friendly and economically viable operations [56]. Fifth, extreme machine learning, a type of artificial intelligence, indicates the industry's move towards innovative and efficient computational models for problem solving, being applied to various aspects of wine production, from predictive analysis in vineyards to quality control in wineries [6]. Sixth, the current use of the keyword collaborative platforms reflects a digital transformation in the wine industry, which facilitate communication and data sharing between the various stakeholders, fostering a more integrated and transparent supply chain [5].

Table 2 displays the articles that have collected the highest total number of citations both globally and locally. Global citations, on the one hand, refer to the cumulative citations that a given article has obtained from any given area and location. On the other hand, local citations are the citations that each individual article has acquired from other articles within the analyzed repository. Furthermore, normalized total citations are the number of citations attributed to a research article or author, corrected for the number of years since publication and the average number of citations received by articles in the same field and time period. This metric takes into consideration variances in citation practices between fields and time periods, as well as the temporal aspect of citation influence. As can be observed, among the 10 most cited articles globally and among the literature analyzed, four publications coincide. These correspond to the research by Chilla et al. [12], Chang et al. [10], Torrijos and Moletta [67] and Ruíz et al. [53].

Table 2 Most cited papers (globally and within dataset)
Full size table

With regard to the main institutions analyzing the subject (see Table 3), it can be seen that the universities that lead the research on automation in the wine industry are University of Cadiz, University of Castilla La Mancha, University of Cordoba and University of Milan (with 3 articles each). Thus, of the top institutions in terms of scientific production on the subject, Spain and China are the countries with the largest number of institutions (tied with 3). This is related to the data shown in Fig. 6 on production by country, given that the Spain is the country with the highest scientific production on the topic under study (15 articles), followed by Italy (10), Chile (6 articles) and China (5 articles).

Table 3 Most prolific institutions (institutions with more than two records)
Full size table
Fig. 6
figure 6

Source: own elaboration based on Bibliometrix®

Scientific production by country.

Full size image

As for the classification of authors, Table 4 shows the most prominent authors according to the number of articles published. As can be seen, Barroso C, Guillen D, and Perezbustamante J, are the authors with the highest scientific production on the subject with 3 articles, followed by De Castro M, Moletta R, Pérez-Correa J, Pérez-Correa J, Sánchez-Rojas J and Torrijos M (with 2 articles each author). Figure 7 also shows how the main authors on the subject are organized into 13 collaborative clusters, being the one formed by Barroso C, Guillen D and Perezbustamante J. the cluster with the greatest weight in terms of jointly published articles.

Table 4 Ranking of leading authors (authors with more than two records)
Full size table
Fig. 7
figure 7

Source: own elaboration based on Bibliometrix®

Author collaboration network.

Full size image

After the bibliometric analysis, in which the literature under study was examined quantitatively, a systematic analysis of the research was carried out. Thus, as can be seen in Table 5, the objective of the research, its methodological approach, the research context and the phase of the wine value chain on which the study is focused were examined. The results show the preponderance of the quantitative versus qualitative methodological approach, the selection of a global framework versus a specific wine context, as well as the greater study of automation in the wine production phase within the wine value chain, versus the viticulture and wine distribution phase. Likewise, by analyzing the content of the articles, three lines of research were identified in relation to the subject matter under study. The first block of research focuses on the development of automated methods to improve sample preparation, detection and analysis of compounds, especially polyphenolics, in wines. There is growing interest in the use of technologies such as high-performance liquid chromatography and mass spectrometry to analyze compounds in wines, suggesting a focus on improving accuracy and efficiency in wine chemical analysis. Second, there is a block of papers focused on identifying efforts to develop automated systems for sorting cork quality and for monitoring fermentation and aroma production in wine, indicating an interest in optimizing product quality and production process efficiency. The third block of research focuses on sustainability and efficient resource management in the wine industry, with a notable trend towards the implementation of sensors and optoelectronic devices for real-time monitoring of viticulture and winemaking processes. These technological tools provide accurate and constant data, facilitating a more efficient and sustainable management of resources, as well as an improvement in the quality of the final product. This focus on the integration of automation and computerization reflects a paradigm shift in the wine industry, given that, by prioritizing both product quality and the efficiency and sustainability of production processes, a new horizon is taking shape for vine cultivation and wine production, where advanced technology plays a crucial role in harmonizing winemaking excellence with environmental responsibility.

Table 5 Analysis of the scientific production examined in the literature review
Full size table

Conclusions and future research agenda

This research examines the structure of knowledge on automation in the wine industry, becoming a valuable resource for novice and experienced academics interested in exploring the development of the scientific literature on the subject, as well as for wine managers to learn about modern trends in automation and, if appropriate, to integrate these innovations in the different stages of the winemaking process.

In order to advance in the research field of automation in the wine industry, future directions are outlined that address emerging and still unexplored domains in relation to the analyzed topic. First, research is proposed on the integration of advanced robotics in viticulture and enology, with special interest in the development of drones and autonomous robots for pruning, harvesting and monitoring of vine and soil conditions, with the aim of increasing the precision of these fundamental operations. Second, it is proposed to examine the application of nanotechnology in winemaking, proposing the use of sensors for real-time monitoring of soil quality, and the development of nanomaterials to optimize wine preservation and packaging, thus opening new avenues for quality control and wine shelf-life extension. Third, it is also suggested to explore the potential of augmented and virtual reality to enrich the wine tasting experience and in the training of winery personnel, offering sensory simulations and virtual environments for learning winemaking techniques. Fourth, it is proposed to examine the role of artificial intelligence in predicting market trends and wine personalization, using advanced algorithms to adapt products to consumer preferences and market demands. Finally, fifth, it is proposed to investigate energy sustainability in the wine industry, exploring specific renewable energy solutions and wine production methods with lower environmental impact, thus marking a path towards innovation and sustainability in the sector. Thus, while the identification of the current research fronts shown in the results and the future research agenda proposed in this section enable to address RQ 1, the quantitative examination of the research examined presented in the results section enables to answer RQ 2.

This research contributes significantly to the theoretical, practical and policy domain in the study of automation in the wine industry. From a theoretical perspective, first, it provides a comprehensive overview of the scientific literature devoted to the analysis of automation in this sector; second, it facilitates researchers to identify key organizations and geographic regions linked to the topic, thus promoting research visits and collaboration on joint projects; third, it makes it possible to locate other experts with whom to collaborate and, potentially, to organize specialized conferences; fourth, it provides guidance in the selection of appropriate journals and publishers for the dissemination of research results; and fifth, the study the study sets the stage for academics to identify emerging trends in the automation of the wine value chain, thus enriching the existing theoretical corpus and opening up new avenues for future research.

From a practical perspective, this study highlights the importance for winery managers to recognize the potential of automation to improve efficiency and quality in wine production. Automation of processes such as fermentation monitoring, cork selection and wine component analysis can increase the accuracy of the final product, vitally important in an industry where quality and product characterization are critical to market success. In addition, the study highlights the crucial role of automation in the sustainability of the wine industry, urging winery managers to consider adopting automated technologies for efficient resource management and minimizing environmental impact. Further practical implications focus on the integration of artificial intelligence and data analytics in winemaking decision-making, as the ability to process large volumes of data to predict market trends, optimize production processes and improve vine disease management opens up new possibilities for competitiveness in an ever-changing marketplace. Ultimately, this study emphasizes the importance of senior management commitment to training and continuing education in the wine industry, because as automation and new technologies gain ground, it is crucial that wine professionals are equipped with the skills and knowledge necessary to use these tools effectively. This would not only improve the efficiency of wine production, but also ensure that the industry can evolve and adapt to technological advances.

From the policy implications angle, this research underscores the need to formulate policies that encourage innovation and the adoption of automated technologies in the wine industry. This would involve incentives for research and development, as well as financial support for wineries, especially small and medium-sized enterprises, to adopt new technologies. It would be also crucial to promote the regulation of automation and artificial intelligence in the wine industry, covering aspects such as security, data privacy and ethics in the use of artificial intelligence. Policies should ensure that the adoption of these technologies is done in a responsible and transparent manner, protecting both consumers and producers. Similarly, policy makers could direct their efforts to promote sustainable practices in viticulture, such as reducing the use of pesticides and efficient water management, supported by automation technologies, which would also imply the need for policies that encourage the adoption of cleaner and more environmentally friendly production practices. Moreover, it suggests the need to consider how automation affects the competitiveness of the wine industry in the global market and how national interests can be balanced with the need to compete in an increasingly technologically advanced marketplace. In this way, the theoretical, practical and policy implications derived from this research provide answers to RQ 3.

While this bibliometric and systematic literature review offers valuable insights, there are certain limitations that must be acknowledged. On the one hand, while the review effectively highlights key research themes and trends, it does not extensively explore the practical implementation and challenges of automation within the wine industry. This gap suggests an opportunity for further research, particularly through a multiple case study approach to better understand how automation impacts operational efficiency in wineries. On the other hand, the analysis in this review is based solely on the Web of Science Core Collection database. While this database is renowned for its comprehensive and high-quality coverage, relying on a single database might result in missing significant articles that are not indexed within it. To address this, future research should aim to expand the bibliometric analysis by including a variety of databases, thereby ensuring a more comprehensive coverage of the scientific literature in this field.