The basis of the digital technology categories and most relevant potentials, barriers, and risks for NCES were presented in the technical report of Kliem et al. (2023) as a preliminary project. Therefore, that basis of categories and indicators was used and further developed and compiled for this study. The literature search consisted of four rounds and an expert discussion.

The first selection of digital technologies (“Identification,” Figure 1) of the academic research was done by means of appropriate article abstracts that deal with the review topic between November 2022 and December 2023. Technologies that solely aim at the economic optimization of agricultural production and administration were not considered. Digital technologies used for indoor livestock management and farming were excluded (e.g., milking, cleaning, and feeding robots). A few virtual fencing sensor technologies were incorporated into the study because of their relevance to NCES. Technologies which are available in German language but only on the market for farmers in Austria and Switzerland, were sorted out. From a total of 993 peer-reviewed articles on agricultural-related digitalization and internet search, an extensive list of technologies with a total of 189 entries was compiled (Figure 1; see Supplementary materials).

Considering that many of the commercially available digital products have been investigated in research on a limited scale or not at all, an internet search via Google was used to complement the list of technologies (“Identificaition, Screening,” Figure 1). In the next step, the data portals of ministries, authorities, agricultural research institutes, and associations of the individual federal states in Germany were browsed. Individual search queries were performed for each technology group in both English and German. Additionally, in Supplementary materials we presented the most important information on each technology, manufacturer, availability, and application areas.

At the time the review was conducted, there was a boom of announcements and abstracts from technology manufacturers about new products, research projects, and case studies on digital technologies that could support NCES. However, in most cases no further information on NCES indicators was disclosed or the research projects were in the early stages and no results were available. Therefore, we did not compile an empirical impact or a comparative analysis of scenarios from real-world NCES oriented digital technology use in regions within Germany.

Projected potentials in this manuscript refer to potentials not yet practiced on a large scale, which, however, were outlined, during the study, as great opportunities for NCES-related measures in agriculture in general.

The NCES indicators were additionally compiled after an expert talk conducted in June 2022 (“In-depth Review,” Figure 1), whereby experts from agriculture, research, policy making, industry and non-profit organizations were asked to describe current NCES issues regarding digital technologies in Germany. This expert talk was used as a screening of the situation on NCES potentials, barriers, and risks through digitalization and used for the further search. The talk was done online with a total of 23 experts (eight researchers, four farmers, four policy makers, four representatives from civil society and three technology developers from industry). A table with the results of this expert talk is available under Supplementary materials.

Afterwards, the projected potentials, barriers, and risks were reviewed with additional keywords “Sustainability” and “Nature Conservation.” Overall, the results of relevant potentials, barriers, and risks were compiled in NCES indicators as listed under Figure 1. We also included additional information from the screening process of websites of technology manufacturers, which cannot be classified with keywords. We used relevant information and examples of technology manufacturers in the assessment of potentials, barriers, and risks.

For this study, two classifications related to the original mode of operation of the technologies listed, were elaborated in a first step: software- and hardware-based technologies (Figure 1). The categories were expanded to a total of 7 in a second step after the literature review: (1) Farm management information systems and decision support systems (FMIS/DSS); (2) Digital technologies for guidance and steering (DTGS); (3) Digital information platforms (DIP); (4) Citizen science applications and platforms (CSAP); (5) Sensors; (6) Field robots (FR) and (7) Unmanned aerial vehicles and systems (UAV/UAS) (Figure 1). Figure 2 schematically illustrates the hierarchy of the digital technologies’ categories in this study toward machine learning (ML), artificial intelligence (AI) and Big Data (BD). In the middle are ML and AI, which are the cornerstones for the extensive functions of software-based technologies. ML and AI are independent concepts, whereby the ML configurations have hardly any points of contact with the everyday life of farmers – it is a task for computer and mathematics experts, who rely on the data generated by farmers (De Jong and De Boer, 2009). Farmers provide the input for the ML algorithms, but they do not design these algorithms themselves. BD is a term for the processing of very large and heterogeneous data volumes at high speed. Therefore, it is combining both ML and AI as predictive analytics and as an important component of business intelligence (Bhat and Huang, 2021).

The potentials are presented in two levels in Table 3. In the first section of Table 3, the projected potentials are listed as the top category. In the sub-category, the digital technologies that could be assigned to the potentials are listed by mode of operation.

The mitigation of GHG emissions from crop cultivation can be supported by digitalization (Soto et al., 2019). DSS can predict and proceed information on nitrous oxide and carbon dioxide emissions from different cropping systems and help farmers with crop management (Volpi et al., 2020). DTGS and sensors can have a positive impact on NCES through a reduction of fuel consumption (Cillis et al., 2018), which can lead to 0.3 to 1.5% less carbon dioxide of the total GHG emissions of the EU in agriculture (Soto et al., 2019). Germany is ranked second in the EU for successfully reducing GHG emissions through the use of digital technologies (Soto et al., 2019) and was ranked with high carbon dioxide mitigation potentials at relatively low costs (Fellmann et al., 2021).

Sensors offer site-adapted fertilizer application through chlorophyll measurements for improved nitrogen (N) use efficiency (Ali et al., 2020; Edalat et al., 2019; Noack, 2018; Rogovska et al., 2019; Spiegel et al., 2021). If crops are fertilized according to their nutrient requirements, less surplus N is released into the environment and causes less negative impacts, such as GHG emissions (especially during soil cultivation after harvest) (Haas et al., 2022; Simionescu et al., 2019). Furthermore, nitrous oxide emissions and nitrate leaching can be reduced through time-adjusted fertilizer application by means of sensor-based data (Alshihabi et al., 2020; Fellmann et al., 2021).

Site-adapted or reduced application of pesticides also represent a NCES approach. Sensors, FR, and drones work at the plant level: they use an own camera or newer models a data source (e.g., satellite imagery and AI) that can differentiate crop plants from weeds or healthy from pest-infested crop plants. Sensors also detect fertilizer requirements and water stress in plants to support site-adapted management for healthy plant stands. Site-adapted pesticide application relies mainly on DTGS. Reduced pesticide application shows favorable NCES-related effects (Kuhn et al., 2022; Ludwig-Ohm et al., 2023). DSS can recommend the optimal herbicide rate and time of application (Van Evert et al., 2017). Sensor-assisted applications can reduce the amount of herbicide by 20 to 40% (Kempenaar et al., 2017; Kliem et al., 2023).

Planning a crop rotation is a time-consuming process in which many different aspects must be considered: previous crops, the plant-available mineral content of N in the soil, the soil characteristics of the site, weather conditions, pests, and disease, among others. As planning must take place separately for each field, digital technologies can facilitate the process by collecting and partially analyzing the necessary data to decide which crops to plant to diversify crop species and crop rotations. FMIS/DSS help farmers to choose the optimal site-specific crop rotation in potato cultivation based on data and laboratory analysis of soil sampling (Haverkort and Kempenaar, 2016; van Evert et al., 2018).

UAVs offer high-definition image-processing in combination with object and pattern recognition, which can be enhanced, restored and analyzed for calculating and mapping of NCES (Da Silva and Mendonça, 2005; Patrício and Rieder, 2018). Drones’ most important usage is in weed detection, detection of nutrient and water stress, mapping and management (Boursianis et al., 2022; Tsouros et al., 2019). Other applications include predicting of crop development, yield and plant health (Brugger et al., 2023; Tsouros et al., 2019). Multispectral images from sensors can also be useful for calculating and assessing agroecosystem services (Gascuel-Odoux et al., 2022). Although weeds are reluctantly tolerated among crops, they should be considered a direct measure for biodiversity conservation (Steinmann, 2020). In this sense, a direct contribution to NCES is also possible with the help of field robots: they can distinguish between crop plants and other plants when a certain degree of ML is reached (Mathanker et al., 2010; Steward et al., 2019). If a sufficient database associated with NCES becomes available, FR will be able to contribute to direct biodiversity conservation by targeting endemic and protected plants and sparing weeds for insect feeding during mechanical weed control. UAVs can be successfully used in wildlife ecology, especially to observe bird species and their nests (Ogawa et al., 2021; Santangeli et al., 2020), to detect mammals (De Kock et al., 2022; Psiroukis et al., 2021) and indicator plants of, e.g., high nature value (Basavegowda et al., 2022). By combining FMIS with other hardware-based technologies, farm management and system planning can be used to contribute to nature conservation (Mouratiadou et al., 2023).

FMIS/DSS, DTGS, FR and sensors can help farmers to improve soil management. Measures for soil protection and preparation are closely linked to the mitigation of GHG emissions (Fellmann et al., 2021). DTGS and sensors can help farmers to adapt and implement precise tillage management (Gabriel and Gandorfer, 2022; Kliem et al., 2023). Sensors and FMIS/DSS process large datasets that include numerous variables, such as soil temperature, soil humidity, weather, and crop plants, and can structure the results as recommendations to reduce soil defects (Javaid et al., 2023). Mechanical lightweight FR weed control can significantly increase soil bulk density in the topsoil layer compared with conventional heavy agricultural machinery (Bručienė et al., 2022).

The major climate challenges farms are facing in Germany are increasing winter precipitation along with a higher risk of erosion and nutrient leaching, an increase in dry periods during the main growth stage, and heavy rainfall events that lead to soil erosion and flooding’s (BMEL, 2022b). Climate change management includes a change in crop rotation, e.g., with the help of crop managing through FMIS/DSS (Mukhamedova et al., 2022; Novkovic et al., 2017) and landscape diversification (Donat et al., 2022; Hernández-Ochoa et al., 2022). Adaptation actions such as improved soil organic carbon management can also have mitigation co-benefits (Lehmann and Dwerlkotte, 2023). Any adaptation measures that increase the resilience of NCES to climate change – for example, reduced fragmentation or extending natural habitats – can allow species to persist (IPCC, 2023).

Digital technologies cannot stop climate change alone, but they can help farmers implement innovative cropping systems. The digital platform AgoraNatura² aims at enabling anyone who manages land and wants to implement a nature conservation project to finance it via crowdfunding or through partnerships with companies. Private investors and companies can specifically promote biodiversity and nature services by purchasing nature conservation certificates. The price of a single certificate (as a donation for a certain project) is between 3 and 20 euros. The donations will be used for the development of, e.g., an herb- and species-rich grassland and for the selective introduction of important plant species in areas with open patches of soil until 2027 (Geppert et al., 2024a). The Uckerbot field robot has been developed in cooperation between industry and research. It is a system that promotes ecological sugar beet cultivation under unfavorable soil conditions while simultaneously supporting weed biodiversity. The weed diversity on the field is examined from the beginning of the field robot development as an adaption measure to poor soil conditions (Steinherr et al., 2023). Furthermore, FR can work 24 h a day, 7 days a week, allowing farmers to adopt diverse small-scale agroecological-friendly approaches (Daum, 2021). CSAP and DIP are particularly relevant for NCES, as biodiversity indicators and the diversity of flora and fauna must be first recorded and assessed before NCES can begin and lead to profound decision-making (Fischer et al., 2020; Jones, 2020).

The communication and knowledge-sharing among different stakeholders are important for the development of new digital tools and the collection and analysis of data for NCES (BMEL, 2022a). New learning opportunities emerge in the context of agricultural knowledge and innovation systems (Ingram and Maye, 2020), where communication and interaction between different actors is a crucial component to push innovation processes (Knierim et al., 2015; Van de Gevel et al., 2020). These actors can be individual farmers, whole farms, and extension services for farmers. Knowledge sharing is necessary in the same environment (e.g., in same the geographical region) as well as at the national and EU levels. A survey from Germany showed that 89% of the participating farmers use smartphone in their daily life and work, with 79% of them agreeing that mobile and digital communication will make it easier to check farm workflows (Fecke et al., 2018). Farmers want to show society that they are improving their skills and are willing to profit from all new digital technologies and information in order to improve their production and sustainability (Schnebelin et al., 2021). With the help of DIP experts can supply farmers, e.g., with knowledge on crop management options excluding or mitigating the use of pesticides (Heimstädt, 2023).

In this section, in contrast to the potentials (Table 3), the barriers and risks are shown in Table 4 in three stages. The barriers are listed in the first section of the table in orange, with the first level representing the top category, which has been projected into three to four sub-categories. In this case, the categories listed could mostly be assigned to all technologies. The projected risks are in red in Table 4.

The average income of farms with arable area under 100 ha is rather low, whereby almost 69% of all farms in Germany manage less than 50 ha (Destatis, 2021). High costs for acquisition, maintenance, and service are the most impeding factor preventing the use of all digital technologies (Table 3; Gabriel and Gandorfer, 2020) and especially for NCES purposes, because farmers cannot afford to invest in expensive technologies, such as FR or UAV/UAS (Geppert et al., 2024a). Since many technology manufacturers do not provide prices online, the authors of the study wrote to four companies that offer drone services for agriculture in Germany. The names and locations of the companies are kept anonymous. A drone survey with RGB, multispectral and thermal camera for a one-day flight for a maximum of 300 ha costs between 6,000 and 9,000 euros, depending on the company’s offer. The evaluation of the raw data is an additional service, which can vary between 1,500 and 3,000 euros. In this case, a farmer will not have a direct income source for providing NCES, as the NCES monitoring will not increase the price of the agricultural products. For this reason, satellite images from current and previous years are mostly used. Furthermore, policy makers are not ready to finance acquisition costs for FR and contractors do not offer FR hire yet. A single FR acquisition starts between 50,000 and 70,000 euros, as found in the review process.

The administrative workload on farms for NCES through digitalization is an increasing obstacle in Germany (Table 3; Gabriel and Gandorfer, 2020). This burden realistically results in little time for the implementation of NCES, especially when the practical implementation is associated with additional administrative workload (Brown et al., 2021) and there is a threat of strict sanctions in case of administrative gaps (Joormann and Schmidt, 2019). Digital administration platforms do not necessarily facilitate farms as the regulations remain very complicated, the sanctions strict, and the controls time-consuming (Reissig et al., 2022). Based on our review the requirements for applying for NCES subsidies from CAP are equally as complicated in the digital format. Furthermore, payments for specific NCES are currently quite low (Batáry et al., 2015; Regulation EU, 2013).

Autonomous technologies, such as FR, are facing legal barriers as they are still in a grey area according to EU law (Basu et al., 2020). In Germany, the European Machinery Directive (European Parliament, 2006) applies to manufacturing and marketing: autonomous robots must be operated by a person. According to Basu et al. (2020), the applicable laws for autonomous field robots are not so clear and the introduction of the term field robot (such as an agribot) is necessary, as the working conditions in the field are not equivalent to other areas of operation, such as road traffic. In December 2022, the European Commission announced that a new proposal for a Regulation of the European Parliament and of the Council on machinery products had been filed (European Commission, 2022). From the current point of view, the liability regime is challenged by technical legal arguments.

EU-wide regulations for the operation of UAVs have been in force since the beginning of 2021, following a decision by the European Commission on 24 May 2019 (Commission Implementing Regulation EU, 2019). For the operation of drones ≥250 g, proof of competence and registration of the drone is mandatory (Commission Implementing Regulation EU, 2019). Based on our request to the responsible authority, a processing time of 6 weeks is estimated for the approval of one drone flight for monitoring purposes over an agricultural area.

The data management of digital technologies is characterized by a difficult legal situation and controversial discussions about data protection (Lutz, 2017; Vogel, 2020). The problem is mainly the General Data Protection Regulation (Regulation EU, 2016), currently in force within the EU, which does not guarantee data sovereignty for non-personal data. As the application of digital technologies involves the collection and storage of very sensitive data (e.g., location, engine operating hours, and fuel consumption), there is still skepticism among farmers regarding data protection (Schleichler and Gandorfer, 2018). There is also a danger in handling digital data, which is a valuable commodity for data traders (Clasen, 2021). Large-scale data collections and analyses are often not directly accessible to farms, but rather to global and financially strong actors (Zscheischler et al., 2021). Farmers do not want their data to be stored on unknown/foreign servers that are not a subject to the EU law, as pointed out during our second expert talk. Systems that can be used on local computers are considered to be trustworthy for farmers (Gabriel and Gandorfer, 2020). However, with increasing use of AI the storage of data is moved to servers. It is therefore important for policy makers to intervene and endorse state-subsidized technologies or demand more transparency from large companies. The DIP named OdiL (abbreviation in German for “open software platform for service innovations in an agricultural value network”), which is currently in development, aims to enforce data ownership to each farm during data exchange with different stakeholders to win the trust of each user (Hertzberg et al., 2020). The Dutch data platform Akkerweb similarly promises farmers that they can decide for themselves exactly which farm data they want to pass on for further use (Van Evert et al., 2018). Another aspect is the fear of data collection, as mentioned by farmers in our second expert talk – farmers are afraid that if they allow NCES indicators to be monitored by authorities and a rare or severely endangered plant or animal species is recorded, their farmland will be blocked. For this, farmers do not receive any subsidies in Germany.

There are various limitations of digital technologies in practice in arable farming. The data collection and recording of biodiversity CSAP are in a very initial phase (Ball-Damerow et al., 2019). While sensors generate a great amount of data, only a very small fraction of these data can be used in agricultural DSS or software-based technologies. The main reasons are compatibility problems when processing data from different sensors, excessive time and cost expenditures, and a lack of multicausal DSS (Kehl et al., 2021). Public agricultural data are collected for administrative purposes only and there is currently no intention to collect and/or use it for decision-making regarding NCES (Luyckx and Reins, 2022). Furthermore, it is not yet clear whether large datasets will be able to capture and assess the complexity of agricultural systems, NCES as a whole (Delgado et al., 2019).

Further challenges for practice arise from data quality. The GreenSeeker optical sensor for instance, is for large-scale measurements, as smaller measurement units can lead to measurement errors (Ali et al., 2020). Limitations also emerge from the calibration of sensors, which are primarily intended to detect the genetic diversity of solely a few common crops, such as wheat (Khadka et al., 2021; Pratap, 2019). Currently, the genetic diversity of other crops and/or weeds in the stand remains unaddressed for cost reasons. Furthermore, plant disease are often caused by more than one pathogen and can show different symptoms, including different colors and patterns, which is a challenge for sensors, since they rely on image material (N. Zhang et al., 2020). UAVs have some disadvantages for species’ protection: they can cause noise pollution and disturb bird species (Schrader, 2017). The comprehensive use of drones is set back by the necessity of complex and expensive image processing and analysis software for application in practice. In addition, real-time recordings and analyses with UAVs are still not possible. The delayed provision of the results of drone pictures limits their implementation in agriculture (Kehl et al., 2021).

The maturity level of FR in practice is an impeding factor for NCES use (Geppert et al., 2024a). The current working speed of FR varies between 1 and 4 km/h (Gil et al., 2023). Furthermore, the effectiveness of weeding can vary greatly between the FR, depending on the site, soil type, weed infestation, weed composition and weeding time (Ahmad et al., 2014; Bručienė et al., 2021; Fountas et al., 2015). Especially on loamy soils, FR must have enough power and weight to successfully plough and further cultivate the soil such as farm tractors. Up to now, FR are still representing niche products (Kehl et al., 2021).

Digitalization could push the further intensification to cultivate agricultural residual areas that are important for NCES as ecological stepping stone biotopes, insofar as they are not under protection or ecological priority areas within the framework of the CAP. This phenomenon could lead to the loss of valuable areas. At the same time, there are concerns that an increase in the number of large-scale farms alongside digital technologies would have a negative impact on NCES. There is fundamental skepticism regarding the use of algorithms that are being developed primarily for yield enhancement (Reichel et al., 2021). If NCES is not considered from the beginning by policy makers and farmers, it will be more difficult to add new drivers into such a complex system at a later time (Geppert et al., 2024a; Kliem et al., 2023). The precise and intensive use of all land areas can lead to a reduction in habitat niches in the marginal areas (Reichel et al., 2021). Currently, site-specific functions in FMIS are less represented (Melzer et al., 2023). During the expert discussion in this study, experts saw a higher risk for intensification on conventional farms through digitalization, if financial support of NCES is not sufficient. For organic farms the risk of NCES losses through digitalization were described as low. Organic farming, as an agricultural concept, concentrates on the promotion of NCES mainly through agro-ecological practices and not through digitalization.

A great challenge is also the interoperability among digital technologies from different vendors. In general, there is a lack of compatibility among data formats and available systems. There are issues with data from sensors and sensor networks when a transfer in a DSS or FMIS application is needed, which is why a variety of isolated solutions are offered (Kehl et al., 2021). Due to non-uniform federal systems, it is for instance not possible to use public geospatial data on a national scale in Germany, which leads to data fragmentation in digital technologies. The Shapefile (.shp) from ESRI is meanwhile a common data format for software-based technologies in agriculture for the storage of geometric locations. However, its size is limited to 2 GB and it does not support topologic information. ISO-XML commonly used as a communication protocol between agricultural machine terminals (MICS, e.g., sensors, robots) and agricultural software (e.g., FMIS/DSS). There are different ISO-XML versions available, so it is not always possible to use together hardware and software from different vendors. With large volumes of data, it is practically impossible for farmers to find time to look at many databases and check data sets or formats.

Digital technologies can lead to trade-offs. An increased use of digital tools can lead to a higher demand of energy and certain raw materials. The mining and installation of raw material can cause massive damage to the environment and wastes are costly and difficult to dispose (Hoiß, 2023). Furthermore, it is not known with certainty whether data storage is relying on renewable energy sources (Van der Velden, 2018). Moreover, data centers produce high amounts of heat loss, which negatively contributes to global warming. Likewise, the production and use of digital and smart technologies demands a lot of energy (Gensch et al., 2019).

Germany is one of the world’s leading nations in the production and use of modern technologies. With the help of digitalization, the competitiveness of German agriculture can be advanced (BMEL, 2022a). However, the ambitious goals set regarding digitalization as a solution for NCES problems have not yet been fulfilled, and future success depends heavily on agriculture as well as social acceptance and economic and political conditions.

The main participants in the digitalization discussion currently are researchers and the smallest cohort of participants are farmers (Martens and Zscheischler, 2022). In our project, we also found that farmers were rather unwilling to participate in talks on NCES promotion through digital technologies, while researchers and industry already expressed ideas for the direction of digitization. The opinions of the experts from the competence areas of research and farms differ greatly regarding digitalization and NCES. Researchers tend to see a broader spectrum of obstacles and preconditions for NCES measures through digital technologies before, during and after the technologies reach the field. Farmers were interested in practical and financial preconditions for NCES measures with digital technologies during and after applications on field. For the future development and use of digital technologies, as well as many research gaps, it is important for farmers to address the economic aspects that currently characterize everyday life on a farm in public discussions (Kliem et al., 2023). According to a survey of 500 farmers in Germany (2022), some of the biggest obstacles for farmers are (1) the implementation of sustainability measures (67%) and (2) digitalization is an economic challenge for the farmers’ business (51%). Furthermore, surveyed participants from rural areas in Central Europe were asked to describe socio-cultural impacts and found that there are concerns about an exclusion of those subjects who cannot keep up with digitalization (Ferrari et al., 2022).

Alongside farmers, society must also be included in the design of NCES supported by digitalization (Kliem et al., 2023). Beyond that, however, the public perception of digitalization in agriculture has hardly occurred. Assuming that farmers were to actively implement NCES measures using digital technologies, it is not known whether consumers would increasingly access products from innovative systems. The question of how to measure and monetize NCES indicators remains unanswered. This would require a case study that compiles a cost breakdown from the acquisition of a new technology to the commercialization of an agricultural product and examines the monetary value of all NCES implemented measures though digital technologies for consumers, e.g., through choice experiments when buying the agricultural product. However, the problem with NCES value remains a broader discussion and the facilitation of NCES measures through digitalization in agriculture can offer solely a partial solution.

In a study it was concluded that the companies dictating the agricultural digital technology market in Germany, as well as in the rest of the world, and are allied with powerful national and international policy actors (Hackfort, 2023). The author added that digitalization pushes low-tech approaches into the background and argued for a structural transformation on policy and funding (Hackfort, 2023) Whole systems could be marginalized by claiming primacy of “sustainable” digitalization in agriculture, while in practice more time will be needed to achieve any declared sustainability goal. Currently, there are two development paths for the further transformation of the agricultural sector: digitalization of existing management methods and administrative processes, where existing large-scale farm machines and technologies are supplemented with digital and automated applications supported by industry; and cultivation of fundamentally new small-scale technologies (e.g., FR and UAV) that deviate from previously used agricultural machinery and cultivation methods supported by researchers (Gaus et al., 2017; Kliem et al., 2023). Both paths have one thing in common for the future - digital technologies must be cost-, yield- and NCES-effective.

Digital technologies can enable an ecological utopia for sustainability or trigger an ecological dystopia and support unsustainable intensification of agriculture (Daum, 2021). Therefore, the industry alone should not set the guiding goals, nor should policy makers only look at food security. Responsible innovation in agriculture is possible only when a wide range of stakeholders is included in the entire design and development process of a technology (2021). As we found, the current digital transformation is strongly focused on global food security, sustainable intensification, and climate-smart agriculture measures. However, digitalization in organic farming, permaculture, or regenerative agriculture is not of interest due to a lower yield potential (Geppert et al., 2024a). Currently, research does not provide sufficient information on the potentials and risks of how the use of digital technologies affects NCES in Germany. This research gap is an obstacle to the implementation of more NCES-enhancing measures in agriculture. On the other hand, the difficulty analyzing the rapidly developing digitalization based on publications could be behind the development of digital technologies in practice. Either way, without a well-founded knowledge base from research, it is difficult to formulate regulatory measures or policy recommendations (Geppert et al., 2024a). There is a great deal of potential in the use of digital technologies. However, the potentials need to be recognized, developed, and combined for practical NCES so that NCES can be widely implemented in agriculture in the next 10 years (Geppert et al., 2024a; Lange et al., 2023). For NCES transformation of agriculture, however, more far-reaching changes are needed along agricultural and ecological concepts that promote the diversification and long-term resilience of agricultural systems (Geppert et al., 2024a).

Future research needs to determine, based on practical investigations, whether the trade-offs in the production of hardware-based digital technologies undercut or exceed NCES potentials in their use in Germany. Furthermore, aspects of beneficial insect promotion and monitoring in the field using digital technologies are currently neglected, but need active research to promote NCES. It would also make economic sense for commercially available technologies in Germany to be analyzed more closely in practice on their NCES potentials and risks before more and more new technologies are launched.