application of AI in the industry has been increasingly optimizing its results, in an attempt to reach its maximum degree of efficiency. AI advent is the arrangement of several technologies, which allow machines digitally to perceive, understand, act, and learn on their own actions or complement human activities, which has become a broad technology used for machine learning, predictive analytics, augmented reality, robotics, performance diagnostic software, and many others.
With entire procedures performed by machines capable of making decisions based on data, agility and increased productivity are natural consequences. Through AI, industrial production has become faster and more effective compared to human labor. Still considering the possibility of the machines performing tasks that a person would not be able to do, as is the case with dangerous raw materials or microscopic components.
AI works through the integration of factors such as the use of IoT sensors, Cloud Computing, and other technologies present in Industry 4.0, working in sync, devices equipped with AI create complex systems, which correlate the information collected and, with this, seek the best ways to carry out the activities for which they were scheduled. These new technologies are developed to work using the least amount of resources possible, whether in terms of raw material or energy consumption, still relating the point of cost reduction, the mitigation of errors, and waste of the operation.
When addressing AI applications, it is worth mentioning IIoT as a critical technological layer added to the production chain, which allows even the connection with suppliers and the analysis of the performance of its raw materials, still pondering the potential of AI in relation to security alerts, which point to the need for maintenance and performance reports in real time, indicating the best measures to be taken.
Still pondering the aspect in which machines can withstand extreme conditions that would be harmful even if perceived only in the long term for the health of the employees of industry, such as cooling cameras, chemical processes, and management of explosive materials, among others, that can be carried out almost entirely through automation.
The aspects in Industry 4.0 in relation to the digitization processes that guarantee the collection of data that were previously lost, the mitigation of risks in decision making, the optimization of operations, and the gain of agility, among others, are also mentioned. The implementation of complex AI algorithms has been enabling industries to assess and enable problem-solving and decision-making in a more complex and secure way.
Assessing that each sector of the industry receives contributions from AI in a different way, as a logistics and inventory structure can benefit from technology for identification and control of demand, for example; or industries with production chains that have different machinery, as is the case with the automotive industry, since with the use of predictive analysis, they can identify the need for maintenance on their machines.
The benefits are not the only ones since the industry receives an extremely positive impact on the use of AI in its processes. Given that it is possible to point out an increase in the quality of products and services, since AI reduces execution errors and subsequently uses operation data to analyze performance and make improvements; or even more effective new products and services, since the development can also be supported by AI to evaluate the proposed designs, identifying the material variables, the weaknesses to be improved, and the possibility of using augmented reality to make tests before actually putting it into production; or even through data analysis, it is possible to get an agile response to new market demands, considering that the needs and interests of consumers are changing with great velocity.
AI brings great advantages to the industry related to the reduction of errors, because after being trained, intelligent algorithms are able to perform very well tasks that are susceptible to errors in processes executed by humans. The reduction of costs since e-commerce stores or banks use robots (chatbot) for customer service, this allows employees to be allocated in more strategic areas, which can increase profit. So, with fewer errors and employees focused on more important processes, the company will have more time to think about the business and leave other tasks to AI.
Thus, AI through an automated process uses large volumes of data to make decisions, dispensing with human intervention and increasing productivity in different activities.
1.5 Trends
Adaptive Intelligence is about helping to generate better business decisions by integrating the computational power of internal and external data in real time with the computing infrastructure and highly scalable decision science. In this type of systems, relating the adaptive learning, the characteristics are monitored so that there is an adjustment in order to improve the process. The efficiency of these systems depends on methodologies adopted to collect and diagnose information related to needs and characteristics, in relation to how this information is processed to develop an adaptive context. These applications essentially make businesses smarter, allowing them to provide customers with better products, recommendations, and digital services, all of which generating better business results [55].
Digital twins are related to the practice of creating a computer model of an object, such as a machine or even a human organ, or yet a process like a climate. By studying the behavior of the digital twin, it is possible to analyze, understand, and predict the behavior of its counterpart in the real world and to solve issues before they occur. However, to take full advantage of the digital twin’s potential, real systems need not only be networked with each other but also need to develop the ability to “think” and act autonomously [56].
This development tends toward AI, from simple mutual perception and interaction to independent communication and optimization, also requiring integrated information systems that allow a continuous exchange of information, still demanding powerful software systems that can implement them along the entire value chain, and planning and designing products, machines, and plants, in addition to operating products and production systems. The technology of digital twins allows users to act in a much more flexible and efficient way, as well as personalize their manufacturing [57].
Intelligent Edge refers to the place where data is digitally generated, interpreted, analyzed, and treated, i.e., the use of this technology means that analyses can be managed more quickly and that the probability of data being unduly intercepted or leaked is considerably less. This technology refers to the analysis of data and the development of solutions in the place where the data is generated, reducing latency, costs, and security risks, making associated businesses more efficient, still pondering that the three largest categories of Intelligent Edge are the edges of operational technologies, IoT edges, and IT edges [58].
The use of Intelligent Edge technology helps to maximize business efficiency, since instead of sending data to a data center or even to a third party to perform processing, the analysis is performed at the location where the data is generated. This means not only that the analysis can be performed more quickly, but it also means that companies are much more self-sufficient and do not depend on potentially flawed network connections to do their job [58].
Predictive maintenance is one of the most promising branches for industrial applications based on the use of data received from the factory to avoid production failures. This type of system eliminates unnecessary maintenance and increases the probability of avoiding failures, which involves a machine or even a component with sensors capable to collect and transmit data and then analyze it, and perform storage in a database. Then, this database offers comparison points for events, as they occur [59, 60].
The predictive maintenance model aims to periodically monitor the operation of machinery, equipment, and parts in a factory, in order to detect failures before they occur and prevent interruptions in the production line, relating IoT and AI in order to assist in the survey and management of data from all sectors of production, integrating the company’s departments, performing analyzes
