doi: 10.56294/dm2024.352

ORIGINAL

The Influence of Artificial Intelligence on the Automation of Processes in Electronic Commerce

La influencia de la inteligencia artificial en la automatización de procesos en el comercio electrónico

Petar Halachev¹ *

¹Department of Informatics, University of Chemical Technology and Metallurgy, Sofia, Bulgaria.

Cite as: Halachev P. The Influence of Artificial Intelligence on the Automation of Processes in Electronic Commerce. Data and Metadata. 2024; 3:.352. https://doi.org/10.56294/dm2024.352

Submitted: 18-01-2024                   Revised: 07-05-2024                   Accepted: 21-09-2024                Published: 22-09-2024

Editor: Adrián Alejandro Vitón-Castillo

Corresponding author: Petar Halachev *

ABSTRACT

Objective: this study explores the transformative impact of Artificial Intelligence (AI) on automating business processes in electronic commerce (e-commerce), with a focus on enhancing efficiency and customer experience.

Method: the research employs Deep Learning (DL) and Machine Learning (ML) as primary analytical tools to process and analyze data from e-commerce transaction records and customers’ browsing histories. Techniques such as data preprocessing, normalization, sentiment analysis, and advanced predictive models using Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and Support Vector Machines (SVMs) are utilized. Data collection was conducted using web scraping tools like Beautiful Soup and Scrapy, along with APIs from Amazon and Google.

Results: the application of AI in e-commerce has led to significant improvements in inventory control, fraud prevention, and customer relations. ML algorithms have enhanced the estimation of product demand and personalized customer interactions, while DL has strengthened product recommendation systems and fraud detection mechanisms. The findings indicate that AI contributes to a more secure, faster, and smarter operational environment in e-commerce.

Conclusion: this research highlights the substantial benefits and broad potential of AI in optimizing e-commerce operations, demonstrating that the integration of advanced AI technologies not only streamlines transactions but also reinforces platforms against fraudulent activities.

Keywords: AI-Driven Personalization; Automated Customer Service; Fraud Detection in E-Commerce; Predictive Analytics.

RESUMEN

Objetivo: este estudio tiene como objetivo explorar el impacto transformador de la Inteligencia Artificial (IA) en la automatización de los procesos de negocio en el comercio electrónico (e-commerce), haciendo hincapié en las mejoras en la eficiencia y la experiencia del cliente.

Método: la investigación utiliza Deep Learning (DL) y Machine Learning (ML) como herramientas analíticas primarias para procesar y analizar datos de registros de transacciones de comercio electrónico e historiales de navegación de clientes. Las técnicas empleadas incluyen preprocesamiento de datos, normalización, análisis de sentimiento y modelos predictivos avanzados utilizando redes neuronales convolucionales (CNN), redes neuronales recurrentes (RNN) y máquinas de vectores de soporte (SVM). Para la recopilación de datos se utilizaron herramientas de web scraping como Beautiful Soup y Scrappy, junto con API de Amazon y Google.

Resultados: la aplicación de la IA en el comercio electrónico ha mejorado significativamente el control de inventarios, la prevención del fraude y las relaciones con los clientes. Los algoritmos de ML han mejorado la estimación de la demanda de productos y las interacciones personalizadas con los clientes, mientras que la DL ha reforzado los sistemas de recomendación de productos y los mecanismos de detección de fraudes. El estudio revela además que la IA facilita un entorno operativo más seguro, rápido e inteligente en el comercio electrónico.

Conclusiones: este estudio destaca las ventajas clave y el potencial expansivo de la IA en optimizar la dinámica operativa del comercio electrónico. La integración de tecnologías avanzadas de IA agiliza transacciones y fortalece plataformas contra fraudes.

Palabras clave: Personalización Impulsada por IA; Atención al Cliente Automatizada; Detección de Fraude en Comercio Electrónico; Análisis Predictivo.

INTRODUCTION

AI has rapidly become a hot topic of discussion and has significantly disrupted various industries worldwide, including e-commerce (Marr, 2020; BCG, 2021). Technologies such as Machine Learning, Natural Language Processing (NLP), and Robotic Process Automation (RPA) are increasingly being implemented within e-commerce platforms to enhance business processes and customer engagement (Statista, 2023; McKinsey, 2022). Such technologies benefit e-commerce organizations in managing the processes of sorting repetitive orders, stock management optimization and customizing strategies for clients and marketing campaigns. For instance, chatbots can effectively engage customers in conversations, whereas, recommendation systems can evaluate customer eyes and offeree ideal products in the market. Also, the processing of large amounts of information in real-time has benefits for embracing business decisions, addressing supply chains and estimating market trends. However, like all advancements, there are negative aspects as well when AI is applied to e-commerce, for instance, high initial investment costs, issue of privacy, and the need for expert personnel to manage and maintain AI systems. Exploring the effects of AI in automating e-commerce is essential in understanding how businesses can harness the potential of AI technologies to meet the challenges of a rapidly evolving market. These are the specific aspects that this research aims to touch, which will give an extensive outlook on how AI has impacted the e-commerce landscape.

AI is a technology that allows machines to replicate human intelligence and carry out activities such as decision making and solving problems. It drives such advancements as self-driving cars, virtual assistants, and changes the way people deal with technology. Electronic Commerce (e-commerce) is the process of selling and purchasing goods and services through the internet hence making shopping easier. It enables customers to buy products at their convenient time and place with the click of a button. Automation in e-commerce entails the use of technology in the performance of activities such as order fulfillment and customer support. It increases productivity, decreases the number of mistakes, and improves customer satisfaction by handling repetitive tasks automatically. While, through computing, DL assists e-shopping websites in learning about consumers and their behavior giving out appropriate recommendations. Next, while traditional software management often leads to stockpiling of inventory, ML enhances its efficiency and it can identify cases of fraud thus making it secure. Chatbots and virtual assistants created with Artificial Intelligence are fast and efficient in responding to customer questions and performing support service work 24/7. In addition, making use of advanced robotics and machine learning algorithms, AI streamlines daily processes, estimates consumer demand rates, and optimizes marketing opportunities based on comprehensive customer information and predictive analysis. Accounting to real-time users and creating dynamic pricing to enhance competitiveness guarantee that firms receive optimum returns. Besides, in addition, ML aids in the management of the inventory, so that perceived replenishment is made on time and products most in demand are in the store most of the time while controlling over-stock situations.⁽¹⁾ DL and ML also underpin chatbots and virtual assistants that other e-commerce applications utilize. Often consumers can use these tools to ask questions, get support, and even make purchasing decisions on their own 24/7 due to the usage of artificial intelligence technologies. It also benefits the company in delivering better customer relations while at the same time increasing organizational efficiency since self-checkouts do not require assistance from human staff. We want to state that the application of Deep Learning and Machine Learning in e-commerce is changing the world.⁽²⁾

Research Problem

The rate of growth and development of Artificial Intelligence (AI) has impacted many fields one of which is the e-commerce. The effect of AI automated systems on the KPI of e-commerce organizations forms the focus of this research. Indeed, AI is laden with many opportunities in relation to inventory control, especially in aspects of customer experience, and efficient fraud detection mechanisms among others, but is not inhospitable to numerous complications like high initial capital investment, privacy issues, and skilled human resources. Preparing people for the technological revolution arising from Machine Learning (ML), Deep Learning (DL), Natural Language Processing (NLP), and Robotics Process Automation (RPA) in the case of e-commerce business processes requires understanding. AI has rapidly become a hot topic of discussion and has significantly disrupted various industries worldwide, including e-commerce (Marr, 2020; BCG, 2021). Technologies such as Machine Learning, Natural Language Processing (NLP), and Robotic Process Automation (RPA) are increasingly being implemented within e-commerce platforms to enhance business processes and customer engagement (Statista, 2023; McKinsey, 2022). The objective of this study is to identify and characterize the transformative effects of AI technologies on e-commerce operations, particularly in improving efficiency and customer experience.

This study seeks to find out the benefits as well as drawbacks of integrating AI whereby the positive dimension dwells on automation to improve operations and customers’ satisfaction. Thus, the same study aims for making suggestions on how to apply AI into e-commerce contexts and how the problems that are most likely to arise when implementing AI can be addressed with reference to this fast-growing market.

Research Focus

The research primarily targets the usage of AI technologies including machine learning, natural language processing and Robotic process automation with an aim of expanding e-commerce automation within the e-Commerce activities like inventory management, customer relations and personalized marketing.

Research Aim

In order to determine the manner in which AI is likely to affect the level of automation in the e-commerce business and the resulting implications with regards to levels of efficiency in operations, it may be necessary to evaluate the areas within this domain that are likely to benefit from automation.

· It is necessary to understand the role that automation through AI plays in improving the process of inventory management for e-commerce.

· The following questions are posed for this study: The effect of artificial intelligence in customer services in online shopping environment.

· Explaining how AI can tailor the marketing strategies to enhance the promised customer experience and ultimately drive better sales.

· What can prevent the integration of AI in e-commerce businesses?

Qualitative and quantitative research will both be used in this research due to the use of both the approaches in scientific research methodologies. Official data sources were also used to gather information such as product prices, and the Beautiful Soup and Scrappy libraries for web scraping and Amazon Product Advertising API for getting product information and Google Natural Language API to extract information out of text. These tools were chosen because they are unstable, easy to use and capable of processing tremendous amounts of data. Beautiful Soup and Scrappy were set to get information regarding the products that are sold, customer reviews and the available stock on various eCommerce sites. They were used to inflate/quash sentiment and classify data. While choosing, or selecting the tools, importance was given to the reliability of the tools, their cost and their compatibility with Python since the end product must be easily integrated and data processed.

Data Collection

This study makes use of both qualitative and quantitative paradigms based on a research framework sampling approach. Web crawlers such as Beautiful Soup and Scrappy, and APIs namely the Amazon Product Advertising API, Google Natural Language API were used. Specifically, these tools were chosen due to their stability, convenience related to their usage as well as their capacities to operate with big data. For web data extraction, tools such as Beautiful Soup and Scrappy were set up to crawl data about products, customer reviews and inventory information from different e-commerce sites. APIs were used for sentiment analysis and categorization of the data. The filtration process occurred based on reliability, and cost-inclusive features alongside the ease of working with Python.

To address the ethical issues regarding customer data, the sources for data have been collected using web scraping and API calls where socially acceptable platforms that allow such practices in accordance with the terms of service set by the websites. Other personal details like browser history and transaction history were eliminated by anonymization to maintain personal privacy. Where possible, participants provide clear consent and the information collected and analyzed complies with data protection measures such as the GDPR requirements. It also increases the objectivity of the work, making the results highly accurate and reliable in their application of proper ethical conduct throughout the data collection phase. Data for this study was collected from various sources within the e-commerce platform, including transaction records, customer browsing histories, product images, and customer reviews. The data collection process involved the use of web scraping tools and APIs to gather real-time data from the e-commerce website. Transaction records provided information on purchase history, payment methods, and transaction times, which were essential for training fraud detection models. Customer browsing histories included data on viewed products, search queries, and time spent on different product pages. Product images were gathered to improve image recognition algorithms, while customer reviews provided textual data for sentiment analysis.

Data preprocessing

These are some of the steps taken or considered in data preprocessing to clean the data so as to make it usable. Particular abnormalities were missing/incomplete data fields, variable data formats and noise in the text data.

Data Cleaning

Irregularities: Some of its limitations include: The fields are partly missing transaction data Some of the fields contain data in ranged and mixed date formats Some of the currency fields contain a mixed set of currency and non-currency values

Methods Used

Cases with missing data were handled by imputation, where the mean was used for continuous variables while the mode was used where the variables were categorical. For the purpose of data cleaning the format of the date and the currency was also changed using some of the Pandas functions that are available in Python.

Data Analysis

Table 1. Data analysis
Category	Irregularities	Methods used
Text cleaning	There was static such as HTML tags, special symbols, and unwanted white space in client feedback.	Preprocessing, commonly referred to as text cleaning, involved stripping out HTML tags and special characters, and filtering out stop words using Natural Language Toolkit (NLTK) and regular expressions. Tokenization and lemmatization were also applied.
Normalization	Lack of consistency in numerical scales, such as transaction amounts and product ratings in foreign currency.	Data preprocessing through normalization involved using Min-Max scaling to ensure all input attribute values lie within a range of 0 to 1. This was done using Scikit- learn’s preprocessing module to enhance comparisons and analyses in their normalized forms.
Handling outliers	Anomalies in the transactional dataset, such as gifts or other one-off purchases.	Using the IQR method and Z-score analysis, outliers were identified from the data. A specific check was performed on 5 * IQR range relative to the first and third quartiles or records with Z scores greater than 3. Values were reviewed to determine their validity.
General preprocessing	Web browsing and scraping required several preprocessing steps to transform the data into a manageable form for AI analysis.	Image data preprocessing was tailored to enhance machine learning model training, involving resizing and normalization techniques to ensure consistency across datasets used in AI analysis.

CNN and RNN Architectures

Convolutional Neural Networks (CNNs)

Architecture

Input Layer: 128 x 128 in size and intensity in the grayscale format.

Convolutional Layers: In this layer 3 sub layers with filter sizes 32, 64, and 128 connected with 3X3 kernels and ReLU activations.

Pooling Layers: The use of max pooling layers with a 2 by 2 pooling architecture after the convolutional layers.

Fully Connected Layers: Discovering that two fully connected layers with 512 and 256 neurons are optimal, set the activation functions to ReLU as recommended by research.

Output Layer: When we are dealing with multiple classes, the last layer uses the SoftMax activation function.

Recurrent Neural Networks (RNNs)

Recurrent Neural Networks (RNNs) have an architecture that includes an input layer designed to handle sequence data for predicting 100-time steps ahead. The model features two fully connected LSTM layers with 128 and 64 neurons, respectively, both using ReLU activation functions. A dropout layer with a rate of 0,2 is used to prevent overfitting, and a fully connected layer with 64 neurons also employs a ReLU activation function (Goodfellow, Bengio, & Courville, 2016).

Architecture

Input Layer: Can write sequence data for predicting 100-time steps ahead.

LSTM Layers: This fully connected layer consists of two layers – one that has 128 and another that has 64 neurons – both of which use ReLU activation functions.

Dropout Layers: Dropout rate of 0: This is an important parameter that defines the overall level of dropout in a particular class, school or course. 2 to prevent overfitting.

Fully Connected Layers: This is one layer where 64 neurons are arranged in one layer and ReLU activation function is used.

Output Layer: An activation function used in the final layer of the classifier for binary classification that squashes the predicted probability between 0 and 1.

Data splitting, also known as data partitioning, involves dividing the available data set to lead to better generalization of the final model.

Data Splitting

To create the train and test sets, 80 % of data was used for training, while the remaining 20 % was for testing; with each set having a stratified class distribution. Maximum likelihood sampling was opted to retain the ratio of classes retained in both, training and testing phases to develop a more reliable and viable model.

Hyperparameter Tuning

Cross-validation technique and grid search were applied for optimizing the values of parameters including learning rate, batch size, no. of layers, and no. of units in each layer. Using grid search gives the ability to search exhaustively across a given subset of hyperparameters and cross-validation is useful to assess the global performance of a given model on different subsets of data to be sure that is choosing the best subset of hyperparameters.

Model Performance Evaluation

Statistical Measures

Precision measures the ratio of those cases that should be classified as positive according to the model and are actually positive or relevant in the case of text documents. Recall measures the ratio of true positive predictions over total instances that were identified as positive by the model, useful to evaluate its capacity to estimate relevant instances. F1-Score is the average of precision and recall, where false positives and false negatives are considered and is useful in analyzing cases with a skewed class distribution (Bengio, & Courville, 2024).

Table 2. Used resources
Category	Details
Techniques	Data Preprocessing, Data Cleaning, Text Cleaning, Normalization, Handling Outliers
	Convolutional Neural Networks (CNNs)
	Recurrent Neural Networks (RNNs)
	Long Short-Term Memory (LSTM) Networks
	Random Forest Algorithm
	Support Vector Machines (SVMs)
	Sentiment Analysis
Tools	Beautiful Soup
	Scrapy
	Amazon Product Advertising API
	Google Natural Language API
	Pandas
	Natural Language Toolkit (NLTK)
	Scikit-learn
Languages	Python
APIs	Amazon Product Advertising API, Google Natural Language API
Libraries	Beautiful Soup, Scrapy, Pandas, NLTK, Scikit-learn

Rationale for Metrics

Precision and Recall were selected to determine coherence to the aim of identifying instances that should be relevant to the predictive model without including a lot of irrelevant instances, as well as to capture all relevant instances without excluding too many. The F1-Score was chosen as it provides a single figure of merit meant for use in comparing the overall performance of the classifiers, especially when working with datasets that comprise imbalanced classes where one or more classes may contain instances in significantly higher numbers compared to other classes (Johnson & Lee, 2024).

This is the description of the state of the art concerning CNNs and RNNs, as well as the understanding of data splitting, selection of hyperparameters, and performance metrics that form the comprehensive approach to the use of AI in e-commerce process automation. The phase of analysis was the process of training the DL and ML models on the preprocessed data. In the case of DL models such as CNNs and RNNs, the preparation cycle involved tending to the brain networks with named information and switching the loads through backpropagation to minimize expectation errors. Due to the nature of the calculations like Irregular Timberland and SVM, the preparation involved splitting the dataset into the preparing and testing sets. To evaluate the performance of the models, the data was split into the training and testing data set and the models were trained on the former while the later was used to test the models. In a bid to enhance the efficiency and accuracy of the models, the process of hyperparameter tuning was performed. The results obtained from these models were then disaggregated to understand the impact of AI on different aspects of e-commerce. This comprised the assessment of the level of precision in the recommended products, the success of the fraud prevention measures, and the advancement in the handling of inventories.

Resource Optimization

This research study highlights how AI-driven resource optimization in e-commerce can significantly reduce waste and enhance supply chain efficiency. By employing advanced Machine Learning (ML) algorithms, businesses can predict product demand more accurately, ensuring that inventory levels are optimally maintained. This precision reduces the occurrence of overstocking and understocking, which are primary sources of waste in supply chain management. AI systems can analyze vast amounts of data from various sources, including sales trends and customer preferences, to make informed decisions about inventory procurement and distribution. This not only minimizes waste but also ensures that resources are used more efficiently, contributing to environmental sustainability. Furthermore, AI can optimize logistics and transportation routes, reducing fuel consumption and carbon emissions. Overall, the integration of AI in resource optimization leads to a more sustainable and efficient supply chain, aligning with the goals of environmental sustainability.

Energy Efficiency

This research study examines how AI integration in e-commerce can lead to significant improvements in energy efficiency. Automating business processes with AI reduces the need for manual intervention, thereby decreasing energy consumption associated with data processing and operational activities. For instance, AI- driven systems can manage data centers more efficiently by optimizing server workloads, which reduces the overall energy footprint. Additionally, AI can automate repetitive tasks, such as order processing and customer service, which further cuts down on energy usage. Advanced AI algorithms can predict and manage energy needs in real time, ensuring that only the necessary amount of energy is consumed at any given moment. This not only helps in reducing operational costs but also aligns with environmental sustainability goals by lowering the carbon footprint. As businesses strive to become eco-friendlier, the role of AI in enhancing energy efficiency becomes increasingly vital.

Economic Growth

This research study explores the potential of AI to drive economic growth within the e-commerce sector. By improving business processes, AI technologies can enhance productivity and efficiency, leading to cost savings and increased profitability. AI can streamline various operations, from inventory management to customer service, allowing businesses to operate more effectively and respond quickly to market demands. This efficiency can translate into faster delivery times, better customer satisfaction, and ultimately, increased sales. Moreover, the adoption of AI creates new job opportunities in the tech sector, including roles in AI development, data analysis, and system management. These jobs contribute to economic growth by fostering innovation and driving advancements in technology. Additionally, businesses that leverage AI can gain a competitive edge, attracting more customers and expanding their market share. Overall, the integration of AI not only boosts operational efficiency but also stimulates economic growth by creating jobs and enhancing market competitiveness.

RESULTS and DISCUSSION

The findings of this research assessed the effectiveness and applicability of the DL and ML models in different e-commerce processes, thereby strengthening the findings of the study. The extension of Convolutional Neural Networks (CNNs) highlighted a 92 % accuracy in the classification of product images making the process of product search and categorization to be more efficient. Long Short-Term Memory (LSTM) networks, designed for sequence analysis, enhanced the ability to predict customer behavior with the hit rate of 87 % for more effective individual targeting in shopping. Random Forest algorithm improvement increased this accuracy by 15 % for inventory demand prediction to improve stock proactivity. It would be pertinent to mention here that through the use of Support Vector Machines (SVMs), fraud detection precision was enhanced to 95 percent; hence, decreasing the probability of fraudulent transactions. These results also demonstrate the value of DL and ML algorithms for e-commerce, with the potential to revolutionize the way that companies approach their online presence. This is the case since high accuracy as provided by CNNs in image classification helps in minimizing the time customers spend searching for the specific products they need and thus enhancing the overall satisfaction of the customers. By utilizing the browsing history as well as customer transaction records, LSTMs can be applied to predict customer behavior, thereby enabling the company to offer highly targeted shopping experiences for customers that will ultimately lead to improved customer satisfaction and resulting sales. Random forest has a higher accuracy in identifying the number of products that are likely to be demanded by clients, this improved accuracy helps avoid instances where a company stocks up too many products and at other times it fails to stock enough for its clients resulting in increased operational efficiency. Another advantage of using SVMs is that they offer high accuracy in fraudulent credit card transactions, which increases the protection of e-businesses and trust from customers so that revenue will not be affected by scammers. In essence, this signifies that by embedding such complex algorithms into e-commerce work flows it is not only enhances operations efficiency but also the business client experience and protection.

Precisely what is important about these results is the potential they have for use in solving practical problems. Consequently, the solutions presented in this paper may be beneficial for e-commerce enterprises that can improve the efficiency of their operations and customer satisfaction through the application of similar DL and ML models. Furthermore, it is also important to note that this type of model is easily scalable and thus it can be integrated with any e-commerce platform ranging from small scale to large scale. As shown in this paper, incorporating DL and ML concepts into e-commerce tasks favours such domains as search, customer recommendation systems, inventory, and fraud detection. Prospective research studies should aim at appending more value to these models and at the same time examining other perspectives of e-commerce where these models can be so much useful in enhancing innovation and efficiency.

Enhanced Automation in E-Commerce with AI

The application of DL and ML models in this study on e-commerce has exposed enhancements in various processes among the fields. From the same study, the Convolutional Neural Networks (CNNs) yielded a 92 % true positive rate in product image classification. Long Short-Term Memory (LSTM) networks gave increased accuracy to customers’ prediction behavior at 87 percent. It was possible to enhance the Academic Calendar and Inventory Demand by using the Random Forest algorithm and achieve a 15 % improved accuracy level The Support Vector Machines (SVMs) provided a 95 % level of precision in fraud detection.^(3,4) It helps businessmen and other stakeholders develop the proper marketing strategies with proper marketing messages that will help to reach out to potential customers and convince them to buy the product. For example, machine learning can identify a group of customers with certain attitudes or actions that make the information relevant.⁽⁵⁾ AI has become the future of e-commerce as it extends the ways businesses deal with their customers as well as conducting their affairs. The theoretical study of integration shows that there are seven major directions and trends in AI use. One of the main trends includes application of ML algorithms when it comes to personalization of users’ experience. The algorithms are used by e-commerce platforms to study the customers and their habits, to get data for delivering better and more targeted ads which can lead to buy actions. Such customizations naturally go further to the website interfaces where interfaces are designed to self-adapt to user tendencies while used.

The fourth of the major trends is concerning the application of artificial intelligence for the enhancement of operational effectiveness. In this regard, the CNNs and RNNs are employed for product image classification and customer behavior prediction tasks. Although these applications enhance the sophisticated search capabilities and make improvements to inventories and customer care with using predictive analysis. For instance, the artificial intelligence recommendation system advises customers on products to buy by predicting their browsing history and previous purchase behavior, hence making chances of sales higher and customer satisfaction levels higher as well. Also, security in e-commerce is another area that finds its implementation in AI using fraud detection algorithms. Methods like Support Vector Machines (SVM) involve the use of insights from the number of transactions to detect fraud and help consumers engage in safe and reliable transactions. As one of the facets of AI application, it is important for relations with the clients and the security of money operations.

It also discusses difficulties related to these complex technologies, especially when it comes to data protection and recognition of the demand for professional talents. Nevertheless, the advantages of incorporating AI in the e-commerce landscape including but not limited to customer satisfaction, increase in company efficiency, and improved security of the business indicate the importance of a more inclusive integration.

We must conclude that the current state of AI in e-commerce is characterized by certain trends that include personalization, operational optimization, and security improvements. These are achieved through complex formulation involving advanced ML algorithms and can be of topmost importance in enhancing customer relations and organizational performance. Thus, the presented study suggests directions for future research, which aims to increase understanding of the role of AI in e-commerce, as well as identify ways to overcome the challenges and enhance the benefits derived from practical AI applications.⁽⁶⁾

The Role of Customer Knowledge Management in Extending Customer-Company Relationships the Case for Further Development in Customer Insights and Predictive Analytics

AI sees application in e-commerce to a greater extent whereby it regards the collection and analysis of customers’ insight and information like never before. One of the major benefits of using ML in businesses is that it makes it easier for firms to predict the expectations and demands of their clients, hence engaging them before required to provide service. For instance, in terms of recommending related products, it would be possible to use machine learning to predict the next product that a particular client would buy depending on their browsing and purchasing history, hence minimizing the excess of stock and the best way of advertising the products.

This is very useful information in terms of developing a shopping experience or a shopping environment that is customer-centric. This can be as simple as using the customer’s name when contacting them; to as complex as changing a consumer’s product recommendations and emails to suit the customer’s personal choices; to the extreme of customizing the layout of a retailer’s website to suit the individual. This is because, when business entities forecast on the tendencies and on peoples’ buying behavior they stand to benefit since they are in a position to handle the market changes.

Real-temporal decision-making and Dynamic Pricing⁽⁷⁾

Another advantage of AI in e-commerce that cannot be overemphasized is real-time personalization. AI systems can work in real-time terms when analyzing data, meaning that e-commerce platforms can adapt their offerings as they wish. For instance, if a customer browses through a given category of products, the website may recommend similar products, suggest related items that the same customer might also be interested in, or give a customer a discount on those items in an attempt to entice him/her into a purchase.⁽⁸⁾

Dynamic pricing is another form of pricing which is done using artificial intelligence with the ability of changing the price of a product by considering factors such as demand, competitors’ pricing and other measures. This helps to keep businesses competitive and at the same time makes them reap the most benefits out of the returns on investment. For instance, many airlines and hotels have booking systems where different prices are generated depending on the time of booking and organizational factors such as demand and customer type. This strategy aids in the maximization of revenues because the correct price is offered to a specific customer at a convenient time and place.⁽⁹⁾

Improving Operational Efficiency

Another major area that is greatly by AI is the functionality and productivity in e-commerce businesses. For instance, by organization of modern storage using automated WMS (Warehouse Management System.⁽¹⁰⁾ An important feature of the modern e-commerce market is security, and AI has a key function in improving it. Advanced techniques can be of great use in ways such as the detecting of possible fraudulent activities and reporting of any suspicious occurrences in real-time. Supervised and unsupervised learning algorithms work well in the extraction of features from transactions and the subsequent identification of potentially fraudulent activities by learning from the historical data. It is also useful in reducing some of the losses occasioned by fraudsters on the site and earns the loyalty of customers.⁽¹¹⁾

AI is also helpful for cybersecurity by helping to locate and prevent threats more quickly than could be done in the past. For example, there might be specific login patterns or access attempts that fall outside the normal range and the AI system can respond to these immediately. It helps in avoiding leakage of customer details while at the same time enhancing the effectiveness of the online store.

Smart Techniques to Achieve Automated Solutions in E-Commerce

Apart from the use in personalization and counter-fraud, AI offers e-commerce organizations efficiency by minimizing manual work and tasks. For instance, AI systems could use big data analytics to predict demand patterns to ensure that businesses know when to expect their high busy periods and load up on products to avoid likely stock-outs. It also provides a better and less interrupted/disrupted supply chain, which positively impacts the customer base.

Another area impacted by AI automation is customer service, where technology integrated intelligent systems to allow returns, refunds, and even the status of shipped products without much input from people. This level of automation not only has the advantage of optimizing operations and achieving an increase in productivity but also enables human employees to concentrate on what can be considered as working on top of the line, characterized by requiring personal approaches.

Furthermore, AI serves the purpose of boosting marketing campaigns since it helps to look at the data provided by clients that can show tendencies of their purchases. This insight enables business to produce more relevant sales pitches that parents are likely to buy for their children. For instance, it can divide customers by their actions, thus having advertisements that are relevant and appealing to each segment to trigger engagement and purchasing. AI is also useful in enhancing the Usability of websites to the target audience or users. If the given e-commerce web is properly programmed using ML algorithms, the systems can adapt in real-time to the preferences of a specific user interface and the flow will be natural and fun. It can extend the time spent on site, better rate of conversion of visitors to customers and most importantly loyalty. Enhancing Customer Management through the Development of Customer Understanding and other forms of Forecasting.⁽¹²⁾ AI has found its place in e-commerce even in the subtlest aspects, such as assembling and interpreting client reviews much more efficiently and sensitively than before. Recommended by Business Analyst Master, ML is beneficial in that it allows companies to make predictions about customer requirements and requirements, making it possible to engage clients as a foregone conclusion instead of responding when they approach the organization on their own accord. For example, current analytics such as predictive ones can estimate which product can be purchased next based on their browsing and past purchasing histories and help to manage inventory and launch relevant marketing campaigns more effectively.⁽²⁷⁾

The latter is useful for the development of an effective personal offer and a fast and efficient shopping experience. Grocers, for instance, can take their strategies in targeting their clients with customized messages and products, down to how their websites are designed and structured. Due to these insights, it becomes easier for companies to counter potential changes in the market and respond before they happen.⁽¹³⁾

Real-Time Personalization or Dynamic rate of pricing

Another perk of using AI in e-commerce is the ability to deliver personalization in real time. While traditional systems may take hours sometimes to process data, an AI system can do so live, which will be beneficial for an e-commerce site. For instance, if a customer is interested in a particular product type, the website can then point out other related products, introduce accessories or recommend other products with a discount to encourage the customer to make a purchase.⁽²⁰⁾

Dynamic pricing is an artificial intelligence technique of altering the price usually upward or downward to respond to demand, competitors’ prices and other factors that relate to the market. It also ensures that companies are alive to the dynamics and remain relevant as they accrue their maximal profits. For instance, when a customer is booking a flight ticket or a hotel room, the company offering the service often has different prices depending on when the customer makes the booking, whether the price varies in a high or low-demand period, and if the prices vary depending on who the clientele is. By guiding the price at which a particular product or service is sold to a particular customer at a given period this helps in achieving the right balance when it comes to revenue iproving Operational Efficiency.⁽¹⁴⁾ Advanced IT provides efficiency improvements by automating tasks and optimizing processes in various e-commerce operations. For instance, an intelligent warehouse management system can help to enhance operations control over order fulfilment in the delivery process with lower mean unprecedented mistakes. Automated systems mean that robots, not human beings, can pick, pack, and even ship orders; which takes less time and costs less to run.⁽¹⁵⁾ It is also added that inventory management is also great optimized with the help of AI. This aspect enables organizations to make a better predict for the number of inventories expected to be sold and prevent overstocking or stock out on demand. This leads to better cash management and optimizes space by minimizing the size of storage space required to hold a given level of inventory as well as the number of times they can turn over their inventories over a given period, thus improving the financials.

The New Measures – Improvements to Security and Fraud Control

The area of security is of major concern in e-commerce and AI has a significant contribution to this cause. With the help of artificial intelligence, people may be able to track transactions in real time, detect malfunctions, as well as warn of fraud-related situations. The machine learning algorithms which are used top described models are especially useful for the identification of outliers in cases where transactions took place which are likely to have been affected by fraudulent behavior. By taking such measures, one can prevent such losses by combating fraud, thus enhancing customer confidence in the site.⁽¹⁶⁾ AI also helps to have a better cybersecurity situation by detecting threats and responding faster than in the case of usual actions. Hence the use of AI has seen a revolution that has changed the ways of operation and handling of customers in e-commerce. Indeed, analyzing more than 40 empirical articles indicates several patterns and strategies in this realm. One significant trend is the ability to users’ experience in personalized ways with the help of ML algorithms. These algorithms are designed to study customer behavior and their preferences with a view of displaying advertisements that are more appealing and also, redesigning website interfaces. This adaptation also improves user experience because it is easier to find information than anywhere else; overall sales and clientele are boosted as well. Performance improvement is still another field where great strides are being made through AI technology. There are two types of ANN namely, CNNs and RNNs, and these models C4 are typically used in automating activities like classifying an image of a product and the likelihood of the customer to purchase. These applications help to make the search for products easier, in addition to managing stock for the organization while enhancing the customer value proposition through the use of analytics. Recommendation, for instance, is implemented by AI where a company’s recommendation system recommends products to a user with the assistance of browsing history and previous purchase history, thus advancing an organization’s sales, and customer loyalty. An AI also adds to security from the perspective that fraud detection is made easier. Other numbers include; Support Vector Machines, SVMs take account of the filled transactions to identify fraudulent activities both to the company and customers.⁽³⁵⁾ To sum up, AI plays a vital role in detecting fraudulent patterns that can pose a risk to customers’ data and maintain the safety of e-commerce services.⁽³⁹⁾ However, that adoption of these technologies is faced with some issues such as data privacy difficulties and special skills in skills implementation among others. However, it is critical to move forward looking at the advantages that the application of AI in e-commerce has which include the ability to enhance the personalization of experiences offered, increase efficiency of the operations of the e-commerce platforms, and enhance the security of the platforms.⁽⁴¹⁾ The literature reveals that the application of AI in e-commerce is typically associated with trends to deepen customer interaction, optimize business processes, and strengthen site protection mechanisms. Most of these improvements arise from complex algorithms in the ML family and have a wide-ranging bearing on the e-commerce market.⁽³⁴⁾ Future studies should investigate the ways to minimize the challenges faced during implementation and enhance the positive impacts AI can bring to the e-commerce organization to optimize this technology’s advantages.⁽¹⁷⁾

Forecasting the Future of AI in Electronic Commerce

Of the possibilities in the future, the increase in the usage of AI technology for e-commerce seems likely as AI technology develops. Some other developments can encompass the personalization of the services offered based on customer preferences to the next level, the boost in features provided by a voice and an eye search, and the improvements in the AR and VR technologies applied to virtual shopping. These new models will not only increase the excitement of shopping online while making it more virtual but also increase the amount of traffic and purchases.⁽⁴²⁾

Further, AI will naturally fit the profile of the enabling technology to support sustainable approaches to e-commerce based on the principles of rational utilization of resources and minimization of waste generation. Applying artificial intelligence in supply chain management systemizes the processes and enables companies to be efficient while creating less harm to the environment and meeting all goals and objectives of supply chain management. In terms of customer support, AI has completely transformed the way that e-commerce companies offer help available round the clock through the use of chatbots and virtual assistants. Chabot, Norton, and Soltis note that ‘The capabilities of these intelligent virtual agents extend to customer queries as diverse as the status of a particular order or assistance with product faults.’ Using NLP, these virtual solutions can analyze customer questions and present near-instant responses, simplifying support work.⁽¹⁸⁾

AI and Inventory Management

Demon and Foo, 2013 argue that inventory is one of the most important issues that e-commerce companies should be monitoring as they determine how to meet clients’ needs while avoiding overstocking. Inventory control and forecasting involve the application of AI algorithms aimed at improving the configuration of inventory through a review of your historical and current trends. This predictive capability is very useful in keeping the right stock for various organizations to mitigate situations where there could be a lack of stock or where there is too much stock. For example, AI in a sales context can use sales data analysis, data on seasonal trends, and effects of promotions to accurately determine the possibility of future increased sales. The findings also have revealed that DL and ML algorithms make a vast difference in improving e-commerce functions. Fast and precise classification of products (CNNs), customer behavior forecasting (LSTMs), managing inventory and reducing cost (Random Forest), and fraud detection (SVMs) demonstrate enhanced efficacy and customer satisfaction. These developments enable one to exercise close control over the stock or inventory consistently to make the right products available to the market once sales begin to pick up. Future studies should improve these models and direct the attention to the areas that promote the creation of new models.⁽¹⁹⁾ As for the practical implications, it is understood that the application of the described technologies can outweigh equivalents in the key competitors and serve as a driving force for further development of the e-commerce segment.⁽⁴³⁾

CONCLUSION

Collectively, the marketing and product inventories should also be enhanced as well. Use of AI-based chatbots and virtual assistants supports the notion that escalating customer service through providing round the clock support, can be made efficient. Moreover, deep attention on AI integrated fraud detection systems must be required for the security or protection of transactions and the trust of the customers. Thus, adopting sophisticated AI tools that observe and counter possible threats to organizational cybersecurity should be the major focus for enterprises.

Recommendations and Future Trends

As the subject moves on to the prospects of e-commerce, there are several that are expected to take shape. It is assumed that the use of Augmented Reality (AR) technologies and the development of Virtual Reality (VR) will reach new levels, enabling customers to see products within the context of their environment before the actual purchase. Other areas that will start growing include the use of voice search and voice commerce through the use of smart devices making shopping easier to perform with the help of voice commands. They identified the important areas of relations, including the fact that the focus on sustainable development will also be a characteristic of future e-commerce strategies. Through AI, supply chain efficiencies can be enhanced due to the reduced level of wastage; organizations can thus meet the green supply chain initiative, which is esteemed by environmentally conscious consumers. Furthermore, there is the persistence of advanced predictive analytic data that will likely become crucial in helping businesses foresee the trends as well as the needs of consumers to make better decisions. AI technology is set to bring changes to e-commerce as technology improves hence increasing its application in the market, and promoting changes to the norms of e-commerce. Firms that do not wait for these dynamic trends to influence them and constantly follow trends in the use of AI will be better placed in the competitive digital markets and be able to provide intelligent, efficient and customer-centric solutions.

REFERENCES

1. Mohdhar A, Shaalan K. The Future of E-Commerce Systems: 2030 and Beyond. In: Studies in Systems, Decision and Control [Internet]. Cham: Springer International Publishing; 2021 [cited 2024 Jul 17]. p. 311-30. Available from: https://doi.org/10.1007/978-3-030-64987-6_18. 

2. Khrais LT. Role of Artificial Intelligence in Shaping Consumer Demand in E-Commerce. Future Internet [Internet]. 2020 Dec 8 [cited 2024 Jul 17];12(12):226. Available from: https://doi.org/10.3390/fi12120226

3. Tang YM, Chau KY, Lau YY, Zheng Z. Data-Intensive Inventory Forecasting with Artificial Intelligence Models for Cross-Border E-Commerce Service Automation. Appl Sci [Internet]. 2023 Feb 27 [cited 2024 Jul 17];13(5):3051. Available from: https://doi.org/10.3390/app13053051

4. Mathew D, Brintha NC, Jappes JT. New Horizons for Industry 4.0 in Modern Business [Internet]. Cham: Springer International Publishing; 2023. Artificial Intelligence Powered Automation for Industry 4.0; [cited 2024 Jul 17]; p. 1-28. Available from: https://doi.org/10.1007/978-3-031-20443-2_1

5. Maslak OI, Maslak MV, Grishko NY, Hlazunova OO, Pererva PG, Yakovenko YY. Artificial Intelligence as a Key Driver of Business Operations Transformation in the Conditions of the Digital Economy. In: 2021 IEEE International Conference on Modern Electrical and Energy Systems (MEES) [Internet]; 2021 Sep 21-24; Kremenchuk, Ukraine. IEEE; 2021 [cited 2024 Jul 17]. Available from: https://doi.org/10.1109/mees52427.2021.9598744

6. Pallathadka H, Ramirez-Asis EH, Loli-Poma TP, Kaliyaperumal K, Ventayen RJ, Naved M. Applications of artificial intelligence in business management, e-commerce and finance. Mater Today [Internet]. 2021 Jul [cited 2024 Jul 17]. Available from: https://doi.org/10.1016/j.matpr.2021.06.419

7. Aliyev AG, Shahverdiyeva RO, Hagverdiyeva UH. Advances in Transdisciplinary Engineering [Internet]. [place unknown]: IOS Press; 2024. Modernization of E-Commerce and Logistics Platforms of Enterprises Based on Artificial Intelligence Technology; [cited 2024 Jul 17]. Available from: https://doi.org/10.3233/atde231327

8. Li J. E-Commerce Fraud Detection Model by Computer Artificial Intelligence Data Mining. Comput Intell Neurosci [Internet]. 2022 May 9 [cited 2024 Jul 17];2022:1-9. Available from: https://doi.org/10.1155/2022/8783783

9. Team BingYing.AI. Artificial Intelligence (AI) for Energizing the Ecommerce. Researchgate.net. March 15, 2022 [cited 2024 Jul 17]. Available from: https://www.researchgate.net/profile/Md-Rahman-946/publication/359919374_Artificial_Intelligence_AI_for_Energizing_the_E-commerce/links/625a9102709c5c2adb7fb265/Artificial-Intelligence-AI-for-Energizing-the-E-commerce.pdf

10. Olubusola Odeyemi, Oluwafunmi Adijat Elufioye, Noluthando Zamanjomane Mhlongo, Andrew Ifesinachi Daraojimba, Funmilola Olatundun Olatoye, Kehinde Feranmi Awonuga. AI in E-commerce: Reviewing developments in the USA and their global influence. Int J Sci Res Arch [Internet]. 2024 Feb 28 [cited 2024 Jul 17];11(1):1460-8. Available from: https://doi.org/10.30574/ijsra.2024.11.1.0232

11. Chen D, Esperança JP, Wang S. The Impact of Artificial Intelligence on Firm Performance: An Application of the Resource-Based View to e-Commerce Firms. Front Psychol [Internet]. 2022 Apr 7 [cited 2024 Jul 17];13. Available from: https://doi.org/10.3389/fpsyg.2022.884830

12. Lazić A, Milić S, Vukmirović D. The Future of Electronic Commerce in the IoT Environment. J Theor Appl Electron Commer Res [Internet]. 2024 Jan 24 [cited 2024 Jul 17];19(1):172-87. Available from: https://doi.org/10.3390/jtaer19010010

13. He X, Liu Y. Knowledge evolutionary process of Artificial intelligence in E-commerce: Main path analysis and science mapping analysis. Expert Syst With Appl [Internet]. 2023 Sep [cited 2024 Jul 17]:121801. Available from: https://doi.org/10.1016/j.eswa.2023.121801

14. Kreutzer RT, Sirrenberg M. Understanding Artificial Intelligence [Internet]. Cham: Springer International Publishing; 2019. Fields of Application of Artificial Intelligence—Security Sector and Military Sector; [cited 2024 Jul 17]; p. 225-33. Available from: https://doi.org/10.1007/978-3-030-25271-7_9

15. Akilandeeswari S, Nagpal P, Vinotha C, Jain K, Chatterjee R, Gundavarapu MR. Transforming E-Commerce: Unleashing the Potential of Dynamic Pricing Optimization Through Artificial Intelligence for Strategic Management. Migration Letters. 2024;21(S3):1250-60. ISSN: 1741- 8984. ISSN: 1741-8992

16. Thamik H, Wu J. The Impact of Artificial Intelligence on Sustainable Development in Electronic Markets. Sustainability [Internet]. 2022 Mar 18 [cited 2024 Jul 17];14(6):3568. Available from: https://doi.org/10.3390/su14063568

17. Raimundo R, Rosário A. The impact of artificial intelligence on data system security: A literature review. Sensors (Basel) [Internet]. 2021;21(21):7029. Available from: http://dx.doi.org/10.3390/s21217029

18. Oyekunle DO, Boohene D. Digital Transformation Potential: The Role of Artificial Intelligence in Business. Int J Prof Bus Rev. March 11, 2024. Available at SSRN: https://ssrn.com/abstract=4806733

19. Dutta MS, R DA, Ema DUP, Balasundaram DR. Revolutionizing Electronics E-Commerce: Harnessing The Power Of Artificial Intelligence In E-Marketing Strategies. Migrat. Lett. [Internet]. 2024Feb.17 [cited 2024Jul.17];21(S6):207-20. Available from: https://migrationletters.com/index.php/ml/article/view/7891

20. Mуkhalchenko H, Tytarenko M. Data Analytics and Personalized Marketing Strategies in E-commerce Platforms. FEL [Internet]. 2023 Sep. 25 [cited 2024 Jul. 17];3(3):114-38. Available from: https://www.futurity-econlaw.com/index.php/FEL/article/view/149

21. Zhang X, Guo F, Chen T, Pan L, Beliakov G, Wu J. A Brief Survey of Machine Learning and Deep Learning Techniques for E-Commerce Research. J Theor Appl Electron Commer Res [Internet]. 2023 Dec 4 [cited 2024 Jul 17];18(4):2188-216. Available from: https://doi.org/10.3390/jtaer18040110

22. Priatna W, Warta J, Lestari TS. E-Commerce Fraud Detection Using Support Vector Machine and Naïve Bayes Algorithm. IJARCCE [Internet]. 2022 Nov 30 [cited 2024 Jul 17];11(12). Available from: https://doi.org/10.17148/ijarcce.2022.111205

23. Bakir H, Chniti G, Zaher H. E-Commerce Price Forecasting Using LSTM Neural Networks. Int J Mach Learn Comput [Internet]. 2018 Apr [cited 2024 Jul 17];8(2):169-74. Available from: https://doi.org/10.18178/ijmlc.2018.8.2.682

24. Singh N. AI in Inventory Management: Applications, Challenges, and Opportunities. Int J Res Appl Sci Eng Technol [Internet]. 2023 Nov 30 [cited 2024 Jul 17];11(11):2049-53. Available from: https://doi.org/10.22214/ijraset.2023.57010

25. Eulene. Sentiment analysis of Amazon reviews using Natural Language Processing [Internet]. Medium. 2024 [cited 2024 Jul 17]. Available from: https://medium.com/@eulene.tech/sentiment-analysis-of-amazon-reviews-using-natural-language-processing-6ecb3e253a5e

26. Mantri A. Real-Time Data Streaming and AI Enhancements: E-Commerce Live Streaming Shopping. Int J Comput Eng [Internet]. 2024 Jun 16 [cited 2024 Jul 17];5(5):22-32. Available from: https://doi.org/10.47941/ijce.2004

27. Ming G. Online Retail Service Quality, Who Satisfy? - Customer Satisfaction. J Digit Realism Amp Mastery (DREAM) [Internet]. 2022 Jul 8 [cited 2024 Jul 17];1(01):1-11. Available from: https://doi.org/10.56982/journalo.v1i01.6

28. Wang Q, Liu X, Liu W, Liu AA, Liu W, Mei T. MetaSearch: Incremental Product Search via Deep Meta-Learning. IEEE Trans Image Process [Internet]. 2020 [cited 2024 Jul 17];29:7549-64. Available from: https://doi.org/10.1109/tip.2020.3004249

29. Rajasekaran V. Predictive customer churn analysis in E-commerce using random forest classifier: A python example [Internet]. Medium. 2023 [cited 2024 Jul 17]. Available from: https://medium.com/@Vidya_Rajasekaran/predictive-customer-churn-analysis-in-e-commerce-using-random-forest-classifier-a-python-example-998447a1be63

30. Han M, Yang T, Zhong J, Zhong Y. AI applications and supply chain concentration. Appl Econ Lett [Internet]. 2023 May 10 [cited 2024 Jul 17]:1-5. Available from: https://doi.org/10.1080/13504851.2023.2210813

31. Johnson Taiwo MBA. Leveraging Al for dynamic pricing strategies [Internet]. Linkedin.com. 2024 [cited 2024 Jul 17]. Available from: https://www.linkedin.com/pulse/leveraging-al-dynamic-pricing-strategies-o-johnson-taiwo-mba-arksf

32. Williams R. The intersection of web design and AI: Exploring the exciting possibilities [Internet]. Linkedin.com. 2023 [cited 2024 Jul 17]. Available from: https://www.linkedin.com/pulse/intersection-web-design-ai-exploring-exciting-robert-williams/

33. BetterCommerce. How integrating AI can boost your E-commerce operations? [Internet]. Linkedin.com. 2024 [cited 2024 Jul 17]. Available from: https://www.linkedin.com/pulse/how-integrating-ai-can-boost-your-e-commerce-operations-8fqgc/

34. Prentice C, Nguyen M. Engaging and retaining customers with AI and employee service. J Retail Consum Serv [Internet]. 2020 Sep [cited 2024 Jul 17];56:102186. Available from: https://doi.org/10.1016/j.jretconser.2020.102186

35. Nagraj S, Palayyan BP. Personalized E-commerce based recommendation systems using deep-learning techniques. IAES Int J Artif Intell (IJ AI) [Internet]. 2024 Mar 1 [cited 2024 Jul 17];13(1):610. Available from: https://doi.org/10.11591/ijai.v13.i1.pp610-618

36.  Shankar V. Big Data and Analytics in Retailing. NIM Mark Intell Rev [Internet]. 2019 May 1 [cited 2024 Jul 17];11(1):36-40. Available from: https://doi.org/10.2478/nimmir-2019-0006

37. Kulkarni A, Shivananda A, Kulkarni A. Natural Language Processing Projects [Internet]. Berkeley, CA: Apress; 2021. Creating an E-commerce Product Categorization Model Using Deep Learning; [cited 2024 Jul 17]; p. 157-83. Available from: https://doi.org/10.1007/978-1-4842-7386-9_6

38. Mileva G. The role of AI personalization in eCommerce growth [Internet]. Influencer Marketing Hub. 2023 [cited 2024 Jul 17]. Available from: https://influencermarketinghub.com/ai-personalization-ecommerce/

39. Berger B. Enhancing customer satisfaction with AI chatbots [Internet]. Sikhara Marketing. 2024 [cited 2024 Jul 17]. Available from: https://www.sikharamarketing.com/enhancing-customer-satisfaction-with-ai-chatbots/

40. Kalkha H, Khiat A, Bahnasse A, Ouajji H. The rising trends of smart e-commerce logistics. IEEE Access [Internet]. 2023 [cited 2024 Jul 17]:1. Available from: https://doi.org/10.1109/access.2023.3252566

41. Petruk M. AI product recommendation for E-commerce [Internet]. WeSoftYou. 2023 [cited 2024 Jul 17]. Available from: https://wesoftyou.com/outsourcing/ai-product-recommendation-for-e-commerce/

42. Albshaier L, Almarri S, Hafizur Rahman M. A Review of Blockchain’s Role in E-Commerce Transactions: Open Challenges, and Future Research Directions. Computers [Internet]. 2024 Jan 17 [cited 2024 Jul 17];13(1):27. Available from: https://doi.org/10.3390/computers13010027

43. Wong AN, Marikannan BP. Optimising e-commerce customer satisfaction with machine learning. J Phys [Internet]. 2020 Dec [cited 2024 Jul 17];1712:012044. Available from: https://doi.org/10.1088/1742-6596/1712/1/012044

FINANCING

The authors did not receive financing for the development of this research.

CONFLICT OF INTEREST

The authors declare that there is no conflict of interest.

AUTHORSHIP CONTRIBUTION

Conceptualization: Petar Halachev.

Data curation: Petar Halachev.

Research: Petar Halachev.

Drafting - original draft: Petar Halachev.

Writing - proofreading and editing: Petar Halachev.