Do manufacturing companies that use AI to manage supply make more profit than manufacturing companies that don’t?
Author Note
Keywords: writing, template, sixth, edition, APA, format, style, self-discipline
1.0 INTRODUCTION
1.1 Broad Issues and Current Conditions
Over the years, the world has been progressing toward a new transformation, and Industry 4.0 technologies are seen as the way of the future (Agostini & Nosella, 2019). Artificial intelligence (AI) is among the most well-known of these technologies (which includes blockchain, IoT, big data, and others). AI is characterized by the ability of devices to interact with and emulate the capabilities of humans (Dick, 2019). When AI is used, decision-making becomes more accurate, faster, and with a more significant set of layers. AI is neither a new topic nor an academic area of study; nevertheless, technical advancements have only recently revealed that AI has a broad spectrum of uses, getting publicity by adapting systems in various fields (Jarrahi, 2018), particularly supply chain management (SCM). While some aspects of information technology are becoming a strategic necessity, AI is gaining traction as a competitive edge. To enhance their usability, several organizations are moving from remote monitoring to control, optimization, and, eventually, specialized autonomous AI-based systems (Prior & Keränen, 2020). Along with its growing prominence in the industry, AI has gained a more widespread presence in current literature. As a result, several fields, such as market analysis, have taken upon the subject, and AI is now studied from a more comprehensive perspective (Sousa et al., 2019). SCM is identified as one of the fields most likely to benefit from Application domains. Even though there is a lot of interest from practitioners and researchers (as shown by the plethora of studies on AI), further research is needed to see how AI can help SCM improve ROI.
Daily, logistics produces a considerable volume of information, primarily produced by managing and tracking massive flows of products (Fosso Wamba et al., 2018, p. 9). The data produced in this manner has a lot of room for improvement. The intelligent use of DataData is a significant challenge. The worth of DataData is determined by the applications that can be created for it, not by the amount of data obtained. The collected data must be designed so that it can serve as the foundation for optimization measures for this reason. The use of such data necessitates a comprehensive and reliable database. Due to progressively enhanced and less expensive sensor technology, data gathering, for instance, is no longer a significant obstacle. What matters is how this DataData is analyzed and how this analyzed data leads to improving the particular process. Another factor to consider is the amount of time it takes to analyze the collected data. Processing large quantities of data, such as that provided by IoT (Internet of Things) technologies, necessitates the use of a specific system to analyze these massive amounts of data (“Structure, Opportunities, Research Issues for Investigating Big IoT Data,” 2018). Enterprises must deal with an ever-increasing volume of data, which is becoming increasingly practical as big data frameworks are implemented (Gupta et al., 2018, p. 11).
Thus, it is usually no surprise that the application of Ai performs such a vital function in logistics. With its cross-sectional activity, Logistics is a critical element, rendering advanced analytics growingly a tactical tool. Identifying key performance indicators (KPIs) is a refreshing approach that enables various assessments and assessments (Leblanc, 2018). This type of data evaluation and representation is facilitated by the digital depiction of physical assets in a virtual environment. Especially in logistics, the utilization of real-time information is an essential method for envisioning activities instantly (Yagci Sokat et al., 2016, p.5). Nevertheless, industrial use of virtual representation structures can be observed primarily in product marketing, factory floor, and operations management. Moreover, Hopkins and Hawking note that there are inadequate real-life utilization instances in logistics both for IoT systems and data science (Hopkins & Hawking, 2018, p. 569).
1.2 Problem Definition
1.2.1 Formal Problem Statement
Contemporary supply chains are unique compared to the past, and they continue to develop in a highly competitive industry. The recent advancements include the establishment of pop-up storage facilities, ship-from-store designs, and other speed-oriented developments. Supply chains are pushed to successfully control increasing brand portfolios that can constitute many large numbers of stock maintaining units (SKUs). SKU quantities continue to increase as firms struggle to meet customer requirements for various sizes, shades, and other differences in product designs. Presently, just twelve percent of corporations implement AI in the supply chain, as per the latest study from MHI and Deloitte (“2020 MHI Annual Industry Report – Embracing the Digital Mindset”, 2021). Likened to the 2019’s edition of the similar survey, AI has indeed fallen a rating in aspects of increasing significance, consigned to the 4th spot by an added focus on sensors and Automated Data Capture. Artificial Intelligence in SCM is a market that is projected to increase at a rate of 46.3 percent to attain $22.6 billion within seven years.
Nevertheless, much of that expansion will be conditional on institutions’ responsibility to ensure that their supply lines are digitally capable. For firms that can input the capital to push the required digital transformation, innovations like Machine learning And Artificial Intelligence have the power to change almost every facet of supply chain management to stimulate profitability, cash flow, consumer satisfaction, and competitive edge. Inspired by these aspects, this study intends to examine if manufacturing companies that use AI to manage supply make more profit than manufacturing companies that don’t.
1.3 Study Purpose
The purpose of this research is to find out how big companies use AI in their supply chains and what the advantages of profitability are. When sustainability is a controversial concern in many countries, this study helps in understanding the current iteration of the application of AI in l manufacturing firms. Several news publications and journal articles have examined the execution of AI systems. This study aims to analyze what kinds of AI processes big businesses have enforced in their supply chain. The main objective of the study is to determine and analyze what types of AI technologies are used in supply chains and the current state of the field.
1.4 Research Questions
The study’s goal is to find answers to the following questions:
- What types of Artificial intelligence have manufacturing firms utilized in their supply chain management?
- what level of AI maturity have the companies embraced?
1.5 Research Paradigm
This study utilizes a Critical realism point of view. Critical Realism (CR) differentiates between the ‘real’ and ‘observable’ world (Critical realism, 2015). As per this viewpoint, the ‘real’ cannot be ascertained and distinct from individual viewpoints and hypotheses. As we recognize it, the universe is formed from people’s views and experiences, in what is noticeable. Consequently, as per the critical realism standpoint, unmeasurable structures cause measurable occurrences, and the social sphere can be comprehended only if individuals realize the structures that establish the occurrences. This study uses inductive logic, a method of reasoning that begins with a single observation sequence. The reasoning leads to typical arguments by integrating observation items. The layout of the theory is constructed based on evidence, and analysis components are not predetermined. The inductive model assumes that the investigator does not select what is relevant and does not test hypotheses or theories. Data focus necessitates self-discipline in accordance with the data, the elimination of preconceived notions, and the application of a systematic approach. While data-driven research appears to be intuitive and dependent, the investigator must focus on his or her behavior and determine the research’s validity. With qualifications, the reader gets details about the study history and validations during the analysis.
- LITERATURE REVIEW
- Introduction
Previous research and related literature on artificial intelligence and supply chain management are presented in this research section. It aims to provide some context for the definition of the topic area. The first section contains some context information on SCM. After that, the reader will have a good understanding of the critical points of SCM and will be able to concentrate on the results. The second section covers artificial intelligence and its various aspects to gain a broad understanding of the subject. The terms “supply chain,” “supply chain management,” and “operations management” were used to scan the documentation on SCM. The latest versions and materials specific to the subject area were used to evaluate the search results. Artificial intelligence in the SCM context was researched using keyword phrases that were important to the subject, for instance, “Artificial intelligence in supply chain management,.” The study must emphasize that when assessing AI capability, all of the findings are weighted equally.
1.2 Supply Chain Overview
Whether it’s high-end goods, client-tailored services, or typical and low-cost products, profitable companies have developed a concentrated and transparent understanding of value creation. Even so, no matter how good the advertisement is, if the product or service cannot be provided to the customer at a reasonable rate, no one will purchase it (Beamon, 2014). Many businesses should boost their SCM because their goods spend at least 6 to 12 months r or more in inventory. Since the goods spend too much time in inventory, there is a tremendous potential to enhance flexibility, cut prices, improve deliveries, shorten cycle times, and minimize waste properly. Internal processes have helped many businesses boost their supply chains. They’ve realized that it has a link to external clients and suppliers and that by using it, they can boost their operations even further (Li et al., 2020).
While the SCM concept first appeared in the early ’80s, it continued developing and gaining more popularity after the ’90s. The increasing interest in SCM can be attributed to two factors: first, increased globalization, which has created functional SCM potentials for companies, such as global distribution and manufacturing, and has exacerbated business challenges on a global scale (Giunipero et al., 2008). Second, there is a trend toward time and excellence-based competition, which necessitates improved coordination and reconciliation between the business and its vendors. Third, there is a great deal of uncertainty in the environment due to technological differences, an uncertain economic climate, and intense market challenges, necessitating a great deal of supply chain flexibility. The explosive expansion of SCM literature has been aided by some scopes such as purchasing and acquisition, shipping, resource regulation, trade, organizational behavior, infrastructure, and financial reporting (Dubey et al., 2017). The critical concepts of SC risk management, as per Laski (2017), are the identification and avoidance of risks in advance. In reality, it’s doubtful that every possible risk can be identified, so identification refers to the most significant ones that affect the SC. Executives are constantly mindful of their corporation’s challenges and threats, which arise in SC; nevertheless, identifying risks at the appropriate time is challenging. Supply chain risk can be described as possible roadblocks to the original target that, in turn, affect the reduction of profitable operations at different phases. The capability and quality of outputs at different points and times in the supply network will define the main processes. A service outage at any stage in SC could have varying degrees of impact on a few other processes. As a result of having consistently profitable operations, risk assessment has become an essential part of the supply chain planning framework. Requirements, evaluation, and identification of threats and hazards in supply chains, as well as addressing issues uniquely to mitigate collateral harm, are all part of risk assessment (Jabbarzadeh et al., 2018).
2.2.1 Decision Levels in the Supply Chain
The supply chain comprises 3 preparation decision stages in the SCM. There are strategic, tactical, and organizational levels. The distinction between these stages is the period of the relevant decisions (Rai, 2019). Strategic choices are typically taken over a lengthy period. At the strategic stage, choices must be made concerning premises where to serve production technologies and then choose the list of distributors to hire in the supply network (Ben & Krichen, 2020; Rai, 2019). Informatics networks are linked to strategic choices since it helps the SC processes and strategic alliances (Ben & Krichen, 2020). (Ben & Krichen, 2020). Therefore, strategic choices describe the supply chain in which assembly, production, and delivery support the marketplace (Baryannis et al., 2018) after the tactical choice stage precedes the strategic choice stage. Tactical level choices are medium-period choices, and the duration is from a month or two to 12 months (Baryannis et al., 2018). At this stage, the supply network is controlled to react on a tactical and operational scale. Such choices focus on consumers’ demands, and it passes to control and to plan procedure (Ben & Krichen, 2020). Production plan is normally given in this stage which is established on estimation and prediction (Baryannis et al., 2018). The strategic phase comprises actions that are planning choices focused on capacity and balancing cost. These activities include developing inventory, procurement contracts, and buying choices (Ben & Krichen, 2020).
If the tactical stage is established, the operational stage tackles comprehensive coordination, stock deployment, and deliveries (Ben & Krichen, 2020). The operational level works with the situations which are rendered at the strategic and tactical stages. That is to say; it offers a daily working and productive function. The operational stage presumes to make short-time choices, such as purchasing, delivery, and processing. The operational stage can be named “flow management” (Ben & Krichen, 2020).
The following figure display the time of preparation stages.
Figure 1: planning levels (Wassim et al., 2012)
2.3 An Overview of Artificial Intelligence
Artificial intelligence has piqued interest in the field of supply chain management in recent times. But since the late ’70s, AI growth has emphasized expanding market efficiency and interpreting business events and trends. Robotic processes and machine learning can perform time-consuming and repetitive work duties as models develop from advanced analytics. Customer experience management software reveals details that allow an organization to represent a customer (Soleimani, 2018). As per Bughin et al. (2017), businesses spent $25-$40 billion on AI in 2016, with big-tech firms spending ninety percent of their money on research and development (R&D) and implementation and ten percent on AI purchases.
AI is described as a machine’s capability to solve problems independently when it has not been specifically trained to do so. Current AI applications can gather information from around the world. This type of AI is designed to use logic and likelihood to select and function in the most effective way possible. AI acts intelligently and recognizes with exceptional precision using large data sets, images, and sounds (Pandian, 2019).
The Internet of Things (IoT), as per Ukil et al. (2014), uses network connections to build a large web of intelligent objects. While the conventional Internet connects individuals to facilitate contact, the Internet of Things connects machines and infrastructure with embedded sensors and allows them to communicate meaningfully through the Web. Machine to machine (M2M), sensors, intelligent world, pervasive and ubiquitous computing are some of the terms used to describe the Internet of Things. As a result, IoT is frequently described as a method of initiating intelligent facility identification, dynamic cognition, device capabilities, and customer interaction. Machine, interface, and application are typically the three parts of an IoT design.
2.3.1 DataData
In an Internet-enabled environment, consumer data capturing activities are collected by real-time con-function in a quantitative perspective, as per Ahlmeyer & Chircu (2016). There are numerous data types, including text, picture, video, sound, snapshots, and existing identity file-based information classes. A recipient is typically the one who establishes, produces, or registers the information. In an IoT-enabled environment, some gadgets keep track of and record all kinds of data related to user behavior in the current, non-numerical state.
Consumers in an Internet-enabled environment strongly influence appliances, causing them to react quickly to the device. An individual can, for example, use a device to request flight tickets electronically, monitor websites to find appropriate flights or carriers, and then purchase via the network using a credit card, all of which is achieved in real-time (Picone, 2020). Individuals can take advantage of IoT devices, but they usually do not enter DataData correctly. IoT devices separately control and collect relevant data from the Internet and recent human activities. Even though certain companies share data with others, user information related to Internet behavior is usually assigned inwardly within an organization or outwardly with affiliated individuals or representatives. On the other hand, DataData is shared with providers and other different devices when using IoT technologies.
2.3.2 Machine Learning
Deep learning is the foundation of machine learning, and the result is artificial intelligence. Among the most successful methods of AI is machine learning, in which a machine develops without being explicitly programmed (Vanchurin, 2021).
Figure 2: AI basic structure (Rouhiainen, 2019).
Machine learning is a term used to describe devices that can learn without being configured in a particular way. As a result, it investigates potential solutions and methods for the machine to solve the problem using existing evidence (Ni et al., 2019). Machine learning is divided into 3 parts: supervised, unsupervised, and reinforcement. In the supervised category, computer-implementable commands are used to obtain data that has already been arranged and labeled, and human intervention is needed to provide feedback to the algorithm. Unsupervised machine learning, in which DataData is not labeled or ordered, uses algorithms to find patterns in the data without the need for human interaction. Algorithms in reinforcement ML are versatile and can learn from their mistakes (Salminen, 2019). Machine learning methods attempt to replicate human actions based on previous experience and understanding. In terms of application, it can be a valuable method for determining the motivations of SC partners for collaboration and strengthening partnerships through organizational processes. As a result of a lack of cooperation, machine learning has recently been used to project unreliable demand data (Forrester effect) (Ni et al., 2019).
The machine learning process consists of 5 steps that must be completed for a particular function to be learned and evaluated effectively. The 1st step is data order, which involves transforming sorted data into random data. The 2nd step is to select a framework, after which an algorithm should be selected. In the 3rd step, the design must be conditioned, with the algorithm calculating each factor’s scales. The design is then evaluated in the 4th stage. Algorithms are monitored and assessed in this phase based on the findings and the actual purpose. The final step is to fine-tune the learning parameters to make the process as fluid as practicable. Predictive maintenance, hiring staff, improving customer satisfaction, finance, and customer support are a few examples of how machine learning is utilized (K et al., 2019).
2.3.3 Deep learning
Deep learning is a branch of ML that is one of AI’s fastest-growing technologies. It has the ability to learn from unmonitored, unmarked, and qualitative data. It is used to comprehend events and issues that are too complicated or difficult for humans to address. It usually involves a large amount of data. Neural networks are used in deep learning to identify complex connections and trends in data. It necessitates a large dataset and a lot of computing resources. Vehicle detection, image processing, speech and voice recognition are all areas where deep learning is commonly used. (Najafabadi and colleagues, 2015)
2.3.4 Neural Network
The structure of a neural network is similar to that of a biological organism’s brain cells. It can identify trends, analyze unstructured data, gain knowledge, discern attributes, and group objects based on conceptual information. A neural network is made up of nodes that are linked to each other and store long-term memory in the connections. Based on the weight of the links, data connections with the primary purpose may enhance or inhibit nodes. The process of learning entails the placement of links based on their weight. Using data structures, a neural network is expected to respond to a user’s preferences effectively. It’s also expected to discover intimate interplay between the data (Wang et al., 2021).
2.3.5 Expert System
Expert systems can mimic human cognitive abilities. It can perform problem-solving, language comprehension, and sensory perceptions, with a large amount of human intelligence. Expert systems are made up of 4 parts. Information base, search engine, justifier/scheduler, and user interface are the four components. The knowledge base is built on learned rules and human knowledge. The inference engine is referred to as the “brain” of this system. It’s a set of problem-solving software whose aim is to find and infer rules from the knowledge base. The justifier explains why and how the system arrived at that particular answer while the scheduler keeps track of and manages the sequential guidelines. The user interface aims to make communication between the user and the system as smooth as possible when dealing with user questions (Sari, 2018). Nevertheless, as the expert system grows in size, it becomes more challenging to handle the whole system’s data. When this system was checked, it proved to be a complicated process with more inconsistent data than positive outcomes. Furthermore, it was discovered that the machine did not learn over time, and by the late ’80s, the corporate community had lost interest in developing it (Taulli, 2019).
2.3.6 Genetic Algorithm
The genetic algorithm system emulates evolutionary biology convictions and collects natural selection laws. It generates species that are compatible with their surroundings. Besides, a genetic algorithm can be used to solve combinatorial optimization problems. It creates a structure that can be used to calculate unique representative values in a given context. There are 5 sections of a genetic algorithm. These are: a) A genetic illustration of a solution to the issue, b) A method for implementing a demographic, c) A evaluation function that assesses the tracking of solutions to observe if they survive, d) Genetic operators such as mutation, fusion, and replication that change the genetic makeup of descendants, and e) model parameters that describe total population, mutation rate, and biological makeup of descendants (Min, 2015).
1.2 Artificial Intelligence in Supply Chain Management
Increased competition, increased supply risk, and increased demand volatility put pressure on the ability to integrate and orchestrate the end-to-end process, to purchase parts and products, transforming them into finished goods, and then delivering them to customers. As a result, big companies exchange real-time data with SC partners, enriching their data sources (Min, 2015). In their study, Baryannis et al. (2018) divided AI’s assistance to companies into four categories. These 4 areas of value development are critical for gaining a competitive edge.
These areas include things like:
- a) Getting near-perfect predictions, like consumer demand and projections.
- b) Increase output while lowering costs and improving quality by maximizing R&D.
- c) Assisting with marketing, such as determining the price, demographics, identifying target buyers, and crafting the appropriate message, among other things.
- d) Give consumers a better experience
SCM is among the most dynamic fields of industry, focusing on cross-sector collaboration, marketing, manufacturing, and logistics. AI has been shown to be a critical feature of SCM in recent years. Modern machines with AI frameworks will collect data and utilize it to select the most likely and reasonable action with the best chance of success.
AI incorporation to SCM can be split into 3 parts, as per Min’s (2015) study. Inventory preparation, make-or-buy decisions, and supplier selection are all part of expert systems. Request scheduling, forecasting, consumer interaction management, negotiations, and order selection are all handled by a genetic algorithm that includes network architecture and agent-based models. AI is viewed as a decision-making tool that can assist businesses in connecting with clients, vendors, and network stakeholders to improve informational awareness. The application of AI, particularly in places where prediction is critical, such as replenishment, is technically and functionally advanced. Although rivals invest heavily in innovative ideas, AI leaders have incorporated a wide range of applications into their daily businesses. On the other hand, some businesses do not consciously use or make any attempt to implement such technology.
2.3.1 Demand Forecasting and Optimization
Keeping supply and demand in check has indeed been a challenge for businesses. With computer-based systems, projecting and predicting demand is nothing new, but improved output and supply chain prediction is needed. The use of an AI system to analyze data instantly results in more precise and accurate demand forecasts. Businesses may eliminate costs associated with supply chain activities by minimizing duplication and optimizing procurement with more reliable forecasting. AI assists in the creation of improved production and retailing approaches by recognizing industry trends and patterns. Weather-related applications (for example, Google’s DeepMind) determine the optimal supply and demand variance based on the nearby weather outlook on the shipment day. The AI solution considers prices, campaigns, regional climate predictions, past market records, and several other factors.
2.3.2 Production
AI has had a significant effect on manufacturing, allowing for better utilization of procedures and properties. AI will plan and model the best robot and human solutions for dependable and high-quality manufacturing. AI can even forecast service downtime and help avoid it. Robotics and automated technologies have resulted in sophisticated technology applications such as camera-equipped machines that can identify items and products and be trained to identify vacant shelf space. When contrasted with traditional approaches, this significantly increases the speed of selecting items.
2.3.3 Inventory
According to Malik & Sarkar (2020, pp. 515-516), logistics visionaries have predicted for many years that inventory’s position in new supply chains will be abolished or, at the very least, drastically altered. Since supply and demand will be flawlessly in harmony in the future, inventories will not require any cushion. This translates to a significant decrease in logistical costs. Most businesses have not perfected their technology and systems to the point that they might forego their core values, such as inventory. Inventory management can be the most noticeable aspect of SCM to end-users. Inventory control is the most critical task of operations planning since it ties up resources and influences the supply of products to consumers. Inventory management affects a variety of business functions.
1.3 AI’s Challenges and Opportunities in Supply Chain Management
1.3.1 Challenges
AI, such as robots, IoT, or smart devices enabling decision-making, will enhance the human experience. But it can malfunction, resulting in bodily harm, monetary loss, and more delicate harms such as the manifestation of human prejudice and a failure of personal dignity. These problems can lead to untrustworthiness because strange, unexpected, and new hazards can create overall discomfort and contribute to AI abandonment. It is a profoundly revolutionary technology that is rapidly developing and becoming omnipresent in everyone’s existence. The AI framework must be comprehensive, and it must account for the many ways in which AI can fail. Data accuracy is at the peak of the list of AI problems, according to Birkel & Hartmann (2019). Data and technologies are not yet mature enough for AI solutions to be implemented.
The following are the current difficulties of integrating AI tools in SCM:
- Because the user has no independent will, he or she is heavily influenced by computer programs, which can lead to incorrect judgments if coded incorrectly.
- Applications are challenging to develop since they are complex and difficult to adopt for regular decision-makers.
- SC decision contexts are cross-border and cross-functional, which AI can be unable to perform adequately due to information acquisition congestion (Min, 2015).
As per The World Economic Forum, using AI to optimize computers to fulfill people’s demands has drawn awareness to ethical concerns and risk assessments surrounding AI: Is AI causing a rise in unemployment? Is AI widening the divide between the rich and the poor? Can AI and robots affect people’s attitudes and interactions? How do we stop ourselves from making mistakes? Is it possible that computers can learn to be biased? (“Top ethical issues in artificial intelligence”, 2021).
1.3.2 Opportunities
According to recent research, well-organized AI tools in SCM are restricted to tactical and functional issues. In SCM, agent-based processes have the highest capacity to address strategic challenges such as customer experience management, contracting relationships, business-to-business agreements, strategic partnerships between SC partners, and joint demand scheduling to remove the bullwhip effect.
Manufacturing Firms may use AI to take over decision-making, regular planning, and tasks in SC to identify the drivers of new demand trends. When responding to volatile demand trends, demand planning usually is inefficient. Deep learning identifies trends in external signals instantly and can differentiate between irrelevant and essential signals. It can fine-tune demand estimates using signals. AI advancements include weather monitoring, spotting market capacity, identifying key demand factor variables, brand quality input, and gathering information from manufacturing machines to improve planning. Batches relevant to SC scheduling and decision-making phases can be identified using genetic algorithms. These requests are rerouted to avoid supply shortages in the immediate future. Using genetic algorithms to identify batches aims to recognize in-house expenditures and automate the acquisition of alternative supplies (Monahan, S. & Hu, M., 2018).
The alternative is not to avoid obtain the most recent planning software from an AI firm. A comprehensive AI approach includes the correct algorithms, a combination of data from various sources, and decision-making power. Sustained strategies facilitate end-to-end organizational change. Companies must find innovative technical technologies that assist them in a dynamic business environment to achieve effective SC planning.
- Summary and Conclusion of Literature Review
According to the literature review, AI focuses on prediction, demand planning, and efficiency at the planning phase in the SCM area. These aspects benefit the customers and allow for more accurate evaluations of processes and properties. Agent-based applications, which can work in various SCM applications, are the most promising spots of AI in SCM. AI can make decisions at the strategic, tactical, and operational stages, but it primarily operates at the operational stage, such as predicting output and warehouse activities.
Manufacturing companies spend vast sums of money in R&D as new AI technology is introduced. They seek to determine the best AI solutions for market behavior due to increasing competition, market volatility, and, more significantly, supply uncertainties. Nevertheless, firms should not invest in AI technologies blindly but rather evaluate what will provide them with the most long-term and extensive benefits. They must also consider the ethical and data maturity issues that AI could pose to the business.
AI-assisted SCM value development aims to achieve near-perfect forecast accuracy while lowering production costs. It also improves service by improving R&D, as well as assisting in the identification of target customers, and providing a better customer experience. Businesses will reduce waste and thereby be more efficient if they have reliable forecasts. They can also cut costs and improve procurement. AI in manufacturing can forecast service downtime and improve product reliability and quality. Robots with cameras can identify objects and materials, speeding up the picking process. When supply and demand are in complete balance, AI can have a significant effect on inventories. As a result, inventory capacity decreases, and consumer demands are met more quickly.
3.0 Research Methodology
3.1 Methodology
The thesis will be a qualitative exploratory study in which preliminary data will be gathered to aid in the definition of problems and the formulation of hypotheses. Exploratory experiments are designed to help people better understand how things work and how to handle them practically. It promotes public awareness of the phenomenon (Helo, Tuomi, Kantola & Sivula, 2019). If the research needs to explain an understanding of a topic, phenomenon, or problem, an exploratory study may help. A literature quest, expert interviews, in-depth person interviews, or focus group interviews are all part of this approach. Exploratory research is noted for its adaptability to change and versatility.
References
Ajournalarticle, R. H., Spud, P. T., & Psychologist, R. M. (2016). The title of the journal article goes here. Journal of Research in Personality, 22, 236-252. doi:10.1016/0032-026X.56.6.895*