Next Article in Journal
AMF Inoculation Modulates Plant Physiology, Rhizosphere Processes, and Uranium Uptake in Sunflower Under Uranium Stress
Previous Article in Journal
A Lightweight and High-Precision Citrus Detection Model for Unstructured Orchard Environments
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Horticulturae
Volume 12
Issue 6
10.3390/horticulturae12060719
horticulturae-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Review

Smart Technologies and Artificial Intelligence in Sustainable Viticulture: Applications, Benefits, Barriers and Governance for High-Quality Grape Production

by
Evangelia Zoi Nathena
1,
Kyriakos Psyllakis
1,
Despoina Petoumenou
2 and
Emmanouil Kontaxakis
1,*
1
Department of Agriculture, School of Agricultural Sciences, Hellenic Mediterranean University, P.O. Box 1939, GR 71410 Heraklion, Greece
2
Laboratory of Viticulture, Department of Agriculture, Crop Production and Rural Environment, University of Thessaly, GR 38446 Volos, Greece
*
Author to whom correspondence should be addressed.
Horticulturae 2026, 12(6), 719; https://doi.org/10.3390/horticulturae12060719 (registering DOI)
Submission received: 20 April 2026 / Revised: 23 May 2026 / Accepted: 9 June 2026 / Published: 11 June 2026
(This article belongs to the Special Issue Advances in Sustainable Growing Systems for High Quality Grape and Olive)
Browse Figures
Versions Notes

Abstract

Smart technologies and artificial intelligence (AI) are increasingly reshaping viticulture by improving vineyard monitoring, supporting data interpretation, and enabling more targeted management decisions. This review examines how sensor networks, remote sensing, machine learning, deep learning, and decision-support systems contribute to more sustainable vineyard management and the production of high-quality grapes. Particular attention is paid to applications in grapevine stress monitoring, disease and pest detection, irrigation and nutrient management, yield estimation, grape quality prediction, and emerging automation. The review also highlights the main barriers that still limit broader adoption in commercial vineyards, including data quality issues, limited transferability across sites and seasons, interoperability gaps, vendor lock-in, and concerns related to governance, privacy, and cybersecurity. Although these constraints remain significant, the available evidence shows that smart viticulture can improve resource-use efficiency, support more precise interventions, and help growers respond more effectively to environmental variability. Future progress will depend on stronger validation under field conditions, better integration into practical vineyard workflows, interoperable digital infrastructures, and decision-support tools that are transparent, reliable, and useful for end users.

Graphical Abstract

1. Introduction

Viticulture is increasingly challenged to reconcile environmental sustainability with the production of high-quality grapes under more variable climatic, economic, and labor conditions. Vineyard systems are now confronted with multiple pressures, including water scarcity, heat stress, phytosanitary threats, rising production costs, and growing demands for traceability and reduced environmental impact. In response, considerable attention has been directed toward smart technologies that can support more precise monitoring and more efficient interventions than conventional uniform management practices [1,2].
The relevance of digital tools in viticulture is closely linked to the intrinsic heterogeneity of vineyards. Spatial differences in topography, soil properties, vine vigor, canopy density, water availability, and disease pressure often occur even within the same block, with direct consequences for grape composition, ripening dynamics and final fruit quality. Under these conditions, digitalization offers more than simple automation; it provides a framework for identifying, interpreting, and managing vineyard variability at finer spatial and temporal scales. Smart technologies can therefore help convert heterogeneity from a source of uncertainty into a basis for more informed and site-specific management [3,4,5].
Within this broader transition, artificial intelligence (AI) has emerged as a key enabling component of smart viticulture. Machine learning and deep learning approaches are increasingly used to classify symptoms, assess vine stress, predict yield, support regulated deficit irrigation, analyze canopy imagery, and improve harvest planning. At the same time, geospatial decision-support systems are strengthening the connection between vineyard monitoring and management by integrating environmental, phenological, and operational data into tools that can better support quality-oriented interventions [6]. Robotics and automation are also gaining ground, expanding the range of vineyard tasks that may, in the future, be assisted or partially automated under real field conditions [7].
Despite this progress, the path from technological promise to reliable commercial adoption remains far from straightforward [1]. Many smart viticulture applications still face important constraints, including strong site dependence, limited transferability across seasons and terroirs, data quality issues, high implementation costs, fragmented digital infrastructures, and unresolved questions related to governance, privacy, and long-term usability [8,9,10,11]. Moreover, the literature is not equally mature across all domains: evidence is generally stronger for vineyard monitoring, stress detection, and image-based diagnosis than for broadly transferable predictive tools, quality forecasting, or field-ready autonomous systems.
Against this background, the present review examines smart technologies and AI in viticulture not only as technical innovations, but as components of sustainable growing systems whose practical value depends on their ability to improve resource efficiency, support grape quality, and remain operationally robust and trustworthy under commercial vineyard conditions. While several recent reviews have addressed precision viticulture, digital monitoring, or specific AI applications, the present review aims to provide a broader and more critical synthesis by linking technological applications with sustainability outcomes, adoption barriers, interoperability challenges, and governance-related considerations. In this sense, the manuscript places particular emphasis not only on what smart technologies can do, but also on the conditions under which they can be credibly adopted and integrated into routine vineyard practice.
More specifically, the objectives of this review are: (i) to examine the main technological layers underpinning smart viticulture; (ii) to assess the current maturity of major AI application domains in vineyard management; (iii) to evaluate how these technologies contribute to sustainability and grape quality; and (iv) to identify the practical, organizational, and governance-related constraints that still limit their broader commercial adoption.

Literature Search and Review Approach

This article was developed as a narrative critical review. The literature search focused primarily on peer-reviewed journal articles and selected review papers addressing smart technologies, artificial intelligence, precision viticulture, vineyard monitoring, decision-support systems, sustainability, and adoption barriers in viticulture. Searches were conducted in major scientific databases, including Scopus, Web of Science, and Google Scholar, using combinations of keywords such as “smart viticulture”, “precision viticulture”, “artificial intelligence in viticulture”, “grapevine disease detection”, “vineyard irrigation decision support”, “yield prediction”, “digital agriculture”, “interoperability”, and “data governance”. Priority was given to recent literature, while earlier landmark studies were retained when they were important for conceptual framing or technological context. Representative case studies were included where they helped illustrate practical applications, quantitative outcomes, or operational limitations under field conditions. Studies not directly relevant to viticulture, duplicate reports, and sources lacking sufficient methodological or contextual information were considered with caution or excluded from the main synthesis.
The review was structured to provide an integrative and critical synthesis, with emphasis on thematic coverage, interpretive analysis, and practical relevance to commercial viticulture.

2. Smart Technologies Underpinning the Modern Vineyard

Smart viticulture relies on the interaction of several complementary technological layers rather than on isolated digital tools [1,12]. At the foundation of this system lies in-field sensing [8,13]. Soil moisture probes, weather stations, thermal and humidity sensors, leaf wetness sensors and connected irrigation components allow continuous monitoring of both vineyard conditions and vine status [8,14,15,16]. Such information is particularly valuable in viticulture, where relatively small differences in water deficit, canopy microclimate, or disease-favoring conditions can lead to substantial variation in berry composition, ripening behavior, and final fruit quality [4,15,17].
A second layer is provided by proximal and remote sensing technologies [1,3,8]. UAVs, satellites, thermal cameras, multispectral systems, and image-based scouting platforms are increasingly used to assess vine vigor, identify temperature anomalies, delineate management zones, and detect early signatures of biotic or abiotic stress [1,3,8,12,18,19]. These approaches are especially useful in vineyards targeting high-value or premium-quality production, where spatially explicit information can support more refined decisions on irrigation, canopy management, disease surveillance, and harvest timing [9,19,20]. Their main advantage is that they provide a broader spatial view of vineyard variability, although their practical value still depends on how easily the resulting information can be translated into management decisions.
The integration of these diverse data streams into analytical frameworks is essential for generating agronomically meaningful outputs [1,8,13,21]. Smart vineyard datasets are inherently heterogeneous, often combining time-series sensor records, weather observations, image data, geospatial information, and management history [2,8,13,21]. Artificial intelligence helps transform this complex information into operational knowledge through pattern recognition, prediction, and classification [1,8,21]. Computer vision has become particularly important in disease detection, phenotyping, and yield estimation, while machine learning models are widely used for forecasting, vineyard zoning, and decision support [1,3,8,9,10,12,21,22,23,24,25]. At this level, the challenge is often not the lack of analytical tools, but the difficulty of making their outputs robust, interpretable, and useful under real vineyard conditions.
The final and perhaps most critical layer is operational integration [4,13]. Digital technologies only acquire real agronomic value when they are translated into practical vineyard actions [4]. Decision-support systems, risk alerts, variable-rate recommendations, and automated prescriptions are the mechanisms through which sensing and analytics influence management in the field [4,8,9,13]. Irrigation DSS, for example, can convert continuous environmental observations into regulated deficit irrigation strategies that support both sustainability goals and grape quality [7,14,26,27]. In this sense, the effectiveness of smart viticulture depends less on the mere presence of digital tools than on their ability to connect observation, interpretation, and action within a coherent management framework, as illustrated in the conceptual workflow shown in Figure 1 [1,3,8,13].
Taken together, these technological layers show that the digital transition of viticulture is no longer limited by data acquisition alone [3,9,12]. In many cases, sensing technologies are already relatively advanced, whereas the more persistent challenge lies in the reliable integration of heterogeneous data into robust, interpretable, and practically usable management systems [9,10,22,28]. This distinction is important, because the real bottleneck in smart viticulture is increasingly not whether data can be collected, but whether they can be turned into decisions that remain useful under commercial vineyard conditions [8,29]. Overall, this suggests that the main challenge is no longer technological availability itself, but the consistent delivery of practical value in commercial viticulture.

3. AI Applications Supporting Sustainable and High-Quality Grape Production

One of the most developed application areas of artificial intelligence in viticulture is the monitoring of vine status and vineyard variability [21]. By combining sensor data, weather information, and image-based observations, AI-assisted systems can estimate canopy vigor, thermal stress, water status, and spatial differences in vegetative development [20]. In heterogeneous vineyards, such information can improve the timing and spatial targeting of management decisions, allowing interventions to be adapted more closely to local conditions rather than applied uniformly across the entire block [21]. This is particularly relevant in sustainable viticulture, where the efficient use of resources depends on identifying where and when action is truly needed [8,21]. Among current application domains, this is one of the areas in which AI appears to be most mature, largely because monitoring and spatial diagnosis are more readily translated into practical management support than more complex predictive tasks.
Another important field of application is disease and pest detection [24,30,31]. Deep learning and computer vision approaches are increasingly used to identify symptoms on leaves, shoots, and bunches from RGB, multispectral, and thermal imagery [1,21,31], while sensor-based systems may help detect disease-conducive microclimatic conditions before symptoms become severe [21,26,32]. The value of these technologies lies not only in faster diagnosis, but also in their ability to improve the timing and precision of plant protection [9,26]. In commercial vineyards, this can support more selective interventions, reduce unnecessary applications, and help protect grape integrity and quality [21,26]. At the same time, the evidence is generally stronger for symptom recognition and early warning than for consistently reliable field deployment across changing canopy, lighting, and seasonal conditions. For example, Kerkech et al. reported that a deep learning segmentation approach using UAV multispectral imagery achieved more than 92% disease detection at grapevine level and 87% at leaf level for mildew detection, highlighting the strong technical potential of image-based diagnosis under well-defined conditions [33]. However, such results also underline the continuing challenge of generalization across vineyard settings and image acquisition conditions. Selected case studies therefore suggest that AI-supported disease detection is promising, but still more robust in controlled or narrowly defined operational contexts than in fully variable commercial environments.
AI is also playing an important role in irrigation and nutrient management [34]. By integrating soil moisture data, weather conditions, canopy indicators, and previous vineyard responses, data-driven systems can support more precise irrigation scheduling and, in some cases, variable-rate water application [21,26,35]. This is especially important in viticultural regions where water scarcity and climate variability are already influencing management choices [12,36,37]. Similar approaches may also contribute to more targeted nutrient management, helping growers balance vine performance, fruit quality, and resource efficiency under increasingly constrained conditions [21,26]. In this area, the agronomic rationale is particularly strong, since digital tools can directly support resource-saving interventions. This potential is supported by field evidence from variable-rate irrigation studies. For instance, Ortuani et al. reported an 18% reduction in irrigation water use compared with a reference sector, without losses in yield or product quality, while grape maturation became more homogeneous over time [38]. These results illustrate how precision irrigation can contribute to both resource efficiency and quality-oriented management when adapted to site-specific variability. At the same time, such outcomes remain context-dependent and should not be generalized without considering climatic, edaphic, and infrastructural differences among vineyards.
A further area of rapid development is yield estimation and grape quality prediction [13,26]. Machine learning models are increasingly used to estimate bunch number, crop load, or expected yield from image-based and phenological information [21,29,39], while spectral and thermal signals are being explored as indicators of ripening and quality-related variability [9,21,40]. These applications are particularly valuable in quality-oriented viticulture, where selective harvesting, improved logistics, and a better understanding of within-vineyard variability can contribute directly to more informed production decisions [1,4,26]. Recent studies also show that predictive performance can be useful but remains uneven across tasks and vineyard contexts. In a multi-variety field study, Palacios et al. reported a mean F1-score of 0.72 for berry detection, with yield prediction R2 values ranging from 0.54 to 0.87 and normalized RMSE values between 16.47% and 39.17%, indicating that early yield forecasting can be informative but still cultivar- and context-dependent [41]. For quality-related traits, Pérez-Pérez et al. showed that vineyard proximal multispectral sensing combined with artificial neural networks achieved R2 values of 0.89, 0.81, and 0.83 for total soluble solids, acidity, and pH in Cabernet and Merlot, and 0.93, 0.97, and 0.94, respectively, in Parellada, supporting the potential of AI-assisted quality assessment while also confirming that model performance remains dataset- and cultivar-specific [42]. These findings reinforce the view that yield and grape quality prediction are promising, but still less mature and less transferable than monitoring and diagnostic applications.
Beyond monitoring and prediction, AI is gradually becoming part of vineyard automation [8,12,21]. Robotics and autonomous or semi-autonomous systems are being explored for scouting, spraying, canopy operations, and repetitive support tasks [8,43]. Although these technologies are still limited by terrain complexity, cost, and operational constraints [21], they address important structural challenges in modern viticulture, especially labor shortages and the need for more precise and timely interventions [16,21]. Their future role is therefore likely to expand as vineyard digitalization becomes more mature and better integrated into practical management systems [8,9,21]. However, compared with monitoring and decision-support applications, automation remains less field-ready and more strongly constrained by engineering, economic, and operational factors. In many cases, the literature still reflects proof-of-concept development rather than routine commercial deployment. Quantitative field evidence for vineyard automation remains more limited than for monitoring applications, but available studies do indicate both promise and current constraints. For example, Silwal et al. reported that a fully autonomous robotic pruning system was able to spur-prune vines from both sides in 213 s per vine, with a total pruning accuracy of 87% under field conditions [44]. These results are encouraging, but they also confirm that speed, plant architecture, terrain variability, and operational complexity remain major constraints for broader commercial deployment.
As summarized in Table 1, current AI applications appear most mature where they improve early detection, spatial diagnosis, and targeted management, particularly in vineyard monitoring and disease surveillance [21,31]. By contrast, broader predictive applications—especially those related to grape quality forecasting, generalized yield prediction across sites, and fully field-ready automation—remain more uneven in terms of transferability and operational maturity [21,29]. This distinction is important, because it suggests that the strongest short-term contribution of AI in viticulture may lie less in full automation and more in supporting better-informed, more timely, and more localized decisions [2,8]. Overall, the literature suggests that AI currently adds greatest value when it supports human decision-making in data-rich but operationally complex contexts, whereas fully autonomous or widely transferable systems remain less convincingly established. Table 1 summarizes the main AI application domains in viticulture, together with their typical data sources, practical relevance, indicative quantitative evidence, and current maturity or main limitations.

4. Contributions to Sustainability and Grape Quality

The contribution of smart technologies to viticulture extends well beyond automation or data collection alone [8,19]. Their practical significance lies in improving the precision of vineyard interventions [1]. By supporting more targeted decisions in canopy management, irrigation and plant protection, AI-enabled systems can help reduce unnecessary inputs, improve resource-use efficiency and lower the environmental burden associated with conventional uniform management [9,21,26]. This is one of the clearest ways in which digital viticulture contributes to sustainability [9,26,43]. Evidence from precision viticulture case studies supports this general pattern. For example, Balafoutis et al. reported that the application of precision viticulture practices reduced environmental burden in vineyard systems by lowering inputs and improving management efficiency, as assessed through life cycle analysis [35]. Similarly, Ortuani et al. observed an 18% reduction in irrigation water use under variable-rate management without losses in yield or product quality, illustrating how targeted intervention can improve both sustainability and production outcomes [38].
A second important contribution concerns the management of vineyard variability [15]. In quality-oriented production systems, differences in vigor, stress intensity, ripening dynamics and disease pressure often occur within the same vineyard block [9,15]. Smart technologies help make this variability more visible and, crucially, more manageable by providing maps, alerts and predictive outputs that support site-specific action [1,9,10]. In this way, digital tools can contribute not only to resource savings, but also to a better alignment between management intensity and fruit quality potential [10,15]. This is particularly relevant in vineyards where spatial heterogeneity directly affects selective harvest decisions, ripening uniformity, and the consistency of grape lots destined for differentiated or premium-quality production.
These technologies may also improve harvest organization and the uniformity of grape lots [1,47]. More detailed monitoring of ripening patterns and within-vineyard heterogeneity can support selective harvesting, more efficient scheduling of picking operations and a clearer separation of zones with different quality potential [1,9,12]. This is particularly relevant in vineyards targeting premium or differentiated production, where even moderate improvements in fruit uniformity and harvest timing may have a meaningful effect on quality outcomes and market value [9,26,43]. In this respect, the contribution of smart viticulture is not only agronomic but also logistical, since better spatial knowledge can help align harvest operations with the variability actually present in the vineyard rather than relying on block-wide averages alone.
Timeliness is another major advantage [26]. Many viticultural decisions must be taken within narrow operational windows, particularly under rapidly changing environmental conditions [9]. AI-supported monitoring and predictive systems can reduce the delay between observation and response, which is especially valuable for disease management, irrigation scheduling and harvest planning [9,21]. This becomes even more important when growers are responsible for multiple vineyard blocks and must allocate time and labor efficiently [16]. In practical terms, the value of these systems often lies less in replacing management decisions than in helping producers intervene earlier and with greater spatial precision.
Smart technologies may also contribute to the economic dimension of sustainability [48]. By improving the timing and spatial precision of interventions, they can help reduce unnecessary costs associated with water use, plant protection, repeated field operations and labor allocation [48]. Even when the initial investment in digital tools is considerable [9,20,48], their practical value may increase when they support more efficient decision-making, reduce avoidable losses and improve the consistency of vineyard management over time [4,20,48]. Economic studies on digital vineyard technologies suggest that profitability depends strongly on context, but also indicate that site-specific management and improved operational coordination can enhance performance where adoption is matched to production scale and management needs [20,48]. For this reason, the economic relevance of smart viticulture should be understood not as universally guaranteed, but as conditional on appropriate integration into the technical and organizational realities of each vineyard system.
Another important contribution lies in strengthening vineyard resilience under increasingly variable climatic conditions [12,13,49]. Earlier detection of water stress, heat-related canopy responses or disease-favoring conditions can improve the responsiveness of management and reduce uncertainty in decision-making [9,11]. In this sense, smart viticulture does not simply improve efficiency under stable conditions, but also enhances the adaptive capacity of vineyard systems facing recurrent environmental stress [13,49,50]. This adaptive role is especially important under climate-sensitive growing conditions, where timely responses to stress events may influence both grape composition and the stability of production outcomes across seasons.
Finally, smart technologies can reinforce traceability, documentation and cumulative learning at farm level [10,22]. Digital records of observations, interventions and outcomes make it easier to compare seasons, evaluate management choices and identify recurring patterns [23,47]. Over time, this can support better strategic planning and more resilient vineyard management, particularly in production systems where grape quality, sustainability targets and efficient resource use must be balanced simultaneously [13,26,51]. The integration of sensing, AI and digital traceability tools may also strengthen transparency along the vineyard-to-winery chain, especially where data continuity and interoperability are maintained across platforms [23].
At the same time, the extent of these benefits should not be overstated [20,21]. The evidence for improved efficiency, input reduction and quality enhancement is generally stronger in applications related to monitoring, diagnosis and targeted intervention than in those requiring broad transferability or long-term autonomous operation [8,21]. This suggests that the strongest current contribution of smart viticulture lies in supporting more precise, timely and informed decision-making, rather than in delivering fully autonomous solutions across all vineyard contexts. Overall, its most important role at present is not to replace vineyard management, but to make it more responsive, consistent and effective under increasingly complex production conditions [1].

5. Barriers, Risks and Field-Ready Adoption Challenges

5.1. Data Quality and Reliability Issues

Despite the rapid development of smart viticulture, the transition from promising research outcomes to reliable on-farm use remains uneven. One of the most persistent obstacles concerns data quality, as the reliability of any data-driven insight is contingent on the quality of the input data [52,53,54,55]. For sensor-based systems, this is a multi-faceted challenge; sensor measurements can be affected by calibration drift, unstable connectivity, missing values, or inconsistent maintenance [52,53,56,57,58,59]. The physical durability of sensors in harsh agricultural environments, coupled with issues like packet loss and energy constraints in wireless networks, can further compromise data reliability [60,61]. Many rural vineyards also suffer from inadequate digital infrastructure, where limited high-speed internet impedes the real-time data transmission necessary for timely decision-making [62]. At the same time, recent advances in edge AI devices suggest that part of this limitation may be mitigated by enabling at-source processing, reducing network dependency, and improving the timeliness of decision support under field conditions [63]. Related developments in model compression and lightweight deployment frameworks may also help lower computational and operational costs, although their adoption in viticulture remains at an early stage [64].
Image-based datasets are equally vulnerable to quality issues. Datasets often vary in illumination, canopy structure, phenological stage and symptom expression [5]. Factors such as shadows, occlusion from leaves or wires, and variable lighting can alter image quality, directly affecting the performance of computer vision models for tasks like yield estimation or disease detection [29,39,65]. The accuracy of AI-driven disease recognition, for example, is highly sensitive to image quality and can decline under real-world conditions where symptoms are subtle or partially obscured [66,67,68,69]. This problem is compounded at the data-source level, where inconsistent record-keeping by farmers can lead to a lack of sufficient or high-quality historical data for training and validating models [70]. Data may also be fragmented across different formats, with inconsistent naming conventions that complicate aggregation and analysis [52]. Even when relying on third-party sources like global climate datasets, the information may lack the necessary spatial resolution to be meaningful for a specific, topographically complex vineyard, failing to capture critical microclimatic variability [13]. Under such conditions, model performance can decline quickly, especially when systems are expected to operate beyond the controlled settings in which they were originally developed [67]. In practical terms, this means that even technically advanced tools may generate outputs that growers find difficult to trust if the underlying data are not stable, representative and consistently collected [71,72]. For this reason, data quality is not just a technical issue, but a central condition for credibility and adoption.

5.2. Limited Transferability Across Vineyard Contexts

A second major challenge hindering the widespread adoption of AI in viticulture is the limited transferability of models across diverse vineyards, growing seasons and production contexts [21,26,29]. Viticulture is inherently complex, defined by unique interactions between local terroir, grapevine cultivars, rootstocks and canopy architecture [2,29]. Furthermore, significant year-to-year climatic variability means that conditions remain fluid, creating an environment where a model performing optimally in one site or season may not necessarily maintain the same level of reliability elsewhere [8,45,46].
Currently, a substantial portion of AI-driven research, particularly in yield prediction and disease detection, relies on models developed under specific environmental or site-specific conditions [21,26,73]. Research indicates that these models often struggle to maintain reliability when deployed in new regions because they lack the necessary generalisation capabilities [5,74,75,76]. For instance, neural networks trained on a specific grape variety or canopy structure frequently exhibit performance degradation when applied to different cultivars or microclimates, as the underlying biological signals are misinterpreted [21,41].
This issue of limited robustness constitutes a primary barrier to wider commercialization. Vineyard managers are understandably reluctant to depend on diagnostic or predictive systems that demand frequent, resource-intensive recalibration or fail to deliver consistent results under local conditions [26,48,49,77]. The long-term value of AI in viticulture is therefore contingent not merely on achieving high accuracy in controlled trials, but on ensuring robustness across the highly heterogeneous real-world environments that define modern grape production [21,26,29,48]. Overcoming this will require developing adaptive frameworks and standardized data protocols capable of accounting for regional variability, thereby bridging the gap between niche research and operational scalability [13,21,26,28,78]. At the same time, this need for standardization should not be understood as an attempt to homogenize viticulture itself, but rather as an effort to improve data comparability and model transparency across inherently diverse terroir conditions. In practice, this remains one of the clearest dividing lines between promising prototypes and systems that are genuinely ready for broader commercial use.

5.3. Economic and Organizational Constraints

Economic and organizational barriers are equally important [79]. The adoption of smart technologies often requires significant investment in hardware, software, subscriptions, data services, maintenance and user training [8,10]. Prohibitive initial capital outlays—ranging from €20,000 to €50,000 per hectare for integrated systems—represent a major obstacle to adoption, particularly for small- and medium-sized vineyards, where economic returns may remain uncertain or indirect [20,21,49]. Because such figures may vary substantially depending on region, vineyard scale, level of technological integration, and labor costs, they should be interpreted as indicative rather than universally representative. This financial risk is compounded by persistent uncertainty regarding the tangible benefits of digitalization, leading many producers to question whether the investment will translate into recognized value for the consumer [49]. Although a full return-on-investment analysis falls beyond the scope of the present review, the literature suggests that economic viability is strongly context-dependent and more favorable where digital tools are aligned with specific management bottlenecks, labor constraints, or quality-oriented production goals.
At the same time, digital tools can add management complexity if they are not well integrated into routine practice [49]. Interoperability gaps between proprietary platforms and legacy equipment create fragmented data ecosystems that complicate integration and increase the workload for farmers [10,21]. Even useful systems may remain underutilized when they require too much time, technical expertise, or interpretive effort from growers and advisors [49]. This issue is amplified by a widespread lack of digital literacy and the skills required to interpret complex data, which can lead to data overload and “technostress” [49]. In this sense, the issue is not only whether a technology works, but whether it fits the practical and cultural realities of vineyard management, where many growers remain skeptical or prioritize traditional methods to preserve their winemaking heritage [49]. This means that adoption depends as much on organizational fit and perceived usefulness as on technical performance alone.

5.4. Interoperability, Governance and Ethical Considerations

Another important limitation concerns interoperability. Many digital agriculture tools operate within closed or semi-closed platforms that do not communicate easily with other systems [10,62]. This can create fragmented workflows, duplicated data handling and long-term dependence on specific providers [10,80]. Closely related to this is the problem of vendor lock-in, where growers may become tied to a particular digital ecosystem because their data, devices or decision-support tools cannot be easily transferred elsewhere [81,82,83]. Such constraints can discourage adoption, especially when producers remain uncertain about the long-term flexibility, ownership or usability of the data they generate [84]. Reviews on agricultural data sharing have repeatedly shown that trust, transparency and clear governance arrangements are central to digital uptake [55,85,86]. Recent work on digital integration in commercial vineyard workflows further suggests that interoperability is not simply a technical convenience, but a prerequisite for practical usability and long-term adoption, especially where multiple platforms, advisory actors, and management decisions must interact coherently across the production chain.
Beyond these technical and operational issues, the broader deployment of AI in viticulture also raises governance and ethical concerns [34,87,88]. Smart vineyard systems increasingly rely on continuous data collection, integration across digital platforms and algorithm-supported decision-making [14,22,23]. Under these conditions, questions related to data ownership, access rights, privacy and responsibility become more important, particularly when operational data may reveal sensitive information about vineyard performance, management practices or commercial strategy [81,89]. These concerns can influence not only legal compliance, but also the willingness of growers to adopt and trust digital tools at farm level [84,86].
Ethical considerations are equally relevant when AI systems begin to influence management decisions more directly [77,89]. If recommendations are difficult to interpret, insufficiently transparent or poorly aligned with local agronomic realities, growers may perceive them as unreliable or difficult to integrate into practice [77,90]. Trust in smart viticulture therefore depends not only on predictive performance, but also on transparency, explainability and clear accountability regarding how digital outputs are generated and used [89,91]. As digital agriculture becomes more interconnected, broader regulatory developments related to AI and data governance may also become increasingly relevant [34,62]. In viticulture, however, their practical significance will depend on whether they support secure, transparent and grower-oriented deployment rather than adding further complexity to adoption [26,49]. From a stakeholder perspective, this suggests that future digital strategies in viticulture should not focus only on technical innovation, but also on practical measures such as open data standards, clearer contractual rules on data ownership and access, advisory support for technology interpretation, and governance arrangements that protect user trust while preserving flexibility in technology choice.
Taken together, these barriers suggest that the future of smart viticulture depends on much more than technical innovation alone [49]. Successful field-ready adoption requires robust data, transferable models, affordable and user-friendly tools, interoperable digital environments and governance frameworks that support transparency and trust [1,49]. Without these conditions, even highly promising technologies may remain confined to pilot studies or niche applications rather than becoming part of routine commercial vineyard management [8,21]. Overall, the literature points to a clear conclusion: the main obstacles to adoption are no longer purely technological, but increasingly practical, organizational and governance-related. Table 2 summarizes the main barriers limiting broader adoption of smart viticulture and highlights their typical manifestations, practical implications, and priority responses.

6. Toward Trustworthy and Scalable Smart Viticulture

For smart viticulture to move beyond promising pilot applications, greater emphasis must be placed on field readiness and practical reliability [13]. Many current studies report encouraging results, but these often come from relatively controlled conditions that do not fully reflect the complexity of commercial vineyards [21,29]. In practice, growers need tools that remain reliable across seasons, sites and management systems, not only under ideal testing scenarios [5,21]. Stronger validation frameworks, clearer reporting of model performance and broader testing with external and multi-site datasets are therefore essential [5,13]. In viticulture, where local conditions strongly influence vine behavior and management outcomes, robustness is often more valuable than isolated gains in model accuracy [68]. This is particularly important in a sector where management decisions must remain both timely and context-specific.
Scalability also depends on whether these technologies are genuinely usable in everyday vineyard management [1,16]. Growers and advisors do not need sophisticated algorithms for their own sake; they need tools that fit within real production routines and support better decisions at the right time [20]. Recommendations must therefore be understandable, relevant and easy to act upon [92]. If digital outputs are too opaque, too technical or disconnected from practical vineyard realities, even accurate systems may remain underused [49,81,92]. Transparency, intuitive interfaces and clear communication of uncertainty are therefore important conditions for wider uptake [26]. In this sense, practical usability is not a secondary design issue, but a central requirement for adoption.
Another major requirement is the development of open and interoperable digital systems [8,21]. Vineyard data lose much of their long-term value when they remain trapped within isolated or proprietary platforms that do not communicate well with one another [10,13,21]. Better interoperability would make digital workflows more efficient, reduce dependence on single providers and allow growers to adopt technologies more flexibly according to their own needs [28]. Without this level of openness, even technically effective systems may remain difficult to scale in practice.
At the same time, the future of smart viticulture depends on people as much as on technology [48]. Growers, consultants and technical staff need the knowledge and confidence to interpret digital outputs, judge their relevance and use them effectively in practice [5,48]. Training, technical support and collaboration between researchers, developers and end users are therefore just as important as algorithmic improvement [5,48,79]. In this sense, AI in viticulture should be understood as part of a broader socio-technical transition rather than as a stand-alone digital solution [68]. This broader perspective is essential if digital tools are to become embedded in routine vineyard management rather than remain confined to specialist or pilot use.
Finally, trust will be essential for long-term adoption [1]. As digital systems become more embedded in vineyard management, issues such as accountability, transparency, data rights and the acceptable degree of automation will become increasingly important [68]. The long-term success of smart viticulture will depend not only on what the technology can do, but also on whether users perceive it as fair, understandable and reliable within the realities of commercial production [68,91]. Overall, scalable adoption will depend on aligning technological development with agronomic relevance, practical usability, and grower confidence. Taken together, these conditions show that wider adoption in viticulture depends on balancing the potential benefits of digital technologies with the practical, organizational, and governance-related constraints summarized in Figure 2.

7. Conclusions

Smart technologies and artificial intelligence are becoming increasingly important in viticulture because they offer new ways to monitor vineyard conditions, interpret variability and support more precise management decisions. Their potential is especially relevant in a sector where grape quality, environmental sustainability and production efficiency must increasingly be addressed at the same time. Across the literature, some of the most promising contributions relate to plant stress monitoring, disease surveillance, irrigation and nutrient management, spatially targeted interventions and more informed harvest planning. These are also the areas in which digital tools appear to offer the clearest short-term practical value for commercial viticulture.
At the same time, the wider adoption of these technologies is not simply a matter of improving model accuracy. Their long-term value will depend on whether they can perform reliably under real vineyard conditions, remain useful across different production contexts and fit within the everyday realities of commercial management. Data quality, transferability, cost, interoperability and user trust remain central challenges that still limit broader uptake. In many cases, these factors are more decisive for adoption than technical performance alone.
Future research should therefore move beyond proof-of-concept development and place greater emphasis on multi-site validation, cross-season robustness, interoperability among digital platforms, and the practical usability of decision-support systems under commercial vineyard conditions. More work is also needed to clarify the economic viability of smart viticulture for different production scales and to strengthen governance frameworks around data ownership, transparency, and responsible AI deployment. These priorities are particularly important if smart technologies are to move from promising innovations to credible tools for routine vineyard management.
Taken together, the available evidence suggests that the future of smart viticulture will depend on more than technical innovation alone. Meaningful progress will require robust validation, practical usability, stronger integration into vineyard workflows, and digital systems that remain transparent, flexible, and trustworthy. Under these conditions, smart technologies could play an important role in supporting more sustainable viticulture while helping growers maintain or improve grape quality in increasingly demanding production environments. Overall, their greatest contribution is unlikely to lie in replacing vineyard management, but in making it more precise, responsive, and informed under real commercial conditions.

Author Contributions

All authors contributed equally to the conceptualization, literature review, writing, revision, and final approval of the manuscript. All authors have read and agreed to the published version of the manuscript.

Funding

This publication is financed by the Project “Strengthening and optimizing the operation of MODY services and academic and research units of the Hellenic Mediterranean University”, funded by the Public Investment Program of the Greek Ministry of Education and Religious Affairs.

Data Availability Statement

No new data were created or analyzed in this study. Data sharing is not applicable to this article.

Conflicts of Interest

The authors declare no conflicts of interest.

Abbreviations

The following abbreviations are used in this manuscript:
AIArtificial Intelligence
ANNArtificial Neural Network
CNNConvolutional Neural Network
DSSDecision-Support System
IoTInternet of Things
MLMachine Learning
NRMSENormalized Root Mean Square Error
R2Coefficient of Determination
RGBRed–Green–Blue
RMSERoot Mean Square Error
ROIReturn on Investment
UAVUnmanned Aerial Vehicle

References

  1. Tardaguila, J.; Stoll, M.; Gutiérrez, S.; Proffitt, T.; Diago, M.P. Smart applications and digital technologies in viticulture: A review. Smart Agric. Technol. 2021, 1, 100005. [Google Scholar] [CrossRef]
  2. Newlands, N.K. Artificial Intelligence and Big Data Analytics in Vineyards: A Review; IntechOpen: London, UK, 2021. [Google Scholar]
  3. Ferro, M.V.; Catania, P. Technologies and innovative methods for precision viticulture: A comprehensive review. Horticulturae 2023, 9, 399. [Google Scholar] [CrossRef]
  4. Ammoniaci, M.; Kartsiotis, S.-P.; Perria, R.; Storchi, P. State of the art of monitoring technologies and data processing for precision viticulture. Agriculture 2021, 11, 201. [Google Scholar] [CrossRef]
  5. Mucalo, A.; Matić, D.; Morić-Španić, A.; Čagalj, M. Satellite solutions for precision viticulture: Enhancing sustainability and efficiency in vineyard management. Agronomy 2024, 14, 1862. [Google Scholar] [CrossRef]
  6. Terribile, F.; Bonfante, A.; D’Antonio, A.; De Mascellis, R.; De Michele, C.; Langella, G.; Manna, P.; Mileti, F.A.; Vingiani, S.; Basile, A. A geospatial decision support system for supporting quality viticulture at the landscape scale. Comput. Electron. Agric. 2017, 140, 88–102. [Google Scholar] [CrossRef]
  7. Tziolas, E.; Karapatzak, E.; Kalathas, I.; Karampatea, A.; Grigoropoulos, A.; Bajoub, A.; Pachidis, T.; Kaburlasos, V.G. Assessing the economic performance of multipurpose collaborative robots toward skillful and sustainable viticultural practices. Sustainability 2023, 15, 3866. [Google Scholar] [CrossRef]
  8. Passias, A.; Tsakalos, K.-A.; Kleitsiotis, G.; Tsipas, E.; Rallis, K.; Fyrigos, I.-A.; Pantazi, X.E.; Sirakoulis, G.C. Recent Advances in Precision Viticulture: A Review. IEEE Trans. AgriFood Electron. 2025, 3, 308–319. [Google Scholar] [CrossRef]
  9. Yang, S.; Feng, X.; Huang, D. Case Study on Precision Viticulture: Implementing Technology for High Yields and Sustainability. Tree Genet. Mol. Breed. 2024, 14, 304–312. [Google Scholar] [CrossRef]
  10. Bastard, A.; Chaillet, A. Digitalization from vine to wine: Successes and remaining challenges—A review. In Proceedings of the BIO Web of Conferences; EDP Sciences: Les Ulis, France, 2023; p. 01034. [Google Scholar]
  11. Fuentes, S.; Tongsonand, E.; Gonzalez Viejo, C. How artificial intelligence (AI) is helping winegrowers to deal with adversity from climate change. In Proceedings of the IVES Conference Series; Enoforum: Zaragoza, Spain, 2021. [Google Scholar]
  12. Sapaev, J.; Fayziev, J.; Sapaev, I.; Abdullaev, D.; Nazaraliev, D.; Sapaev, B. Viticulture and wine production: Challenges, opportunities and possible implications. In Proceedings of the E3S Web of Conferences; EDP Sciences: Les Ulis, France, 2023; p. 01037. [Google Scholar]
  13. Simeunović, M.; Ratković, K.; Kovač, N.; Racković, T.; Fernandes, A. A Knowledge-Driven Framework for a Decision Support Platform in Sustainable Viticulture: Integrating Climate Data and Supporting Stakeholder Collaboration. Sustainability 2025, 17, 1387. [Google Scholar] [CrossRef]
  14. Pascoal, D.; Silva, N.; Adao, T.; Lopes, R.D.; Peres, E.; Morais, R. A technical survey on practical applications and guidelines for IoT sensors in precision agriculture and viticulture. Sci. Rep. 2024, 14, 29793. [Google Scholar] [CrossRef]
  15. Ozdemir, G.; Sessiz, A.; Pekitkan, F.G. Precision Viticulture tools to production of high quality grapes. Sci. Pap. Ser. B Hortic. 2017, 61. [Google Scholar]
  16. Rivera Chavez, Z.B.; Porcaro, A.; De Simone, M.C.; Guida, D. Improving Sustainable Viticulture in Developing Countries: A Case Study. Sustainability 2025, 17, 5338. [Google Scholar] [CrossRef]
  17. Fuentes, S.; Gago, J. Modern approaches to precision and digital viticulture. In Improving Sustainable Viticulture and Winemaking Practices; Elsevier: Amsterdam, The Netherlands, 2022; pp. 125–145. [Google Scholar]
  18. Sassu, A.; Gambella, F.; Ghiani, L.; Mercenaro, L.; Caria, M.; Pazzona, A.L. Advances in Unmanned Aerial System Remote Sensing for Precision Viticulture. Sensors 2021, 21, 956. [Google Scholar] [CrossRef] [PubMed]
  19. Matese, A.; Filippo Di Gennaro, S. Technology in precision viticulture: A state of the art review. Int. J. Wine Res. 2015, 2015, 69–81. [Google Scholar] [CrossRef]
  20. Sofia, S.; Agosta, M.; Asciuto, A.; Crescimanno, M.; Galati, A. Unleashing profitability of vineyards through the adoption of unmanned aerial vehicles technology systems: The case of two Italian wineries. Precis. Agric. 2025, 26, 41. [Google Scholar] [CrossRef]
  21. Moganapathi, B.; Kavitha, P.; Muthuvel, I.; Senthil, A.; Balachandar, D.; Mahendran, P. The AI-viticulture nexus: Robotics and precision technologies for sustainable vineyards. Plant Sci. Today 2025, 12, 1–11. [Google Scholar] [CrossRef]
  22. Gebresenbet, G.; Bosona, T.; Patterson, D.; Persson, H.; Fischer, B.; Mandaluniz, N.; Chirici, G.; Zacepins, A.; Komasilovs, V.; Pitulac, T. A concept for application of integrated digital technologies to enhance future smart agricultural systems. Smart Agric. Technol. 2023, 5, 100255. [Google Scholar] [CrossRef]
  23. Adamashvili, N.; Zhizhilashvili, N.; Tricase, C. The integration of the internet of things, artificial intelligence, and blockchain technology for advancing the wine supply chain. Computers 2024, 13, 72. [Google Scholar] [CrossRef]
  24. Seng, K.P.; Ang, L.-M.; Schmidtke, L.M.; Rogiers, S.Y. Computer vision and machine learning for viticulture technology. IEEE Access 2018, 6, 67494–67510. [Google Scholar] [CrossRef]
  25. Poblete-Echeverría, C.; Hernández, I.; Gutiérrez, S.; Iñiguez, R.; Barrio, I.; Tardaguila, J. Using artificial intelligence (AI) for grapevine disease detection based on images. In Proceedings of the BIO Web of Conferences; EDP Sciences: Les Ulis, France, 2023; p. 01021. [Google Scholar]
  26. Izquierdo-Bueno, I.; Moraga, J.; Cantoral, J.M.; Carbú, M.; Garrido, C.; González-Rodríguez, V.E. Smart viniculture: Applying artificial intelligence for improved winemaking and risk management. Appl. Sci. 2024, 14, 10277. [Google Scholar] [CrossRef]
  27. Kang, C.; Diverres, G.; Karkee, M.; Zhang, Q.; Keller, M. Decision-support system for precision regulated deficit irrigation management for wine grapes. Comput. Electron. Agric. 2023, 208, 107777. [Google Scholar] [CrossRef]
  28. Adam-Blondon, A.F.; Alaux, M.; Pommier, C.; Cantu, D.; Cheng, Z.M.; Cramer, G.R.; Davies, C.; Delrot, S.; Deluc, L.; Di Gaspero, G.; et al. Towards an open grapevine information system. Hortic. Res. 2016, 3, 16056. [Google Scholar] [CrossRef]
  29. Barriguinha, A.; de Castro Neto, M.; Gil, A. Vineyard yield estimation, prediction, and forecasting: A systematic literature review. Agronomy 2021, 11, 1789. [Google Scholar] [CrossRef]
  30. Oliveira, R.C.d.; Silva, R.D.d.S.e. Artificial intelligence in agriculture: Benefits, challenges, and trends. Appl. Sci. 2023, 13, 7405. [Google Scholar] [CrossRef]
  31. Gatou, P.; Tsiara, X.; Spitalas, A.; Sioutas, S.; Vonitsanos, G. Artificial Intelligence Techniques in Grapevine Research: A Comparative Study with an Extensive Review of Datasets, Diseases, and Techniques Evaluation. Sensors 2024, 24, 6211. [Google Scholar] [CrossRef]
  32. Pero, C.; Bakshi, S.; Nappi, M.; Tortora, G. IoT-driven machine learning for precision viticulture optimization. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 2023, 17, 2437–2447. [Google Scholar] [CrossRef]
  33. Kerkech, M.; Hafiane, A.; Canals, R. Vine disease detection in UAV multispectral images using optimized image registration and deep learning segmentation approach. Comput. Electron. Agric. 2020, 174, 105446. [Google Scholar] [CrossRef]
  34. Younas, M.; Akhtar, M.N.; Batool, S.; Owais, M.; Sahar, S.; Anum, W. The Integration of Artificial Intelligence in Agriculture: Emerging Trends, Benefits and Challenges. J. Asian Dev. Stud. 2025, 14, 1316–1333. [Google Scholar] [CrossRef]
  35. Balafoutis, A.T.; Koundouras, S.; Anastasiou, E.; Fountas, S.; Arvanitis, K. Life cycle assessment of two vineyards after the application of precision viticulture techniques: A case study. Sustainability 2017, 9, 1997. [Google Scholar] [CrossRef]
  36. Mirás-Avalos, J.M.; Araujo, E.S. Optimization of vineyard water management: Challenges, strategies, and perspectives. Water 2021, 13, 746. [Google Scholar] [CrossRef]
  37. Finco, A.; Bentivoglio, D.; Chiaraluce, G.; Alberi, M.; Chiarelli, E.; Maino, A.; Mantovani, F.; Montuschi, M.; Raptis, K.G.C.; Semenza, F.; et al. Combining precision viticulture technologies and economic indices to sustainable water use management. Water 2022, 14, 1493. [Google Scholar] [CrossRef]
  38. Ortuani, B.; Facchi, A.; Mayer, A.; Bianchi, D.; Bianchi, A.; Brancadoro, L. Assessing the effectiveness of variable-rate drip irrigation on water use efficiency in a Vineyard in Northern Italy. Water 2019, 11, 1964. [Google Scholar] [CrossRef]
  39. Mohimont, L.; Alin, F.; Rondeau, M.; Gaveau, N.; Steffenel, L.A. Computer vision and deep learning for precision viticulture. Agronomy 2022, 12, 2463. [Google Scholar] [CrossRef]
  40. Passias, A.; Kleitsiotis, G.; Tompris, I.; Stavroulakis, E.; Tsipas, E.; Tsakalos, K.-A.; Rallis, K.; Fyrigos, I.-A.; Pantazi, X.E.; Sirakoulis, G.C. A holistic approach to grapevine cultivation with precision viticulture. In Proceedings of the 2024 IEEE 2nd Conference on AgriFood Electronics (CAFE); IEEE: New York, NY, USA, 2024; pp. 1–5. [Google Scholar]
  41. Palacios, F.; Diago, M.P.; Melo-Pinto, P.; Tardaguila, J. Early yield prediction in different grapevine varieties using computer vision and machine learning. Precis. Agric. 2023, 24, 407–435. [Google Scholar] [CrossRef]
  42. Pérez-Pérez, C.A.; Viejo, C.G.; Fuentes, S.; Valiente-Banuet, J.I. Vineyard proximal sensing using multispectral imaging to evaluate grape ripening and quality traits using artificial neural networks modeling. J. Agric. Food Res. 2025, 23, 102252. [Google Scholar] [CrossRef]
  43. Mărculescu, S.-I.; Badea, A.; Teodorescu, R.I.; Begea, M.; Frîncu, M.; Bărbulescu, I.D. Application of artificial intelligence technologies in viticulture. Sci. Pap. Ser. Manag. Econ. Eng. Agric. Rural. Dev. 2024, 24. [Google Scholar]
  44. Silwal, A.; Yandun, F.; Nellithimaru, A.; Bates, T.; Kantor, G. Bumblebee: A Path Towards Fully Autonomous Robotic Vine Pruning. Field Robot. 2022, 2, 1661–1696. [Google Scholar] [CrossRef]
  45. Laroche-Pinel, E.; Cianciola, V.; Singh, K.; Vivaldi, G.A.; Brillante, L. Assessing the spatial-temporal performance of machine learning in predicting grapevine water status from Landsat 8 imagery via block-out and date-out cross-validation. Agric. Water Manag. 2024, 306, 109163. [Google Scholar] [CrossRef]
  46. Brillante, L.; Mathieu, O.; Leveque, J.; Bois, B. Ecophysiological Modeling of Grapevine Water Stress in Burgundy Terroirs by a Machine-Learning Approach. Front Plant Sci. 2016, 7, 796. [Google Scholar] [CrossRef]
  47. MacPherson, J.; Voglhuber-Slavinsky, A.; Olbrisch, M.; Schobel, P.; Donitz, E.; Mouratiadou, I.; Helming, K. Future agricultural systems and the role of digitalization for achieving sustainability goals. A review. Agron. Sustain. Dev. 2022, 42, 70. [Google Scholar] [CrossRef]
  48. Galati, A.; Sofia, S.; Crescimanno, M. Economics and barriers of precision viticulture technologies: A comprehensive systematic literature review. Inf. Process. Agric. 2025, 12, 487–500. [Google Scholar] [CrossRef]
  49. Panero, M.; De Bernardi, P.; Moggi, S.; Pierce, P. Unveiling digitalisation in Italian viticulture: A field study on drivers and barriers. Br. Food J. 2025, 127, 500–516. [Google Scholar] [CrossRef]
  50. Kudryashova, E.; Casetti, M. The internet of things-the nearest future of viticulture. AGRIS-Line Pap. Econ. Inform. 2021, 13, 79–86. [Google Scholar] [CrossRef]
  51. Seyedzadeh, Z.; Jabalameli, M.S.; Dehghani, E. Towards a sustainable viticultural supply chain under uncertainty: Integration of data envelopment analysis, artificial neural networks, and a multi-objective optimization model. Sci. Total Environ. 2025, 970, 178980. [Google Scholar] [CrossRef] [PubMed]
  52. Punt, T.; Snijkers, G. Exploring precision farming data: A valuable new data source? A first orientation. In Proceedings of the UNECE Workshop on Statistical Data Collection’New Sources and New Technologies, Geneva, Switzerland, 14–16 October 2019; pp. 14–16. [Google Scholar]
  53. Mansouri, T.; Sadeghi Moghadam, M.R.; Monshizadeh, F.; Zareravasan, A. IoT data quality issues and potential solutions: A literature review. Comput. J. 2023, 66, 615–625. [Google Scholar] [CrossRef]
  54. Abdipourchenarestansofla, M.; Schroth, C. The Importance of Data Quality Assessment for Machinery Data in the Field of Agriculture; VDI-Berichte; VDI Wissensforum GmbH: Düsseldorf, Germany, 2022; p. 2395. [Google Scholar] [CrossRef]
  55. Šestak, M.; Copot, D. Towards trusted data sharing and exchange in agro-food supply chains: Design principles for agricultural data spaces. Sustainability 2023, 15, 13746. [Google Scholar] [CrossRef]
  56. Soussi, A.; Zero, E.; Sacile, R.; Trinchero, D.; Fossa, M. Smart Sensors and Smart Data for Precision Agriculture: A Review. Sensors 2024, 24, 2647. [Google Scholar] [CrossRef]
  57. Horstmann, L.P.; Wagner, M.; Scheffel, R.M.; Fröhlich, A.A. Dealing with incomplete datasets with a confidence attribution algorithm. Measurement 2022, 199, 111509. [Google Scholar] [CrossRef]
  58. Haseeb, K.; Ud Din, I.; Almogren, A.; Islam, N. An Energy Efficient and Secure IoT-Based WSN Framework: An Application to Smart Agriculture. Sensors 2020, 20, 2081. [Google Scholar] [CrossRef]
  59. Alahe, M.A.; Wei, L.; Chang, Y.; Gummi, S.R.; Kemeshi, J.; Yang, X.; Won, K.; Sher, M. Cyber security in smart agriculture: Threat types, current status, and future trends. Comput. Electron. Agric. 2024, 226, 109401. [Google Scholar] [CrossRef]
  60. Catelani, M.; Ciani, L.; Bartolini, A.; Del Rio, C.; Guidi, G.; Patrizi, G. Reliability Analysis of Wireless Sensor Network for Smart Farming Applications. Sensors 2021, 21, 7683. [Google Scholar] [CrossRef]
  61. Patrizi, G.; Bartolini, A.; Ciani, L.; Catelani, M. System reliability of fault-tolerant Wireless Sensor Network for precision agriculture. In Proceedings of the 2024 IEEE International Symposium on Systems Engineering (ISSE), Perugia, Italy, 16–18 October 2024; pp. 1–6. [Google Scholar]
  62. Kandel, G.P.; Poláková, J.; Hamouz, P.; Hruška, A.; Varvaris, I.; Manikas, I. The role of digitalization in supporting farmers and strategic policies for food security and sustainability in Europe: A review. Sustain. Futur. 2025, 11, 101702. [Google Scholar] [CrossRef]
  63. Gong, R.; Zhang, H.; Li, G.; He, J. Edge Computing-Enabled Smart Agriculture: Technical Architectures, Practical Evolution, and Bottleneck Breakthroughs. Sensors 2025, 25, 5302. [Google Scholar] [CrossRef]
  64. Wang, T.; Guo, J.; Zhang, B.; Yang, G.; Li, D. Deploying AI on Edge: Advancement and Challenges in Edge Intelligence. Mathematics 2025, 13, 1878. [Google Scholar] [CrossRef]
  65. Íñiguez, R.; Wolela, F.; Gonzalez Pavez, M.I.; Barrio Fernández, I.; Tardáguila, J.; Venter, T.; Poblete-Echeverría, C. Artificial intelligence-driven classification method of grapevine major phenological stages using conventional RGB imaging. OENO One 2025, 59, 9306. [Google Scholar] [CrossRef]
  66. Mohanty, S.P.; Hughes, D.P.; Salathe, M. Using Deep Learning for Image-Based Plant Disease Detection. Front. Plant Sci. 2016, 7, 1419. [Google Scholar] [CrossRef]
  67. Khan, K.H.; Aljaedi, A.; Ishtiaq, M.S.; Imam, H.; Bassfar, Z.; Jamal, S.S. Disease detection in grape cultivation using strategically placed cameras and machine learning algorithms with a focus on powdery mildew and blotches. IEEE Access 2024, 12, 139505–139523. [Google Scholar] [CrossRef]
  68. Cividino, S.R.S.; Cappelli, A.; Belluco, P.; Rinaldi, F.; Avramovic, L.; Zaninelli, M. Preliminary Study of Vision-Based Artificial Intelligence Application to Evaluate Occupational Risks in Viticulture. Sensors 2025, 25, 6749. [Google Scholar] [CrossRef]
  69. Íñiguez, R.; Palacios, F.; Barrio, I.; Hernández, I.; Gutiérrez, S.; Tardaguila, J. Impact of leaf occlusions on yield assessment by computer vision in commercial vineyards. Agronomy 2021, 11, 1003. [Google Scholar] [CrossRef]
  70. Mark, R. Ethics of using AI and big data in agriculture: The case of a large agriculture multinational. ORBIT J. 2019, 2, 1–27. [Google Scholar] [CrossRef]
  71. Guth, F.A.; Ward, S.; McDonnell, K. From lab to field: An empirical study on the generalization of convolutional neural networks towards crop disease detection. Eur. J. Eng. Technol. Res. 2023, 8, 33–40. [Google Scholar] [CrossRef]
  72. Fenu, G.; Malloci, F.M. Evaluating impacts between laboratory and field-collected datasets for plant disease classification. Agronomy 2022, 12, 2359. [Google Scholar] [CrossRef]
  73. Adão, T.; Chojka, A.; Pascoal, D.; Silva, N.; Morais, R.; Peres, E. Synthetic Data-Driven Methods to Accelerate the Deployment of Deep Learning Models: A Case Study on Pest and Disease Detection in Precision Viticulture. Computers 2025, 14, 327. [Google Scholar] [CrossRef]
  74. Zhang, J.; Guan, K.; Chen, Z.; Huang, Y.; Zhao, K.; Peng, B.; Wang, S.; Wu, X.; Wang, S.; Banerjee, A. Transfer learning for improved crop yield predictions in a cross-scale pathway: A case study for Brazilian national soybean. Int. J. Appl. Earth Obs. Geoinf. 2025, 145, 104981. [Google Scholar] [CrossRef]
  75. Guo, G.; Zhou, F.; Wu, Q.; Zhai, D. Multi-granularity alignment for crop diseases detection. Plant Phenomics 2025, 7, 100001. [Google Scholar] [CrossRef]
  76. Wang, H.; Ye, Z.; Yao, Y.; Chang, W.; Liu, J.; Zhao, Y.; Li, S.; Liu, Z.; Zhang, X. Improving cross-regional model transfer performance in crop classification by crop time series correction. Geo-Spat. Inf. Sci. 2025, 28, 1581–1596. [Google Scholar] [CrossRef]
  77. Alexander, C.S.; Yarborough, M.; Smith, A. Who is responsible for ‘responsible AI’?: Navigating challenges to build trust in AI agriculture and food system technology. Precis. Agric. 2024, 25, 146–185. [Google Scholar] [CrossRef]
  78. López-Morales, J.A.; Martínez, J.A.; Skarmeta, A.F. Digital transformation of agriculture through the use of an interoperable platform. Sensors 2020, 20, 1153. [Google Scholar] [CrossRef]
  79. Biondo, A.; Rizzo, G.; Migliore, G.; Galati, A. Wine growers’ propensity to adopt digital precision farming technologies: Integrating risk attitudes to the Technology Acceptance Model. Res. Glob. 2025, 11, 100298. [Google Scholar] [CrossRef]
  80. Ahoa, E.; Kassahun, A.; Verdouw, C.; Tekinerdogan, B. Challenges and Solution Directions for the Integration of Smart Information Systems in the Agri-Food Sector. Sensors 2025, 25, 2362. [Google Scholar] [CrossRef]
  81. Rotz, S.; Duncan, E.; Small, M.; Botschner, J.; Dara, R.; Mosby, I.; Reed, M.; Fraser, E.D. The politics of digital agricultural technologies: A preliminary review. Sociol. Rural. 2019, 59, 203–229. [Google Scholar] [CrossRef]
  82. Hackfort, S. Patterns of inequalities in digital agriculture: A systematic literature review. Sustainability 2021, 13, 12345. [Google Scholar] [CrossRef]
  83. Atik, C.; Martens, B. Competition problems and governance of non-personal agricultural machine data: Comparing voluntary initiatives in the US and EU. J. Intell. Prop. Info. Tech. Elec. Com. L. 2021, 12, 370. [Google Scholar]
  84. Jouanjean, M.-A.; Casalini, F.; Wiseman, L.; Gray, E. Issues Around Data Governance in the Digital Transformation of Agriculture: The Farmers’ Perspective; OECD Food, Agriculture and Fisheries Papers; OECD Publishing: Paris, France, 2020; Volume 146. [Google Scholar] [CrossRef]
  85. Sullivan, C.S.; Gemtou, M.; Anastasiou, E.; Fountas, S. Building trust: A systematic review of the drivers and barriers of agricultural data sharing. Smart Agric. Technol. 2024, 8, 100477. [Google Scholar] [CrossRef]
  86. Wiseman, L.; Sanderson, J.; Zhang, A.; Jakku, E. Farmers and their data: An examination of farmers’ reluctance to share their data through the lens of the laws impacting smart farming. NJAS-Wagening. J. Life Sci. 2019, 90, 100301. [Google Scholar] [CrossRef]
  87. Mishra, A.C.; Das, J.; Awtar, R. An emerging era of research in agriculture using AI. J. Sci. Res. Technol. 2024, 2, 1–7. [Google Scholar] [CrossRef]
  88. Gavai, A.K.; Bouzembrak, Y.; Xhani, D.; Sedrakyan, G.; Meuwissen, M.P.; Souza, R.G.-S.; Marvin, H.J.; van Hillegersberg, J. Agricultural data Privacy: Emerging platforms & strategies. Food Humanit. 2025, 4, 100542. [Google Scholar] [CrossRef]
  89. Dara, R.; Hazrati Fard, S.M.; Kaur, J. Recommendations for ethical and responsible use of artificial intelligence in digital agriculture. Front Artif. Intell. 2022, 5, 884192. [Google Scholar] [CrossRef] [PubMed]
  90. Linsner, S.; Kuntke, F.; Steinbrink, E.; Franken, J.; Reuter, C. The role of privacy in digitalization–analyzing perspectives of German farmers. Proc. Priv. Enhancing Technol. 2021, 2021, 334–350. [Google Scholar] [CrossRef]
  91. Schieck, M.; Braun, B. The effect of transparency on trust and adoption of AI: A survey in German viticulture. In Precision agriculture’25; Wageningen Academic: Wageningen, The Netherlands, 2025; pp. 1151–1158. [Google Scholar]
  92. Posadas, B.B.; Ogunyiola, A.; Niewolny, K. Socially responsible AI assurance in precision agriculture for farmers and policymakers. In AI Assurance; Elsevier: Amsterdam, The Netherlands, 2023; pp. 473–499. [Google Scholar]
Horticulturae 12 00719 g001
Figure 1. Conceptual workflow of smart viticulture, showing the progression from data acquisition through in-field sensing and remote sensing to data integration, AI-based analytics, decision support and vineyard action. The figure highlights how these interconnected stages support targeted interventions, improved resource-use efficiency, enhanced grape quality and greater resilience under variable production conditions.
Figure 1. Conceptual workflow of smart viticulture, showing the progression from data acquisition through in-field sensing and remote sensing to data integration, AI-based analytics, decision support and vineyard action. The figure highlights how these interconnected stages support targeted interventions, improved resource-use efficiency, enhanced grape quality and greater resilience under variable production conditions.
Horticulturae 12 00719 g001
Horticulturae 12 00719 g002
Figure 2. Conceptual framework linking the main potential benefits of smart viticulture with the principal barriers that still limit broader adoption in commercial vineyards. The figure emphasizes that scalable and trustworthy implementation depends not only on technological performance, but also on robust validation, practical usability, interoperability, grower-oriented decision support and transparent governance.
Figure 2. Conceptual framework linking the main potential benefits of smart viticulture with the principal barriers that still limit broader adoption in commercial vineyards. The figure emphasizes that scalable and trustworthy implementation depends not only on technological performance, but also on robust validation, practical usability, interoperability, grower-oriented decision support and transparent governance.
Horticulturae 12 00719 g002
Table 1. Main AI application domains in viticulture and their relevance to sustainable growing systems.
Table 1. Main AI application domains in viticulture and their relevance to sustainable growing systems.
Application DomainTypical Data SourcesTypical AI MethodsMain Vineyard ObjectiveRelevance to Sustainability/Grape QualityIndicative Quantitative EvidenceCurrent Maturity/Main Limitation
Vine status and stress monitoring IoT sensors, thermal imagery, multispectral imagery, weather data.ML regression/classification, deep learning, anomaly detection.Identify water stress, heat stress and spatial variability.Supports targeted irrigation and localized management; helps protect fruit composition.Quantitative performance is often promising, but remains strongly dependent on local calibration, temporal coverage, and validation across sites and seasons [45,46].Relatively mature; performance still depends on data quality and local calibration.
Disease and pest detectionRGB images, multispectral/thermal images, leaf wetness and weather data.Computer vision, CNNs, image classification.Early diagnosis and risk-aware plant protection.Improves treatment timing and may reduce unnecessary spray applications.UAV-based deep learning mildew detection has been reported at >92% grapevine-level and 87% leaf-level detection under well-defined conditions [33].Among the most advanced domains; field performance can decline under variable real-world conditions.
Irrigation and nutrient managementSoil moisture, microclimate, plant indicators, historical responses.Predictive ML, DSS optimization.Refine irrigation and fertilization decisions.Improves resource-use efficiency and helps balance yield with quality.Variable-rate irrigation studies have reported 18% lower irrigation water use without yield or quality loss under site-specific field conditions [38].Agronomically valuable, but often limited by infrastructure, transferability and site-specific responses.
Yield and quality predictionPhenology, imagery, weather records, historical yield data.ML prediction, feature extraction.Forecast yield, maturity and spatial harvest variability.Supports logistics, selective harvest and quality-oriented decisions.Early yield prediction studies have reported R2 = 0.54–0.87 and NRMSE = 16.47–39.17% [41], while quality-trait prediction has reached R2 up to 0.97 depending on trait and cultivar [42].Promising, but less robust across cultivars, seasons and vineyard conditions.
Automation and roboticsVision systems, navigation sensors, machine interfaces.Object detection, robotic control, AI-assisted automation.Assist repetitive and time-sensitive vineyard operations.Potential to improve precision and reduce labor constraints.Autonomous robotic pruning has been reported at 213 s/vine with 87% pruning accuracy under field conditions [44].Still less mature in practical terms due to terrain, cost and operational complexity.
Table 2. Key barriers to large-scale adoption of smart viticulture tools.
Table 2. Key barriers to large-scale adoption of smart viticulture tools.
BarrierTypical Manifestation in VineyardsIndicative Evidence/ImplicationPractical ConsequencePriority Response
Data quality and reliability issues Missing values, sensor drift, unstable connectivity, variable image quality, inconsistent annotation.Model performance often declines when systems move from controlled datasets to unstable field conditions, particularly under variable image quality, incomplete records, or inconsistent annotation [53,56,71,72].Reduced model reliability and lower confidence in digital outputs.Standardized data collection, calibration protocols, quality control and robust validation.
Limited transferability across production contextsPerformance declines across terroirs, cultivars, rootstocks, seasons and training systems.Cross-site and cross-season deployment remains one of the clearest weaknesses of current AI systems, especially when models are trained on specific cultivars, canopy structures, or local environmental conditions [41,45,46,74,75,76].Restricted applicability outside pilot conditions and reduced practical value.Multi-site and multi-season datasets, external validation and context-aware model development.
Economic and organizational constraintsHigh costs of equipment, software, subscriptions, maintenance and training; limited technical capacity and time availability.Adoption is often constrained not only by initial investment, but also by uncertain returns, time demands, and the practical burden of integrating digital tools into existing management routines [20,48,49,79].Slow adoption, partial use of digital tools, or abandonment after initial uptake.User-friendly systems, advisory support, training and clearer evidence of cost-effectiveness.
Interoperability gaps and vendor lock-inPoor compatibility among platforms, limited data portability and dependence on proprietary ecosystems.Weak interoperability can reduce long-term usability, increase dependence on single providers, and complicate routine decision-making across the production chain [28,78,80,81,84].Fragmented workflows, reduced flexibility and long-term dependence on specific providers.Open standards, interoperable platforms, portable data formats and transparent data access policies.
Governance, privacy and cybersecurity concernsUnclear data ownership, uncertain access rights and vulnerabilities in connected systems.Uncertainty about data ownership, access rights, and digital security can reduce willingness to share data or rely on AI-supported tools in routine management [55,59,84,85,86,88].Low willingness to share data and hesitation toward broader digital integration.Clear governance frameworks, secure infrastructures, transparent responsibilities and trust-building mechanisms.
Limited transparency and explainability of AI outputsRecommendations that are difficult to interpret or poorly aligned with vineyard realities.When AI recommendations are difficult to interpret, growers may hesitate to trust them even when technical performance appears promising [77,89,90,91,92].Reduced user trust and lower integration into practical decision-making.Explainable models, transparent decision-support interfaces and stronger human oversight.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Nathena, E.Z.; Psyllakis, K.; Petoumenou, D.; Kontaxakis, E. Smart Technologies and Artificial Intelligence in Sustainable Viticulture: Applications, Benefits, Barriers and Governance for High-Quality Grape Production. Horticulturae 2026, 12, 719. https://doi.org/10.3390/horticulturae12060719

AMA Style

Nathena EZ, Psyllakis K, Petoumenou D, Kontaxakis E. Smart Technologies and Artificial Intelligence in Sustainable Viticulture: Applications, Benefits, Barriers and Governance for High-Quality Grape Production. Horticulturae. 2026; 12(6):719. https://doi.org/10.3390/horticulturae12060719

Chicago/Turabian Style

Nathena, Evangelia Zoi, Kyriakos Psyllakis, Despoina Petoumenou, and Emmanouil Kontaxakis. 2026. "Smart Technologies and Artificial Intelligence in Sustainable Viticulture: Applications, Benefits, Barriers and Governance for High-Quality Grape Production" Horticulturae 12, no. 6: 719. https://doi.org/10.3390/horticulturae12060719

APA Style

Nathena, E. Z., Psyllakis, K., Petoumenou, D., & Kontaxakis, E. (2026). Smart Technologies and Artificial Intelligence in Sustainable Viticulture: Applications, Benefits, Barriers and Governance for High-Quality Grape Production. Horticulturae, 12(6), 719. https://doi.org/10.3390/horticulturae12060719

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop