What ‘smart’ dairy farming technologies are Irish dairy farmers using and why? Exploring perceived benefits and barriers to technology adoption in a pasture-based system

1 Introduction

A rise in the availability of ‘smart’ dairy farming technologies has promised solutions to myriad pressures (Abbasi et al., 2022), such as improving productivity, providing labour savings, reducing environmental impacts and enhancing social wellbeing (O’Brien et al., 2015; Garcia-Covarrubias et al., 2023; Hogan et al., 2023). Answering calls to simplify language around how different advanced technology terms are used (Rotz et al., 2019), this study uses the term ‘smart dairy farming technologies’ to discuss a range of automated, precision, digital and robotic devices, apps and integrated solutions that help dairy farmers manage their animals and farm. In alignment with previous studies (Palma-Molina et al., 2023b), the study classifies smart dairy farming technologies into categories based on the area of dairy farm management in which they are used. The categories are cleaning, feeding, financial management, grass management, herd management and milking technologies. For example, herd management technologies include the use of smart wearable technologies (ear tags and collars) to monitor individual cows for improved health, welfare, breeding and nutrition (King et al., 2021). Financial management technologies include web-based decision support systems, which evaluate physical and economic performance, aiding in financial planning (Aquilani et al., 2022; Papadopoulos et al., 2025). Furthermore, decision support tools that integrate farm grassland data to help farmers determine appropriate actions around pasture use (e.g. PastureBase Ireland) are included in the grassland management technologies category.

The level at which Irish dairy farmers, operating in a predominantly pasture-based system, have adopted these smart dairy farming technologies (i.e. automatic feeders and cleaners, herd and grass management apps, milking solutions) remains varied, with some being adopted more slowly than expected (Palma-Molina et al., 2023b). The slower uptake of individual smart farming technologies by Irish dairy farmers, especially among small- and medium-sized farmers, is partly explained by economic factors. High investment costs make many technologies like milking robots unaffordable (Shalloo et al., 2021), while insufficient cost–benefit evidence can create uncertainty (Kopler et al., 2023). Long-term financial risks from expensive investments may not provide a clear return for multiple reasons, such as a change in market conditions (Aquilani et al., 2022; McGrath et al., 2023) or if a technology is poorly suited to the specific production system (Losacco et al., 2025). Data privacy is also a concern as some farmers want to protect sensitive animal health or production cost data to mitigate for financial exploitation (Regan, 2019; Aquilani et al., 2022). Additionally, some smart farming solutions require higher technical literacy, increasing labour costs for additional training or to recruit specialised staff (Kopler et al., 2023). Other farmers report a lack of financial and training supports, which impact their ability to adopt certain technologies (Drewry et al., 2019; van der Weerdt et al., 2024).

On the flip side, there is growing evidence that farmers increasingly perceive the benefits of adopting smart technologies. For example, in QuantiFarm, a Horizon Europe project that motivated this study, livestock test-case results showed that some of the tested ‘smart technologies’, referred to as Digital Agricultural Technology Solutions, helped farmers solve problems. German dairy farmers found the use of automated sensors reduced daily physical labour requirements through remote observation capabilities. Irish dairy farmers expressed milking robots provided them with more flexibility in their daily tasks, giving them better work–life balance, while the associated data generated at point-of-care improved their ability to proactively respond to cow health issues (Parce and Donnellan, 2025). In Regan (2019) case study analysis, some farmers who expressed data privacy concerns, also saw the need to balance privacy with space for innovation, supporting the benefits of data collection advancements. Other studies (Aquilani et al., 2022; Papadopoulos et al., 2025) have shown that precision livestock farming tools can facilitate a better work–life balance by reducing physically demanding tasks and paperwork burdens. Studies in New Zealand (Dela Rue et al., 2020; Yang et al., 2021), another country with predominantly pasture-based dairy systems, have shown higher farmer preference for investment in automation (labour saving) technologies than data capture (for decision support) technologies. This is consistent with previous studies in Ireland (Palma-Molina et al., 2023b) and is related to the expectation of improving labour efficiency, especially in farms with larger dairy herds (Dela Rue et al., 2020).

Varied levels of technology adoption are observed across Europe and in other countries with pasture-based dairy systems such as New Zealand and Australia, with multi-faceted sociodemographic and economic determinants like herd size, farm size, financial capacity and labour availability influencing technology adoption decisions (Gargiulo et al., 2018; Dela Rue et al., 2020; Yang et al., 2021). These factors operate alongside cultural and behavioural determinants like technology affinity, peer influence and awareness of training supports to create a complex matrix of factors influencing farmers’ decision-making from first encounter with a technology to investment and use, post-adoption (van der Weerdt et al., 2024).

To understand this study’s context, the more recent evolution of Irish dairying, given the impact of milk quota abolishment in 2015, is reviewed, and the state of Irish dairying in 2024 is explained. Since milk quotas were abolished in 2015, Irish dairying fundamentally shifted towards bigger herds and farms to increase productivity (Läpple et al., 2017; Dillon and Moran, 2022). Labour challenges increased alongside these changes in farming practices, prompting farmers to seek technologies, which helped to address expanding labour needs (Deming et al., 2018; Hogan et al., 2022a) and increase flexibility of work practices to achieve improved work–life balance (Regan, 2019; King et al., 2021). Furthermore, Irish dairy farmers, and the wider European farming sector, continue to face challenges, including the negative impacts of geo-political instabilities, market volatility and changing environmental regulations (POLIMI School of Management, 2026).

In 2024, there were 15,131 dairy farms in Ireland with an average herd size of 99 cows (unchanged from 2023 but up by 155% since 2013) operating on an average enterprise area of 47.8 hectares (up by 3% from 2023) (Teagasc, 2025a). Average dairy income in 2024 was €108,700 on an average size farm of 70 ha, up by 113% from 2023 (Buckley and Donnellan, 2025). The average national age of Irish farmers was 58, with dairy farmers included in this figure, while 19% of Irish dairy farmers were identified as having a ‘high age profile’—households in which the farmer is over 60 years old and no one else in the household is under 45 (Teagasc, 2025b). Average 2024 prices paid for milk (actual fat and protein) increased on 2023, up by 20% to €0.515/L. However, 2024 input prices, although fairly stable, remained high with uncertainties continuing across regulatory, climate change and market conditions (Teagasc, 2025a). Farmers in 2024 faced another unknown, related to the renewal of the Nitrates Derogation, a policy monitored by the European Council to protect water quality and its impact on the levels of organic and chemical nitrogen (N) they could spread. Budget allocations for the new Common Agricultural Policy (CAP, 2028–2034) remained under negotiation, causing further concern for Irish farmers.

Amidst this uncertain environment for Irish dairy farmers, the role of smart farming solutions is more important than ever. Farmers continue to need practical, proven innovations that help them remain competitive, sustainable and resilient in the face of change. Employing descriptive statistics and hierarchical cluster analysis, this quantitative study aims to contribute knowledge about smart dairy farming technology adoption in an Irish context. To achieve these goals, this study has three research objectives.

• Identify what smart dairy farming technologies Irish dairy farmers are using in 2024 and at what rates.
• Examine and describe associations between some sociodemographic factors and the adoption of selected smart dairy farming technologies in 2024.
• Identify Irish dairy farmers’ perceptions of benefits and barriers of a selected smart dairy farming technology used or not used by farmers in 2024 across five categories—animal health/welfare, labour, financial, social and environmental.

A survey questionnaire was used to address the research questions.

2 Materials and methods

This study was based on data collected from the 2024 Irish National Farm Survey (NFS). The NFS is an annual survey conducted in Ireland by Teagasc as part of the Farm Accountancy Data Network (FADN) of the European Union (after 2025, NFS data are supplied to the FSDN). The set of Irish dairy farmers who agreed to participate voluntarily in the NFS was selected using stratified random sampling from the 2024 list of NFS farms. Teagasc National Farm Recorders interviewed farmers face-to-face, usually at their home, typically located on their farm. Farmers gave verbal responses to the questionnaire, and answers were recorded by Farm Recorders.

To create a stratified sample, the entire NFS survey population is divided into homogenised groups based on the criteria of farm type. Dairy is one of four (dairy, beef, sheep, tillage) recognised farm types in the NFS. Population weights based on total farm size (ha) and type of farming system (as revised by Ireland’s Central Statics Office 2023 Farm Structures Survey) are applied to the sample, ensuring that data accurately represents farmers nationally (Teagasc, 2025b). In 2024, the NFS sample was 863 farms, representing almost 88,000 dairy, beef, sheep and tillage farms nationally (Teagasc, 2025a). Multiple economic and sociodemographic factors are captured in the NFS, including family farm income, farm size, herd size, farm location and age of the farm operator.

Approximately every second year, the NFS runs an Additional Survey (AS) in the autumn to gather topic-specific farm-level indicators, supplementing knowledge gained from the annual survey. In the 2024 NFS AS, 320 dairy farms were surveyed, with 271 completing responses in the dairy-only section. To determine smart dairy farming technology adoption rates, the 271 NFS AS responses were merged with the general 2024 NFS dataset (320 dairy farms) by farm codes. This process produced a farm-level dataset (n = 250), representing an estimated 11,880 Irish dairy farms nationally.

The 2024 NFS AS questionnaire consisted of 11 questions, of which three were for dairy farmers only. The three dairy-only questions were close-ended and covered key issues around ‘dairy technology use’ and farmers’ perceptions about associated benefits and barriers for adopting or not adopting selected technologies. First, farmers were asked to indicate their smart dairy farming technology use by ‘ticking’ as many as applicable from a list of 18 technologies grouped into six categories: cleaning, feeding, financial management, grassland management, herd management and milking (Appendix A). The 18 dairy farming technologies surveyed were reviewed by a panel of precision technology experts for the purposes of validation. The panel confirmed the inclusion of these technologies because of their known use in pasture-based dairy systems, a decision also supported by the few existing similar studies conducted in Ireland (Läpple et al., 2017; Shalloo et al., 2021; Palma-Molina et al., 2023b). The included technologies purposefully varied in adoption rates to seek a wide snapshot of technology adoption trends in 2024. This approach aimed to show if some popular technologies continued to be adopted at higher rates and sought to understand any changes in adoption rates for other technologies. For example, some technologies have gained momentum more recently [e.g. PastureBase Ireland (PBI), a national grassland decision support tool available from 2013 has more recent additional features and a mobile app originally released in mid-2018) or have been slower to be adopted in an Irish context (e.g. milking robots, <4% adoption rate compared with fluctuating average EU wide adoption between 15% and 25%). Technologies were grouped (Appendix A) into six typical farm management categories (cleaning, feeding, financial management, grassland management, herd management and milking) in accordance with previous studies (Palma-Molina et al., 2023b).

Participants were next asked about smart dairy farming technologies they used and felt had more benefits than other technologies on their farm (Appendix B). Farmers were asked to focus on one of four technologies available (automatic cluster removers, automatic calf feeder, automatic cow ID system and PastureBase Ireland). To follow best practice for close-ended questions, participants were provided with free boxes in the second and third dairy-only questions, capturing responses about technology adoption factors that were not specifically listed. If participants did not want to focus on any of the four listed technologies, they could choose a technology they were using (selected from the previous question’s list of 18) and felt was beneficial, recording it in a free box, ‘Other Technology’.

The four technologies for the ‘benefits’ question were selected because they represent a sample of tools from four different farm task categories and are technologies used to varying degrees on Irish dairy farms. Knowledge gained from this sample can help develop a baseline of adoption decision-making factors. Labour impacts are cited as potential reasons for adoption of the first two technologies, specifically time savings from the installation of automatic calf feeders in an Irish pasture-based system (Hogan et al., 2023) and reduction in operator stress from the use of automatic cluster removers (ACRs) in a New Zealand pasture-based dairy system (Edwards et al., 2015). Furthermore, ACRs (milking category) have been adopted at a higher rate, as previously indicated, while automatic calf feeders (feeding category) are increasingly adopted for various reasons, including stronger financial support from government schemes. Additionally, these automated technologies offer scope to save time for typical Irish dairy farms using a twice-a-day milking/feeding system, especially helpful to farms with expanding herds (Palma-Molina et al., 2023b). Automatic cow ID systems (herd management category) and PBI (grass management category) are particularly relevant for farmers making decisions to support cow health (Shalloo et al., 2021) and grow sufficient grass within a pasture-based system (Hanrahan et al., 2017; Palma-Molina et al., 2023a; O’Brien et al., 2025). However, more advanced cow ID systems, like cow collars, have a lower adoption rate (8.4% in 2018; Palma-Molina et al., 2023) than other tools, while PBI has enjoyed more saturation with around one-third of Irish dairy farmers using the decision-support tool (DST) to some degree (O’Donovan and Maher, 2026). These categories and benefits were selected based on consultation with a panel of experts in precision technologies in Ireland and on frameworks found in existing literature (Lovarelli et al., 2020; Mata, 2026).

Once a participant selected one technology to focus on, they could ‘tick’ as many of the 13 listed benefits of the chosen technology across five categories: animal health/welfare, labour, financial, social and environmental (Appendix B). Following on from the ‘Other Technology’ option stated above, participants also had a choice to indicate ‘Other Benefit’ in a free box.

Finally, as in the benefits question, participants were asked to choose one technology they did not use and felt had more barriers to adoption than others. The four technologies participants could select from for this ‘barriers’ question were automatic drafting system, automatic manure system, individual cow activity sensors and milking robot (Appendix C). Again, if participants did not want to focus on any of the four technologies, a free box, ‘Other’, could be used to record a technology from the original list of 18. With the selected technology in mind, participants were asked to ‘tick’ as many of the 12 listed barriers to adoption in the context of their farm across the same five categories: animal health/welfare, labour, financial, social and environmental, and ‘Other Barrier’ (Appendix C).

The four technologies for this barriers question were selected in a similar manner to the benefits question—designed to provide a sample of technologies from various categories (Appendix A), which were also validated by a panel of precision technology experts. While it is estimated that between 3% and 4% of dairy farms in Ireland have installed milking robots (milking category), little empirical research has been done to confirm actual numbers of robots in use and the reasons behind these low adoption rates. Automatic drafting gates and individual cow activity sensors (herd management category) were among those technologies that were adopted at the lowest rate in 2018 (Palma-Molina et al., 2023b). This knowledge led to their inclusion with an aim to identify some of the associated barriers with their low adoption rates. Lastly, no known studies have captured adoption rates for automatic manure scrapers (cleaning category) in Ireland, and inclusion in this study could provide a baseline for future comparisons.

The study employs quantitative methods using descriptive statistical analysis to explore 2024 NFS AS survey data across continuous [herd size, farm size, family farm income/ha (FFI/ha), proportion of hired labour, farmer age, household members] and categorical variables (gender, agricultural education, derogation). Eight of the sociodemographic factors—herd size, farm size, FFI/ha, proportion of hired labour, farmer age, household members, gender and agricultural education—were selected to mirror those chosen in previous studies (Balaine et al., 2020; Palma-Molina et al., 2023b, Palma-Molina et al., 2025). Derogation status was added as the ninth factor due to the significance of the topic for Irish dairy farmers (O’Donovan et al., 2021; McCormack et al., 2026).

Hierarchical cluster analysis is also used to analyse benefit and barrier dimensions for technologies, which farmers selected as either a technology they used and felt had more benefits than others on their farm, or did not use, perceiving that tool had more barriers to adoption than others.

Descriptive statistical analysis was performed using R statistical software, Version 4.5.0 (R Core Team, 2025). Hierarchical cluster analysis was conducted using Stata software statistical package version 16.1 (StataCorp, 2019).

2.1 Technology adoption rates and associations with sociodemographic factors

Relative frequency distributions of smart dairy farming technologies were estimated to establish the adoption rates of each technology. Additionally, frequency distributions were estimated to establish the adoption rates within each of the six categories of smart dairy farming technologies: cleaning, feeding, financial management, grassland management, herd management and milking. A farm was considered as an adopter of a category of smart dairy farming technology (e.g. cleaning) if it was using at least one of the technologies included in the category. For example, in the cleaning category, a farm would need to use automatic manure scrapers, automatic milk machine washers or automatic washers on bulk tank to be considered an adopter of cleaning technologies.

To integrate data on technology adoption usage with sociodemographic variables, eight variables were selected, mirroring a 2018 study (Palma-Molina et al., 2023b). The variables included were herd size, farm size, FFI/ha, farmer’s age, number of household members, farmer’s gender, agricultural education (farmers with at least one agricultural course completed) and derogation (farm located on nitrates derogation land). Additionally, a ‘proportion of hired labour’ variable was created by dividing the number of paid labour hours with the total number of labour hours in the farm (paid plus family labour).

To explore sociodemographic characteristics associated with technology adoption, one technology per category was selected for the analysis based on the rates of adoption levels within categories. Specifically, technologies with adoption rates closest to 50% were selected as representative of each of the six categories to allow for more optimised statistical comparison between adopters and non-adopters. This approach helped to reduce sampling bias, which may have existed if technologies were chosen for analysis with skewed rates of either very low or very high adoption rates (Agarwal and Devra, 2024), as was the case with some of the surveyed technologies (e.g. milking robots and virtual fencing were adopted at a very low rate of 2%). The technologies selected for analysis were automatic milk machine washers for the cleaning category (45% adoption rate), automatic in-parlour feeders for the feeding category (70% adoption rate), app for financial management for the financial management category (18% adoption rate), app for grassland management for the grass management category (48% adoption rate), app for herd management for the herd management category (59% adoption rate) and automatic cluster removals for the milking category (57% adoption rate). The app for financial management was the only technology listed under the financial management category, eliminating choice when selecting a representative technology.

Two-sample t-tests were conducted to identify differences in the means of continuous variables (herd size, farm size, FFI/ha, proportion of hired labour, farmer age, household members) between adopters and non-adopters for each of the six smart dairy farming technologies representing the six categories. Chi-squared statistics were estimated to test for differences of categorical variables (gender, agricultural education, derogation) between adopters and non-adopters of smart dairy farming technologies. Significance was determined if p < 0.05.

2.2 Benefit and barrier dimensions associated with adoption decisions of chosen technologies

Hierarchical cluster analysis (HCA) was used to classify farmers according to what they perceived as benefits or barriers for a chosen technology. Each farmer chose one particular technology as particularly beneficial and another technology as being particularly difficult. While the HCA analysis was carried out more broadly for all technologies in the associated survey question, specific analysis is centred on the responses from farmers who reported automatic cluster removers as the most beneficial technology (n = 102) and milking robots as the technology with the most barriers to adoption (n = 149).

HCA enabled homogeneous groups to be identified based on the five pre-determined categories (animal health/welfare, labour, financial, social and environmental) for each of the two technologies (automatic cluster removers and milking robots) selected most often for each question. The Gower distance matrix was calculated for each of these technologies, thereby enabling the estimation of the extent to which farms differ from one another based on their scores for each dimension. The methods used to calculate these scores are described later in this section.

The Gower distance is based on the absolute difference in scores between each farm. It is a suitable approach when the data consist of both binary and continuous variables (Pavoine et al., 2009) as is the case for the data used in this research. The ensuing step involved the choice of clustering method to classify farms into various clusters. The Wards linkage method was applied to classify farms into clusters based on the (dis)similarity of the scores for individual farms. This hierarchical clustering method is based on minimising the variation within each cluster (Ward, 1963). In choosing the number of clusters, the stopping rules based on Duda–Hart indices were used (Duda et al., 2000).

Prior to separate cluster analyses being undertaken for the two most selected technologies (ACR and milking robots), analysis was carried out for all technologies listed in the benefits and barriers questions, discussed below.

To better understand perceived benefits of adopting a chosen technology, response data for each of the 13 potential benefits (Appendix B) were classified according to the five dimensions: animal health/welfare, labour, financial, social and environmental. The first step was to calculate a score for each of the five dimensions and to repeat this for each technology, bearing in mind that farmers should only select one technology. For instance, in the case of the social dimension, there were four potential benefits. If a respondent chose just one of these four benefits for a particular technology, the score for the social dimension was 0.25. Alternatively, if the respondent chose all four of these benefits, the score for the social dimension equalled ‘one’ for that technology. As only one benefit fell under the animal health/welfare dimension, the score for this dimension could only take one of two values (either zero or one). For each of the four technologies, the average score for each dimension was compared. This approach enabled a basic understanding of what perceived benefits were associated with the selection of particular technologies.

For each technology, the scores for each of the five dimensions and the 13 individual benefits are reported. However, scores for the five dimensions were not summed to calculate an overall benefit of each technology. Ultimately, the comparison of scores between the technologies is limited because each farmer respondent selects one technology and does not report the benefits for all of the technologies. The best indicator of the most beneficial technology is the number of farmers choosing that technology as the most beneficial. In this research, the findings indicate that this is clearly the ACR technology.

The analysis of the data about the benefits of a chosen technology is therefore focused on automatic cluster removers. This was the technology reported most by farmers as the technology with the highest adoption benefits (n = 102). Duda–Hart indices (Duda et al., 2000) indicated the suitability of a three-cluster solution.

To better understand perceived barriers of not adopting a chosen technology, response data for each of the 12 potential barriers to adoption were classified according to the same five dimensions discussed above (Appendix C). As in the case of the data from the ‘benefits’ question, average scores were first calculated broadly for each of the five barrier dimensions, and this exercise was repeated for each technology, bearing in mind that farmers should only select one technology. This approach enabled an understanding of what perceived barriers were associated with farmers who chose not to adopt particular technologies. For instance, in the case of the animal health/welfare dimension, there were three potential benefits. If a respondent chose just one of these benefits, the score for the animal health/welfare dimension was 0.33. Alternatively, if the respondent chose all three of these benefits, the score for the animal health/welfare dimension equalled ‘one’ for that technology. As only one benefit fell under the labour dimension, the score for this dimension could only take one of two values (either zero or one). For each of the four technologies, the average score for each dimension was compared. As above, this approach enabled a basic understanding of what perceived barriers were associated with the selection of selected technologies.

Specific analysis of adoption barrier scores was centred on the responses from farmers who reported milking robots as the technology they did not use and felt had the highest adoption barriers (n = 149). For the analysis of data about the barriers to the adoption of milking robots, the same indices (Duda et al., 2000) indicated the suitability of a four-cluster solution. For choosing the best number of clusters, decisions were based on the practicalities of needing a reasonable sample size in each cluster.

3 Results

The 2024 NFS AS dairy-only portion of the questionnaire had a response rate of 84.7% (271/320). To determine smart dairy farming technology adoption rates, and as stated above, the 271 NFS AS responses were merged with the general 2024 NFS dataset by farm codes. This process produced a farm-level dataset sample (n = 250), representing an estimated 11,880 Irish dairy farms nationally.

3.1 Rates of technology adoption in 2024

When asked about smart dairy farming technologies they used, participants reported that they adopted certain technologies at higher rates. Figure 1 shows the rates of technology adoption in 2024. Five technologies were reported as being adopted more often than others in 2024 (≥57%): automatic washers on bulk tank (86%), automatic manure scrapers (78%), automatic in-parlour feeders (70%), app for herd management (59%) and automatic cluster removers (57%). The three technologies adopted at the lowest rates (≤10%) included automatic forage feeders (10%), virtual fencing (2%) and milking robots (2%). Approximately 3% of Irish dairy farmers declared they did not use any of the listed technologies. Figure 2 shows the percentage of adoption of smart dairy farming technologies by category (cleaning, feeding, financial management, grassland management, herd management and milking).

As seen in Figure 2, technologies in the cleaning category were clearly adopted at a higher rate than technologies in other categories, with 93% of respondents reporting they adopted at least one cleaning technology surveyed. The technology most adopted in the cleaning category was automatic washers on bulk tank at 85%. Technologies in the feeding category were adopted at the second highest rate, as compared to other categories, with 74% of participants reporting they adopted at least one feeding technology (Figure 2). Financial management technologies are the least adopted category of smart dairy farming technologies, with 18% of participants responding they adopted the app for financial management. The caveat here is that there was only one choice provided in the financial management category, which could have influenced selection rates.

3.2 Associations of sociodemographic factors with smart technology adoption

Table 1 presents the differences in farmer and farm characteristics of adopters and non-adopters of smart dairy farming technologies. The results show that adopters of all categories of technologies had larger herd sizes, and there was a higher proportion of hired labour than non-adopters. Additionally, it was found that derogation was associated with the adoption of all smart dairy farming technologies.

Table 1. Differences in means (SD in parenthesis) of farm and farmer characteristics between adopters and non-adopters of smart dairy farming technologies, 2024.

Source: Authors’ analysis of 2024 Teagasc NFS Additional Survey data.

Statistical tests based on t-tests for continuous variables and chi-squared test or Fisher’s exact test for categorical variables (distinguished by χ² or odds ratio).

*p < 0.05.

Results (Table 1) also show that adopters of cleaning, feeding, grassland management, herd management and milking technologies had larger farm sizes than non-adopters of these technologies. Adopters of financial management, grass management and milking technologies had higher FFI/ha than non-adopters. Adopters of cleaning, grassland management and herd management technologies were significantly younger than non-adopters. In terms of the number of household members, analysis shows that adopters of herd management and milking technologies had a larger number of household members than non-adopters. It was also found that gender was associated with the adoption of milking technologies, with males being less likely than females to adopt the representative milking category technology—automatic cluster removers. There were no significant associations of agricultural education with any of the categories of smart dairy farming technologies. When nitrates derogation status was analysed, this factor was shown to be associated with higher levels of smart technology adoption across all six technology categories.

3.3 Perceived adoption benefits for a chosen technology

Table 2 contains a comparison of average scores for different technologies according to each benefit dimension. A small number of respondents provided a response for more than one technology (six farms). The most popular choice of technology was ACRs with 102 farms selecting this technology as one they used and felt was more beneficial than other technologies on their farm. A large number of farmers (50 farms) responded with a technology other than the four listed technologies. These other technologies included milking robots, cow feeders and herd management software systems. High, moderate and low score parameters were established, with the nomenclature of ‘high’ corresponding to scores of 0.70–1.0, ‘moderate’ to scores between 0.40 and 0.69 and ‘low’ to scores between 0 and 0.39.

Table 2. Average benefit score by dimension according to the chosen technology.

Source: Authors calculations using 2024 Teagasc NFS Additional Survey data.

Animal health/welfare benefit factors score highly for most farmers adopting selected technologies, as seen in Table 2. The average score for animal health/welfare is particularly high (0.864) for those farmers reporting automatic cow ID system as the most beneficial technology. For automatic cluster removers and automatic calf feeders, improved animal health/welfare was also viewed as a benefit of adopting this technology. For farmers choosing the grass management app PBI, there was a moderate average score for animal health/welfare as a benefit of this grass monitoring tool.

The same three technologies reporting high scores for animal/health benefits (automatic cluster removers, automatic calf feeders and automatic cow ID systems) also reported labour as an important benefit (Table 2). Meanwhile, farmers who focused on PBI indicated a low association with labour benefits from using this grass management app. Labour and social dimensions (scores all ≥.516) appear important for the adoption of automatic cluster removers and automatic cow ID systems. For the labour dimension, the average score is relatively high (0.683) for those farmers reporting automatic calf feeders as the most beneficial technology and moderately high for farmers using automatic cow ID systems (0.576) and automatic cluster removers (0.516). The labour dimension is strongly associated with perceived benefits of adopting automatic calf feeders (0.683) and moderate for the social dimension (0.476) for the same technology.

Regarding the financial dimension, the PBI technology for grassland management has the highest average score (0.469). Automatic cow ID systems have the second highest score (0.364), which, along with the remaining technologies (automatic cluster removers and automatic calf feeders), is relatively low. For most technologies, the average score is very low for environmental benefits, except for PBI at 0.656, as seen in Table 2.

Animal health/welfare benefit factors score highly for most farmers adopting selected technologies, as seen in Table 2. The average score for animal health/welfare is particularly high (0.864) for those farmers reporting automatic cow ID system as the most beneficial technology. For automatic cluster removers and automatic calf feeders, improved animal health/welfare was also viewed as a benefit of adopting this technology. For farmers choosing the grass management app PBI, there was a moderate average score for animal health/welfare as a benefit of this grass monitoring tool.

3.3.1 Perceived benefits associated with automatic cluster remover adoption

In this section, results are reported for the cluster analysis based on the farms who choose ACRs. As stated, ACR was the technology selected more often than others as a smart dairy farming technology participants used and felt was more beneficial. Table 3 provides a comparison between the three clusters, generated using the Wards linkage method described earlier, in the scores for each dimension.

Table 3. Average score by dimension for three clusters of farmers who selected automatic cluster removers as most beneficial.

Only applies to farmers who selected ACR (102 out of 220) as the most beneficial technology.

Source: Authors’ calculations using 2024 Teagasc NFS Additional Survey data.

Results show that most respondents (clusters 1 and 2 representing 79.41%) who selected ACR as a more beneficial technology perceive the greatest benefits from the animal health/welfare and social categories. Cluster 1 scores highest for animal health/welfare with all 36 farms reporting this as important for the decision to adopt ACR. Farms in this cluster can be described as ‘mainly animal health/welfare-focused’ when it comes to the adoption of ACR. As in the case of cluster 1, the farms in cluster 2 score very highly for animal health/welfare (0.983). However, since there is only one sub-dimension for animal health/welfare—”improves animal health/welfare”—the analysis above remains as is. All clusters in Table 4 report low scores for the environmental dimension’s influence on their adoption of ACR.

Table 4. Average score by benefit sub-dimensions for three clusters of farmers who selected automatic cluster removers as most beneficial.

Only applies to farmers who selected ACR (102 out of 220) as the most beneficial technology.

Source: Authors’ calculations using the 2024 Teagasc NFS Additional Survey data.

Cluster 2 stands out as reporting many benefits including a high score for labour (0.819) and social (0.786). Cluster 3 has the lowest number of farms (n = 21), and members of this cluster report fewer benefits from the adoption of ACR. Farms in this cluster do report benefits for labour and social dimensions but do not report animal health/welfare as an influence.

Breaking the benefit dimensions down into their component dimensions provides more nuanced results for each cluster. Table 4 shows results for ACR benefits, including the more detailed sub-dimensions for all five benefit categories. Specifically, results across all clusters indicate that farmers felt particularly strongly that ACR “improves animal health and welfare” (animal health/welfare) and “eases physical demand of milking and related tasks” (social).

As shown in Table 4, farmers in cluster 2, representing 44.1% (the highest number of farmers in the sample), give all three sub-dimensions for labour higher scores than farmers in C1 and C3. Under labour, high score results for cluster 2 become more defined when the sub-dimensional results are examined. Cluster 2 farmers have the highest score (0.97) for “labour—easier to hire workers due to their interest in technology”, giving a score of 0.83 for the “labour—frees up time to do additional tasks, previously contracted”. A score of 0.66 is given by cluster 2 to the third labour variable—“labour — reduces need to hire seasonal labour”. Farms in this cluster tend to have a relatively greater reliance on paid labour. Furthermore, cluster 2 social scores were highest, compared with those of C1 and C3, in the sub-dimension benefit of “social—eases physical demand of milking and related tasks” (0.98) and “social—easier to get younger generation involved” (0.94). Regarding financial benefits, this group of respondents (C2) perceived some financial benefits (with responses for “increasing profits”, 0.473, and “reducing vet bills”, 0.554), which corresponds to the high value this group placed on labour factors and animal health. This group reported a very low score (0.05) for the financial benefit of “reduces feed purchased from off-farm sources”, which is consistent with expectations, as ACR is not associated with feeding benefits.

Differences between C1 and C2 in labour and social categories are notable (Table 4). C1 (35.29% of respondents) did not perceive any of the labour factors as strong benefits for adopting ACR, including “easier to hire workers due to their interest in tech”, “reduces need to hire seasonal labour” and “frees up time to do additional tasks, previously contracted”. Similarly, while C1 valued the ability of ACR to “ease physical demand of milking and related tasks”, they did not perceive any other social benefits very positively, including issues of safety, increasing attractiveness for younger generation and offering work–life balance, differing from higher scores in these sub-dimensions by C2 farmers noted above.

Furthermore, Table 4 shows that farms in cluster 1 also did not perceive ACR as making it “easier to hire younger workers” (labour) and they did not think ACR really improves “getting younger generation involved” with farming (social). Examining the data, it is known that while family labour units are similar between clusters 1 and 2, cluster 1 farmers are less likely to have hired labour than cluster 2 farmers and are older (57 years old), on average than cluster 2 farmers (53.4 years old). Without talking to individual participants, it is challenging to specify reasons for these disparities.

3.4 Perceived adoption barriers for a chosen technology

Table 5 shows the results for the comparison of five technologies according to the self-reported barriers associated with the adoption of particular technologies. Financial factors appear to be the most influential barrier for not adopting one of the smart technologies, particularly for milking robots (average score = 0.626). Table 5 also shows that there are some similarities in the average scores between the different technologies. For instance, the average score for the labour dimension is similar for the four mentioned technologies with a narrow range from 0.320 for individual cow activity sensors to 0.369 for milking robots. High, moderate and low score parameters were again established, with the nomenclature of ‘high’ corresponding to scores of 0.70–1.0, ‘moderate’ to scores between 0.40 and 0.69 and ‘low’ to scores between 0 and 0.39.

Table 5. Average barrier score by dimension according to the chosen technology.

Source: Authors’ calculations using the 2024 Teagasc NFS Additional Survey data.

For all of the technologies, the average score appears relatively low for the social dimension, with milking robots having the highest of these low scores (0.204) and all other technologies scoring ≤0.080. Animal health/welfare scores also fall in the lower range, meaning this dimension does not appear to be a barrier to adoption for farmers citing these technologies as being more difficult to adopt.

3.4.1 Perceived barriers of milking robot adoption

As stated, milking robots was the technology selected most (n = 149) for the barriers question. Table 6 shows results from the cluster analysis of four clusters of farmers who reported milking robots as being particularly difficult to adopt. The four clusters were generated using the Wards linkage method as described earlier. These results indicate some contrasting reasons for farmers to report milking robots as a technology they did not use and felt had more barriers to adoption than others on their farms.

Table 6. Average score by dimension for farmers who selected milking robots as a technology that was more difficult to adopt.

Only applies to farmers who selected milking robots (149 out of 220) as the technology with the most barriers to adoption.

Source: Authors’ calculations using Teagasc NFS Additional Survey data.

Cluster 2 (C2) contains farmers who report many barriers to the adoption of milking robots, while the reported barriers are more selective for the other three clusters (C1, C3 and C4). Regarding results about labour barriers (Table 6), since there was only one factor (“don’t think it [milking robot] will reduce no. of workers needed”), there was a clear division between C1/C2 and C3/C4, although the majority of participants (C3+C4 = 63.09%) did not select labour as a barrier. Regarding results about financial barriers, all four clusters gave financial barriers a moderate to high score, ranging from 0.51 (C3) to 1 (C2). Although cluster 2 is the smallest (n = 18), all farmers in this cluster gave all three financial barriers a score of 1.

Cluster 3 contains the most farmers (n = 62). In this cluster, the average scores for the financial dimension (0.514) and environmental dimension (0.500), as seen in Table 6, indicate that these factors have some influence on the perceived difficulty of adopting milking robots, while other factors appear less influential. Farmers in cluster 4 do not report environmental factors as influencing the choice of milking robots as being particularly difficult. Interestingly, some farmers in the other three clusters (C1, C2, C4) report environmental factors as influential. This contrasts with the findings on the perceived benefits of adopting a chosen technology. Instead, farmers in cluster 4 appear to place emphasis on financial and social barriers as particular difficulties with milking robots (Table 6).

Once again, the interpretation of scores becomes more defined when the sub-dimensional results are examined. Table 7 shows average scores for ‘barrier’ sub-dimensions for four clusters of farmers who selected milking robots for this question.

Table 7. Average score by barrier sub-dimensions for four clusters of farmers who selected milking robots as a technology that was more difficult to adopt.

Only applies to farmers who selected milking robots (149 out of 220) as the technology they did not use and felt had the most barriers to adoption.

Source: Authors’ calculations using Teagasc NFS Additional Survey data.

One financial sub-dimension, “investment costs too high” is given a score of ≥.75 by all clusters (C1–C4), indicating that this financial reason was the most significant financial barrier to their non-adoption of milking robots. C2 scored all three financial sub-dimensions with the highest score (1), although this group represents the smallest cluster (n = 12). Another financial barrier, “financial benefits don’t outweigh financial costs”, was given a high score by C2 (1) and C1 (0.79) and a moderately high score by C3 (0.62), while the sub-dimension “unsure what financial supports available” was scored quite low (≤0.34) by all clusters (C1, C3, C4), bar C2, as explained above. However, the environmental sub-dimension “uncertain impacts to help meet environmental targets” was given a low score (≤0.07) by all four clusters, indicating that few farmers checked this factor.

4 Discussion

This study explored the adoption rates for a selection of 18 types of smart dairy farming technologies in Ireland in 2024 across six functional categories: cleaning, feeding, financial management, grassland management, herd management and milking. This research also investigated sociodemographic factors associated with technology adoption and dairy farmers’ perceptions of benefits and barriers of selected smart dairy farming technologies. Results indicate that technology adoption decisions are multi-faceted and significantly associated with certain sociodemographic characteristics of the farmer and farm, as well as various financial and non-financial factors, supporting research by van der Weerdt et al. (2024). While the study focused on a relatively small number (18) of the hundreds of individual software and hardware options available in the market for today’s dairy farmers, findings show technology adoption trends for some of the most important types of adopted and not-adopted technologies. Furthermore, while this study centred on dairy farms in Ireland and the sample of dairy farmers was limited to 271 who completed the 2024 NFS AS, these farms represent nearly 12,000 dairy farmers nationally and so shed light on farmers’ perceptions and provide insight into some of the sociodemographic factors that are associated with smart dairy farming technology adoption. By paying particular attention to ACR as the technology with the most benefits, selected by 46% of farmers, and milking robots as the technology with the most perceived barriers, selected by 67%, more in-depth knowledge is derived from these farmers’ perceptions about adoption and non-adoption factors.

4.1 Smart dairy technology adoption trends

Comparing this study’s adoption rates with those from the study of Palma-Molina et al. (2023b), which created a baseline for smart dairy technology adoption in Ireland, the rates of 10 technologies surveyed in both studies could be compared (automatic manure scrapers, automatic milk machine washers, automatic calf feeders, automatic in-parlour feeders, app for grass management, electronic plate meters, automatic drafting gates, cow activity sensors, automatic cluster removers and milk meters). There was a significant increase in adoption rates across all 10 technologies. Appendix D shows the outcomes of this comparison, with four technologies (cow activity sensors, automatic calf feeders, automatic drafting gates and electronic plate meters) increasing by over 100%.

Several possible explanations support the increasing levels of smart technology adoption seen in this study, compared with the 2018 rates. The literature cites reasons including global socioeconomic shifts (European Commission, 2025), along with improvements in animal welfare management tools (Herlin et al., 2021) and internet connectivity (Gabriel and Gandorfer, 2023), although lack of adequate broadband services in some regions is also cited as an adoption barrier (Kenny and Regan, 2021). Labour shortages and aging farmers further contribute to the increasing use of automated technologies (e.g., ACR, automatic feeding and cleaning systems), as concluded by Morrone et al. (2022), Gabriel and Gandorfer (2023), and the European Commission (2025). The European Commission Joint Research Centre (2025) found that there was a marked increase in technology adoption by European farmers, particularly from 2008 to 2020. There were notable increases for IT, software and crop technologies, likely due to greater acceptance and availability of these technologies and vastly improved internet connectivity (European Commission Joint Research Centre, 2025). The adoption of livestock technologies was steadier across Europe, with the EU JRC report (2025) indicating that the more gradual life cycles of livestock farming technologies could be a factor. The JRC report (European Commission, 2025) stated that technology adoption slowed, 2020–2024, due to the impact of global events like the Ukraine war and COVID-19, causing farm-level financial uncertainties.

In Ireland, a somewhat different story of technology adoption emerges. A comparison of NFS reports from 2018 (Dillon et al., 2019) and 2024 (Dillon et al., 2025) finds a 60% increase in average gross machinery spending by Irish dairy farmers from 2018 to 2024. Machinery is the largest investment category for Irish farms (Dillon et al., 2025). National reports (Department of Agriculture, Food and the Marine, 2024) have shown increases in machinery investments including automated technologies such as automated manure scrapers, automatic calf feeders and milking robots. Another contributing factor to the increases in adoption for all 10 technologies noted above could be the increase in subsidies for investments in smart dairy farming technologies such as the Dairy Equipment Scheme (Department of Agriculture, Food and the Marine, 2025). This subsidy allows for new dairy farmers to have access to financial support to invest in up-to date technologies, including smart dairy farming technologies. Additionally, there is growing evidence of the benefits of investing in some of these technologies in Ireland in areas such as pasture use (e.g. PBI) (Palma-Molina et al., 2023b; O’Donovan and Maher, 2026) and milk yield (e.g. ACR) (Palma-Molina et al., 2023b, Palma-Molina et al., 2025) and in reducing labour demands through automation (e.g. automatic calf feeders) (Sinnott et al., 2021). These factors could have motivated dairy farmers to invest in smart dairy farming technologies.

4.2 Association of sociodemographic factors and technology adoption

The purpose of this part of the study was to identify if there were significant sociodemographic differences between adopters and non-adopters. Across a wide variety of smart dairy technologies, a strong pattern emerged where adopters tend to be larger farms in terms of dairy cow herd size, farm size and family farm income. Confirming some findings from international literature (Gargiulo et al., 2018), larger herd size was found to be highly significantly associated with the adoption of all technologies. Similarly, higher farm income was associated with increased adoption, but financial constraints may be an issue for lower-income farms (Gabriel and Gandorfer, 2023; Palma-Molina et al., 2023b; Losacco et al., 2025). This raises the question about the potential for improving the sustainability and profitability of smaller farms through increased technology adoption strategies.

Specifically, this study found significant differences in the average herd size and farm size between the adopters and non-adopters of ACR. These differences could be due to the importance of managing workloads on relatively larger farms. Although focused on milking robot adoption, Gargiulo et al. (2018) found that farms in the top 25% for performance were optimising automated milking systems to manage larger herds. Concurrently, there appears to be scope, as concluded by Garcia-Covarrubias et al. (2023), for smaller, less intensive farms to increase their efficiency by using technologies such as ACR and automatic manure scrapers.

Findings show that higher farm income is associated with adoption of some technologies. This confirms studies by Garcia-Covarrubias et al. (2023) and Palma-Molina et al. (2023a) that farmers with stronger economic performance (higher FFI/ha) were more likely to adopt technologies like ACR, seemingly leveraging their capital to invest in tools that help to manage larger scales. Affordability of adopting certain technologies may be an issue for lower income farms (Stevens et al., 2021; Kopler et al., 2023; Mgendi, 2024; European Commission Joint Research Centre, 2025). Interestingly, however, the mean difference in FFI/ha appears notably high between the adopters and non-adopters of the app for grassland management (PBI), underscoring that income variability does not necessarily explain the differences between groups.

A farm’s proportion of hired labour showed a significant positive association with adoption across all six categories. Existing literature offers varying findings, some demonstrating that adopters may increasingly adopt precision and automated technologies to reduce labour demands, helping to manage the increased workload of larger farms (Yang et al., 2021; Hogan et al., 2022b; Garcia-Covarrubias et al., 2023; Hogan et al., 2023).

Correspondingly, farmers operating under the Nitrates Derogation, typically large, intensive operations with higher stocking rates, showed a universally significant positive association with technology implementation in this study. The positive association of certain technologies like ACR with farmers in derogation is logical given this technology is a key automated technology used in the milking process, and automation helps provide consistency, essential for managing larger herds. The technology with the highest association with derogation status was the app for grassland management (PBI). PBI, a national web-based tool helping farmers manage their grass growth and utilisation (Hanrahan et al., 2017), is particularly relevant to derogation farms as it supports farmers in efforts to precisely manage nitrogen, required for compliance under the Nitrates Derogation (O’Donovan et al., 2021; O’Donovan and Maher, 2026).

Sociodemographic factors like age and gender were also associated with adoption, though complexities remain. Adopters of automatic milk machine washers were significantly younger than non-adopters, an association which may support literature (Regan, 2019; Kopler et al., 2023), which discusses the idea that new entrants are looking to technology to reduce the physicality of milking related tasks, aided by the adoption of cleaning technologies. This study found that adopters of management tools such as apps for grassland and herd management are also younger than non-adopters. These results confirm outcomes of other studies showing younger farmers were more likely to adopt reproduction and grassland management technologies (Palma-Molina et al., 2023b; Tur Cardona et al., 2025). This cohort is essential for ‘sustainable transition’, appearing to benefit from policies promoting generational renewal and investment (European Commission Joint Research Centre, 2025; Loughrey et al., 2025).

Gender was found not to be a significant determinant for the adoption of most technologies, except for ACR, which showed a significant association. However, this result could be due to the sample having few female farmers (9 female farmers vs. 240 male farmers). General studies suggest that gender is often not a significant determinant in overall digital technology adoption (Irish Farmer’s Association (IFA), 2019; European Commission Joint Research Centre, 2025). Some PLF studies (Hostiou et al., 2017; Kling-Eveillard et al., 2020) indicate women’s involvement in animal husbandry and using automated technologies increases once smart technologies (PLF) are adopted on farm.

While high agricultural education is often linked to adopting a larger number of technologies per farm (Palma-Molina et al., 2023b), it was not directly associated with individual technology adoption in this study. The literature suggests that while formal education may not be the sole determinant for a single technology investment, high-adopting farms generally possess the practical knowledge necessary to manage such complex technologies and develop strong business plans (Conseil Européen des Jeunes Agriculteurs (CEJA), 2023; Palma-Molina et al., 2023b). Other literature (European Commission Joint Research Centre, 2025) shows that the association between general education levels varies across crop and livestock digital technology adoption while finding farmers with agricultural training adopted livestock-specific technologies at a 21%–24% higher rate. In Ireland, a rolling average statistic (2022–2024) shows 87% of Irish dairy farmers having received some form of agricultural education, compared with 62% for all farmers (Buckley and Donnellan, 2025), suggesting more research could increase understanding about education’s impact on technology adoption trends in Irish dairying.

4.3 Benefit and barrier factors associated with technology adoption

Key insights emerged from study survey questions, which focused on farmers’ self-reported reasons for perceiving technologies as either being very beneficial or having multiple barriers to adoption. Although focused on a subset of smart dairy farming technologies, the data from farmers’ perspectives have allowed for a deeper understanding of the reasons for adoption or non-adoption. Getting farmers’ perspectives are critical to this process, underscored by European-level reports (European Commission, 2025; European Commission Joint Research Centre, 2025) about the importance of direct farmer engagement. Other livestock technology adoption studies (Stevens et al., 2021; Dayoub et al., 2024) further emphasise that without farmer (end-user) input, technology may not address farmers’ needs and their willingness to invest may be compromised.

Farmers perceived varying benefits associated with specific technologies. This study found that the importance of labour dimensions came to the fore in the selection of the most beneficial technologies, particularly for ACR, the technology selected by more farmers when indicating benefit factors. These findings correspond to those found elsewhere in the literature in relation to automatic calf feeders and labour efficiency (Sinnott et al., 2021) and automatic cluster removers and labour efficiency (Hogan et al., 2022b). However, findings related to milking robots, the technology selected by more farmers when indicating barriers to adoption, vary from the wider literature. Findings regarding perceived financial barriers behind non-adoption are supported by others (Shortall et al., 2015; Shalloo et al., 2021). However, outcomes from this study should be understood in the context of the limited uptake of milking robots in Ireland (estimated at ~4%), with this sample containing less than 2% of farmers who had adopted milking robots.

4.3.1 Perceived benefits of automatic cluster remover adoption

Automatic cluster removers were the most popular choice by farmers when considering benefit factors associated with technology adoption. This result makes sense as ACRs are a well-known technology, being adopted by over half (55%) of Irish dairy farmers with adoption rates rising, increasing from a 32% adoption rate in 2018 (Palma-Molina et al., 2023b). Cluster analysis of the benefits of adopting ACR showed that most respondents perceived animal health/welfare factors as strongly associated with their adoption of this technology. Findings underscore previous research which highlighted the positive contribution of ACR adoption to animal health (Wieland et al., 2020).

While animal health/welfare was positively associated with the adoption of ACR, financial benefits appeared less important for most of the technologies considered, including automatic cluster removers, automatic calf feeders and automatic cow ID systems. Furthermore, cluster analysis of ACR benefits indicated farmers’ perceived environmental benefits as having a limited role in their decision to adopt a chosen technology, with the notable exception of PastureBase Ireland. Perhaps this outcome is related to how farmers interpreted and responded to the two environmental factors that were grouped together (“improves grass growth” and “helps meets environmental targets”). Farmers may have perceived the environmental sub-dimension, “improves grass growth”, as unrelated to the adoption of ACR, and so they moved on to the next section instead of decoupling the “helps meet environmental targets” factor. However, both environmental factors were perceived as directly related to the adoption of PBI, indicating farmers perceived both environmental sub-dimensions as associated with its adoption.

Reported low rates of environmental benefits for selected technologies are of concern, suggesting a critical need for clearer communication regarding the role of smart technologies in supporting environmental sustainability targets. However, there could well be some positive environmental benefits from these technologies, which farmers have not considered influential in decision-making. For instance, the use of ACR can reduce milking duration (Prendergast et al., 2024), thereby reducing electricity consumption. Automatic calf feeders can help reduce feed waste and thereby reduce overall GHG emissions. Relative to some other technologies (e.g. protected urea), these technologies may have a proportionally small influence on environmental sustainability but are nonetheless important.

4.3.2 Perceived barriers of milking robot adoption

Milking robots emerged as the technology with the most perceived barriers to adoption, with 67.7% of farms identifying them as problematic. For many farmers, the primary deterrent to not adopting milking robots was financial, specifically factors of “high investment cost” and “lack of awareness of financial supports”. These reasons are not surprising given the significant capital investment required for milking robots (typically over €100,000/robot), supporting other studies that showed financial barriers impede milking robot adoption for farmers in Ireland (Shalloo et al., 2021) and New Zealand, who also operate in an extensive grass-based system (Omoyajowo, 2024).

Interestingly, these farmers did not consider milking robots as creating uncertainty in meeting environmental targets, supporting varied positive and negative environmental effects of adopting AMS by Norwegian dairy farmers highlighted by Martinsson et al. (2024). Some farmers considered labour barriers as important in deciding not-to-adopt milking robots, i.e. “don’t think it will reduce the number of workers needed”, a concern particularly relevant to Irish farms reliant on family labour. Given that milking-related tasks consume a large portion of labour input (O’Donovan et al., 2008), the potential for milking robots to reduce labour requirements is normally considered to be one of the strengths of that technology (Shortall et al., 2015). Milking robots can reduce time spent on milking tasks and offer increased flexibility, changing task management on farms (Shortall et al., 2015; Rodenburg, 2017). However, these benefits may not be very applicable to the many dairy farmers in Ireland reliant on the main farm operator and one or two other family members to undertake milking-related tasks.

Also, some perceived low levels of social support as barriers when considering milking robot investment. This result underscores other research that highlights that low levels of technology learning supports by technology providers and/or farm advisors can impede farmers’ willingness to adopt some technologies (Drewry et al., 2019; Omoyajowo, 2024). Similarly, livestock farmers surveyed by Hartung et al. (2017) reported that insufficient maintenance support by technology providers post-installation is also a drawback to the use of some precision livestock technologies.

Overall, while there is scope for increasing adoption—particularly for technologies posing the least financial strain—decisions remain linked to a variety of financial and non-financial behavioural factors (van der Weerdt et al., 2024).

5 Conclusion

This study provides insights into the adoption of ‘smart’ dairy farming technologies by Irish dairy farmers in 2024. Results offer evaluation of the perceived benefits and barriers in the adoption process for dairy farmers operating largely within a pasture-based system. Through direct input from farmers during the 2024 NFS AS, it is evident that adoption of all types of surveyed ‘smart’ technologies (automated, precision, digital and electronic) by Irish dairy farmers experienced a positive upward trend from 2018 to 2024, confirming wider technology adoption trends in Europe.

However, the technology adoption landscape remains varied, with adoption rates differing between smart dairy farming technology categories and individual tools. The results from sociodemographic data analysis align with existing literature, indicating that dairy farmers with larger herds and farm areas and greater proportions of hired labour are more likely to adopt smart dairy farming technologies. Furthermore, Irish dairy farmers operating under Nitrates Derogation status exhibit a positive association with higher smart dairy farming technology adoption rates, across functional categories. This analysis offers an innovative benchmark of the association between Irish dairy farmers and derogation status for future studies. On the other hand, the study produced mixed results regarding the association of smart dairy technology adoption with family farm income, age, gender and agricultural education. Perceived benefits and barriers to adoption are also diverse, with decisions to adopt or not adopt smart dairy farming technologies shaped by an interplay of multiple behavioural factors (economic, social and environmental), as well as personal and cultural factors. Animal health/welfare were perceived as the most significant benefits of technology adoption across the four provided technologies, including in the specific analysis on ACR adoption. Financial and environmental factors were identified as primary barriers for the four provided technologies, with financial factors perceived as an important barrier for milking robot adoption.

Farmers’ willingness to invest is currently tempered by pressing and complex geopolitical instability, policy uncertainties and climate change. In this anxious atmosphere, the role of smart technologies in the Irish dairy sector represents opportunities for some farmers. Smart dairy technologies offer farmers potential support through automation, which can reduce time spent on milking tasks. The provision of animal health data can assist with early disease detection, and other data available through grass and financial management apps can improve farm-level decision-making. Given the study design, conclusions about deeper motivations behind these decisions could not be drawn. However, results point to the multiple determinants associated with technology adoption, supporting opportunities for further research.

In response to the need for a resilient and sustainable agricultural sector, the EU has prioritised the digitalisation of European agriculture. For the livestock sector, the EU is committed to the role of digitalisation to enhance the sector’s dual needs for competitiveness and sustainability. The Vision for the Future of Agriculture and Food (European Commission, 2025) calls for additional financial and training supports for digital literacy and upskilling. The EU Cohesion Policy promotes integration and harmonisation of digital solutions for social progress, building on its historical focus on infrastructure development. The EU is positioned to continue supporting farmers’ uptake of digital solutions through subsides for investments, research and innovation (European Union, 2026).

While potential benefits exist for farmers as they increasingly adopt a variety of smart dairy technologies, many opportunities remain to help farmers understand the potential benefits and realistic challenges of adopting a particular smart dairy technology for use on their farm. Farmers will continue to need practical, proven innovations that help them remain competitive, sustainable and resilient in the face of change.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fanim.2026.1896469/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Charlotte Reed, DairyNZ, New Zealand

Reviewed by: Sindiso Mpenyu Nleya, National University of Science and Technology, Zimbabwe

Brian Dela Rue, DairyNZ, New Zealand