J.O. Giordano, A. Laplacette
Artificial intelligence-driven tools could transform reproductive monitoring and management. A full reproductive cycle of cattle requires completion of several biological processes and events such as estrus, ovulation, pregnancy, and parturition that could be monitored and predicted by Artificial intelligence tools. Reproductive monitoring and management tasks and decision-making could also be enhanced or fully executed by Artificial intelligence tools. Artificial intelligence-driven tools for reproductive monitoring and management are emerging, with most at a nascent stage and created based on machine learning algorithms. Development of effective AI-driven tools requires a better understanding of animal biology through data and the consequences of management and environmental factors on data used by Artificial intelligence models. Successful reproductive monitoring and management of cattle is the result of a complex interplay between animal biology and execution of several monitoring and management tasks. The reproductive cycle of female cattle involves physiological and endocrine events enabling conception, gestation, and parturition, after which lactation begins. Parturition, recovery of uterine health, return to normal ovarian cyclicity, expression of estrous behavior, ovulation, and pregnancy are the result of a series cyclical endocrine, behavioral, and physiological patterns. Monitoring and managing female cattle through these complex biological processes requires completion of several tasks, interventions, and decisions usually executed by personnel with technical knowledge and a specific skillset. The diversity, complexity, and temporal dynamics of events and interventions during the reproductive cycle coupled with advances in automation and data-driven technology has primed the field of reproduction for disruptive artificial intelligence (AI) tools. Although most tools for AI-driven decision-making and semi- or full automation of management tasks including detection of estrus, synchronization of ovulation, reproductive program selection, and cow reproductive status monitoring are at a nascent stage or under development, these tools are likely to transform reproduction in dairy farms. As for most disciplines in the animal sciences, the field of cattle reproduction has traditionally relied on conventional statistical methods. Mathematical models including linear and generalized linear models, mixed effect models, and time-to-event models were used to explore and test relationships between variables of interest and infer or predict outcomes of interest. This approach provides easily interpretable results, but major limitations included the inability to process large and complex datasets with diverse types of data (e.g., time-series and cross-sectional variables, imaging data) or handle nonlinear relationships between variables. These limitations coupled with the expectation of improved performance of AI models and a surge of data-driven technologies for farming have fueled unprecedented growth in the exploration of AI-driven tools for cattle reproduction. Among other reasons, AI is expected to outperform traditional methods because of its flexibility to work with diverse predictors in “big data”, handle complex relationships in the data, and ignore a priori assumptions of relationship between variables (Sarker, 2021; Heil et al., 2023; Ratkovic, 2023). AI is a field of computer science that enables machines to mimic human intelligence through data-driven methods that use observations as examples to learn how to understand, classify, or make predictions from data. Machine learning is a branch of AI that consists of providing machines with the ability to automatically learn without being explicitly programmed (Samuel, 1959). Although different ways to categorize AI methods exist, machine learning algorithms (MLA) are typically classified as supervised, unsupervised, and reinforcement learning. Supervised learning models use labeled data to learn and tend to be used for classification or regression tasks (Ayodele, 2010). This type of algorithm has been the most explored for application in cattle reproduction. Linear classifiers, logistic regression (LR), support vector machine (SVM), decision trees (DT), random forest (RF), boosting, neural networks (NN), and Bayesian networks (BN) are examples of the most popular supervised MLA used for classification tasks (e.g., estrus vs. no estrus, pregnant vs. nonpregnant) in reproduction. This group of algorithms could also be classified into two main groups. Classic or nondeep MLA rely on human participation to perform feature engineering and to signal differences between variables. On the other hand, deep learning is a subfield of ML based on artificial neural networks. These networks include multiple layers that enable the algorithm to deal with increasingly complex and larger amounts of data such as the time series, high frequency data generated by wearables and nonwearables sensors for cattle (e.g., activity, temperature, rumination, milk yield, body weight sensors). Deep learning models can automatically learn features and interactions between variables even from unstructured data at the expense of requiring large amounts of data to train the networks. More recently, a growing number of AI methods, including convolutional neural networks (CNN) and large language models (LLM) began to be explored for tasks including interpretation of image data such as ultrasonography scans, image-based detection of reproductive events, and dairy records mining (Andrade et al., 2023; Wang et al., 2024; Gontijo et al., 2025). Although these forms of AI might be more powerful, little is known about their value in reproductive monitoring and management because limited research has been conducted with these relatively novel methods. The goal of this review is to explore some of the key opportunities for application of AI-driven tools in dairy cattle reproductive monitoring and management in commercial dairy farming systems using traditional management practices. The review focuses on the type and characteristics of data and tools that will enable advances in reproduction in dairy systems. Special emphasis is placed on research that has already explored and paved the way for using AI-driven methods to monitor reproductive events and status, and to predict the reproductive potential of cows. Finally, challenges encountered during development of AI tools and their subsequent on-farm implementation are discussed. The opportunities to develop AI-driven tools for reproduction are growing rapidly as the number, diversity, and complexity of technologies available and data streams generated in dairy farms expand and diversify. Numerous wearable and nonwearable sensors, herd management and environmental are or will be available for commercial the reproductive wearable and nonwearable sensors a of cow and performance for monitoring reproductive events (e.g., and status (e.g., of or reproductive potential A deep from the of data generated at high frequency and from and the of and technologies and data used for development and implementation of AI-driven tools for monitoring reproductive events and status through the reproductive cycle of dairy and Numerous behavioral, and performance and features can be with herd management and performance records and environmental data to AI-driven that personnel or herd monitoring and management tasks including detection of estrus, pregnancy, and of and technologies and data used for development and implementation of AI-driven tools for the reproductive potential of dairy a reproductive behavioral, and performance and features can be with herd management and performance records and environmental data to AI-driven predictions to cow and herd management interventions and of estrus expression and reproductive the of pregnancy and can be and predicted for interventions on the stage of the reproductive artificial at artificial or which a of physiological and events, in of data or more the and physiological by during reproductive events and different of the reproductive the of and data can that the reproductive cycle as is about events, and more data is generated at high frequency and the available for cow reproductive cycle and data for the and the available for pregnancy to can data and the The more biological a but also a larger and more diverse group of features with potential value of cow reproductive and decision-making in the reproductive cycle of cattle for implementation of AI-driven tools based on the deep generated from the multiple of high data by data-driven tools and monitoring These could to and tasks and decisions between for and from to in or the for personnel these tasks can a growing and the of interventions on cows. reproductive events and outcomes can or to herd management and and management differences between and including physiological to the of milk data for and for and different types of and of on data-driven technologies for challenges and opportunities for development and implementation of AI tools for and cows. at the AI models based on behavioral, and computer data could be used to predict the time of a that enables to at This could of of by using the or data could also predict events and monitor interventions to in lactation is with reproductive performance et al., 2023). and monitoring events after could at of reproductive and enable implementation of to the of on reproductive The which the time from for pregnancy, provides opportunities for AI tools. models based on data during this in monitoring of reproductive physiological status could enable models that to reproductive management et al., Deep of recovery of uterine health, return to ovarian and estrus, and to lactation could be used or with herd management features and environmental data to develop models that effective management for a of the and of pregnancy for or of are decisions with consequences to herd and data through the could be used to predict and of pregnancy that reproductive performance and AI could data for return to estrus after of estrus and number of estrus to predict the of pregnancy with of based on milk and lactation to detection of estrus is or be the of the AI tools could monitoring and management tasks such as detection of estrus and of synchronization of to and of to or Although is of the of models to in estrus based on et al., et al., other into AI models could that predict the of estrus based on features (e.g., of be for decisions at the time of or endocrine and of and activity, rumination, and could be used to the expected of a cow in estrus by estrus detection As the for from dairy cattle opportunities to use AI tools at or the time of also AI-driven systems could be to from the of methods to a artificial with conventional or using or or are available for dairy farms. The type of could also be by AI tools that and the of for value or a could the to a female dairy using based on the of the and available for the cow predicted of pregnancy in with data for herd and data the time of and after pregnancy through artificial or can also opportunities for implementation of AI-driven endocrine could support implementation of and and after pregnancy that could or support development could be through cow or group (e.g., could be to of through (e.g., or or through of ovarian or pregnancy status monitoring through milk and physiological and ultrasonography is major of AI-driven tools could and monitor pregnancy and or pregnancy status at time or challenges of traditional pregnancy methods. models could A of application of tools is the detection of events that reproductive management interventions and estrus events to of pregnancy status and interventions in the and a to a and estrus such as or completion of a of or or wearable tools can personnel to monitoring tools could monitor et MLA Linear and to predict time using activity, and by and The ability to predict on the of with algorithms that to to to to et also that a on features created from of and the expected of to and between and Although in the is for to in the to is of because could result in to or the high performance of some of the algorithms based on or body temperature, research is to develop AI-driven tools that approach and for these are to be used in dairy farms. is a technology for monitoring but has been The in and of and during are through et that a neural on of multiple including and these outcomes of the the algorithm for and systems for for commercial application challenges including and several of estrous has been for more a of estrus detection systems based on wearable sensors that monitor and physiological are available for commercial farms et al., et al., The in with estrus is and of such in most that estrus detection has been with estrus created without Although this might interest in AI-driven are opportunities to estrus detection and multiple and physiological with herd and environmental data could or in in some result in management interventions such as or other that a in the of that during estrus and behavior, body temperature, and milk are for into estrus detection algorithms et al., data such as or other that could might also be of AI-driven estrus detection algorithms including cow et al., multiple features et al., Wang et al., and and et al., have been A of MLA including supervised and and deep learning algorithms were and et al., et al., Wang et al., Although this algorithms were effective et al., et al., Wang et al., the of performance that for of the traditional algorithms et al., novel AI algorithms and were conducted in research research on their value with large and under commercial systems could a major in detection of estrus as this technology could or wearable Wang et for in estrus based on a that used the The improved to cow and and in a with to with The ability to in estrus based on which potential for This and most other that explored for detection of estrus and et al., were conducted in research under with limited number of and for relatively of to be advances in will challenges in commercial farms including high in and of cows. including data and also be for effective implementation in commercial farms. status and detection of pregnancy are key management tasks. to pregnancy using data-driven technologies is of this the milk have been used to train MLA to predict pregnancy status based on in features with et models using milk differences from milk on data to after the performance for on-farm et used deep learning to predict pregnancy status from milk data. The of of and of that models on milk could be a to predict pregnancy Although this approach to pregnancy might traditional methods as performance at of has been milk has been for other such as pregnancy or et al., Although some of the pregnancy status and models on milk data generated results, were in to support reproductive novel methods and other and physiological data by biological pregnancy to into MLA might to the development of more effective tools. with different reproductive potential in of key of the reproductive cycle or execution of management tasks is for reproductive management the of estrus after the of the the of pregnancy to or and the of of pregnancy on can enable implementation of for or of with different reproductive potential et al., novel on-farm technologies and in the field of the and of data and methods to predict reproductive outcomes have rapidly farms have to enhanced predictions of and high frequency data from sensors that complex reproductive and physiological in dairy are herd management and performance records and environmental monitoring tools. to novel data streams and is the way for major in development of and AI-driven tools for interventions and after interventions after pregnancy and through the lactation AI-driven models to or of with different reproductive potential is a goal the of of between behavioral, and performance with cow reproductive potential et al., et al., et al., 2023; et al., 2025). The expectation is that AI models can be more effective traditional at and based on multiple features to multiple These include complex between of biological processes and reproductive biological of the temporal dynamics by some and the value of used research on MLA to predict the of using and data because data from wearable and nonwearable systems or were More recently, a growing number and of cow and data is being used to train to use MLA to predict the of et MLA and with to health, and and cow features from farms that in a of systems. The predictions were for models based on algorithm of and for and of and and in milk and number of and events data were the most Although relatively to most models in the the performance by et were for models through using datasets and ML including methods that temporal and to other from to predictions were this is of the most predictors of reproductive potential because of the of data generated through milk et predicted the of to with MLA and on data for multiple and factors and Although models with from to to of milk data to the value of the models. More recently, et multiple MLA models to of with high using a of milk and and other milk features from The in of and models as as for and models and as high as for the These were a with et and is value of data for of pregnancy As more on-farm milk technologies available will be more opportunities to explore the use of milk and milk features to predict As behavioral, and performance data from monitoring systems using wearable and nonwearable sensors more the of these data into models based on MLA has been from these systems has been with the of reproductive et al., et al., et al., 2023; et al., 2025). has that records of estrus and as as features from the of several activity, rumination, by during the can more likely to estrus and explore the value of high frequency and data for the of the in dairy used data from from a commercial herd in the to train several MLA models et al., data from cow with features from several and temperature, body and milk features yield, and in and to and in multiple or were used to train several The models in the of to which relatively These that including features from into MLA models the ability to predict pregnancy with other types of data. these be with because of a limited number of observations available in the potential of reproductive potential differences to the use of with a and no to use more complex methods to the performance of models. from et with more from et used data from and data from cow from with in the and to predict the of expression of estrus after the of the and pregnancy to or features from data for milk including yield, and in the and and with other cow features and performance data were used to train models. The for the for the models including data or the data with wearable sensors and cow features were to for expression of estrus, and and for in the test Although to from other MLA models using more and data, models data have et better for models that predicted the of pregnancy using data from The linear and algorithms in the of to is to the high ability of these MLA to the ability of the algorithms to from the data or the result of including available to the algorithms included estrus and estrus of which are with the of and but available the decisions by AI tools with data from or several algorithms including data available on the of or a AI are for decision-making at the time of or after such tools could be used to the type of used for or a but enable the for synchronization of estrus or the of the most models for of different reproductive potential in the have been to farms or their performance has been in commercial farms. are two examples of models using MLA that were in commercial farms. these algorithms are and the methods used to the models have been fully et that from commercial farms in and based on a that predicted to pregnancy based on MLA differences for several reproductive in different a different to that at estrus or The algorithm used for included and and with data from predicted and from the for and from a et also that a created based on algorithm on such as activity, rumination, and other features with different reproductive on the methods used to this algorithm These two examples the of AI tools for with different reproductive potential for used in A of these two algorithms is that the potential have been as and data were these type of data into MLA models to predict reproductive potential is of the of models to predict the of using a of and performance from predictions for and several other cow features and 2025). Although predictions for were as high as expected in of predicted of pregnancy from the MLA models in differences in (e.g., to between the and group for and AI-driven tools based on MLA on diverse datasets is and rapidly of research in dairy cattle reproduction. observations from the available are no and methods to MLA for development of models. for which has been used in most and the use of using data, is no use of MLA in research This of and of MLA use but also to most or effective and methods, and the type of data to predict reproductive This is increasingly complex as the and number of and the type and of data available to A in research has been to test several MLA to the most type of to of MLA such as ML were explored for other in dairy et al., 2025). limitations and the limited that have of these tools can the value of this from the is that most of the research has been to nondeep learning in dairy cattle reproduction with other novel and AI methods including deep and other forms of AI is or at its nascent This is likely to a of the that of these AI methods are relatively for data their research has also that multiple of high frequency and data for the of and environmental factors that or are with reproductive is to predict outcomes of biological events and management interventions for (e.g., predict a cow will estrus, pregnant Although in AI methods and of data could this based on of predicted from MLA models is a with potential value for application of AI-driven tools consists of based on of cow predicted from models that These can be based on and opportunities to herd performance and management practices. As of as or can be that expected differences between are to implementation of a herd or could be into of predicted of pregnancy to of and the has differences in outcomes such as AI be for development and implementation of AI tools to predict the reproductive potential of dairy and for reproductive management. in the process include and data for cow and herd behavioral, physiological and performance features and events, and environmental from different including and wearable and Machine learning models are to predict the of or of interest (e.g., estrus, pregnancy, pregnancy for cows. for predicted from MLA are used to of with differences in reproductive potential for implementation of reproductive management can be into with expected and AI based on the generated by a machine learning algorithm machine learning This review several other of reproductive monitoring and management for application of AI tools. MLA for different of reproductive technologies for cattle such as et al., et al., et al., and animal et al., has also been systems are already available for in et al., could be and have been explored for and et al., 2025). to pregnancy and reproductive and ovarian based on ultrasonography have been and et al., 2025). are several to AI-driven tools can cattle reproduction. in reproductive management AI tools are more to AI performance can be and tools be easily generalized because of and their to to or pregnancy methods pregnancy methods, and the of methods are to AI tools to the of management and AI-driven models that can be generalized but also to specific can to these challenges including data, in different of data, AI methods, and of fully time data tools to and AI and will be to these challenges including novel for data and data of data methods to and implementation of data tools et al., 2023; et al., et al., 2025). The is as AI tools are likely to use data from a of and data streams from and the this and data between dairy herd management or data tools with other data systems by multiple will be to enable AI tools. This can be through of data of data between in data and a more diverse of data systems from or or limited to data for at or research and AI development requires large and diverse datasets that in are to of the methods and models are to that data-driven are to fully to and most of these AI tools. The and in more to the development of effective and AI tools. understanding of how data cow reproductive biology and the of management and the on are is to understanding of cow biology through data used or generated by AI-driven tools. or of these is likely to the of AI methods for reproduction. The opportunities to reproductive monitoring and management through AI-driven tools are and multiple physiological events and interventions to a reproductive AI-driven tools can to monitoring and managing through these events, or interventions, and or Among the and growing of AI methods and tools MLA have been used and explored most AI-driven algorithms have been for detection of events, estrus, and pregnancy status MLA models have been to predict cow reproductive potential with of The AI tools a of performance and of to limited to farms. A of work to be to reproductive monitoring and management with AI-driven tools. by surge in the number and types of data-driven and rapidly AI methods, the field of cattle reproduction is primed to be by is a in the of at and in dairy cattle biology and management systems. as of the and of the for and of the and for the of the research animal and data to reproductive health, and of dairy cattle through research focuses on reproductive biology and management of dairy including of the estrous cycle to and development of to and detection of has the development and of monitoring management and data-driven tools that of reproductive and management reproductive machine and systems for dairy systems and enabling to make decisions that and of from the of in of in from the of and in from the of has more and and and and more in research is to and in dairy herd reproductive and with to into that for is a in the of at and a of the and under the of of from the of and of in at research focuses on management to reproductive management and performance in dairy through work the development of machine learning and traditional tools to of with reproductive enabling management to animal research include and that and reproductive performance by estrous detection with synchronization of This approach reproductive data and technologies to for dairy has more and and more at to the between methods and on-farm tools that enable dairy to make decisions about reproductive management. This for by the of The in this are of the and the or of the of the or the of interest The no or of interest. and The and research from the in this are by the of and of the and by the for and and also by the of and program and or in this are of the and the of the of and or the of in this review also by and