1. Introduction

In the wave of digital transformation, Artificial Intelligence (AI) and Machine Learning (ML) technologies are playing an important role in various domains, especially in marketing decision making. The integration of AI and M intelligence has shifted traditional marketing decisions towards a more data-driven approach, enabling decision makers to gain deeper insights into market data and improve their strategies [1]. As a result, AI and M intelligence have received increasing attention in recent years. The combination of AI and ML facilitates the extraction and analysis of large amounts of data, helping marketers to better understand customer needs and provide more personalized interactions. By applying AI and ML, marketers can enhance the customer experience and improve their decision-making process, which not only improves operational efficiency but also increases market competitiveness, creating a win-win situation for both customers and organizations. The purpose of this paper is to explore how AI and ML can enhance customer experience and help marketers make decisions. The main objective is to analyze the current applications of AI and Machine Learning in the market and identify the existing gaps.

2. Methods

2.1. Literature Review

2.1.1. Selection Criteria for Relevant Studies

In order to ensure the relevance and quality of the selected studies, this paper sets the following inclusion criteria: first of all, select the studies published in recent years (for example, in the past five years) to ensure that the articles are up-to-date, and the publication time is more than five years, unless it is a classic literature, which is an important basic contribution to the current study. Primarily, academic journal articles, conference papers, systematic reviews and empirical studies were included, and it was ensured that the subject matter was hugely and highly relevant: the studies had to explicitly deal with the application of AI and ML in marketing, with preference given to studies with a rigorous methodology as well as to quantitative studies consisting of a clear methodology for data analysis.

2.1.2. Purpose of The Study

This paper intends to identify the main purpose and research questions of the study, describe the methodology and data analysis techniques used in the study, summarize the main findings and conclusions of the study, identify the limitations and potential biases mentioned in the study, and finally, through a comprehensive analysis of the extracted information, identify the main areas of application of AI and ML in marketing, the effects, the challenges faced, and the future research directions.

2.2. Data Collection

Based on the results of the literature review and the nature of the ML techniques studied, and according to the classification of ML specific algorithms as well as the scenarios of practical applications, they are categorized into two classes from the architectural level in order to better show the applications of ML. The first category of this classification shows the application of ML technology in the PPAC portfolio at this stage, i.e., predictive analytics (P), personalization (P), ad optimization (A), and customer experience enhancement (C) in the marketing portfolio. In the second half, the algorithms (e.g., supervised learning, unsupervised learning, deep learning) used by ML in this process are fully demonstrated. In this paper, this architecture is used to determine whether the selected article is relevant to its topic or not. The articles that satisfy the PPAC marketing mix and to which ML is applied are then analyzed and classified in the figure 1and table 1.

Figure 1: ML in PPAC marketing mix.

Table 1: ML applications in marketing based on PPAC marketing mix.

3. Applications of Artificial Intelligence in Marketing

Continuing from the previous section, this section divides the application of AI and ML in marketing into four segments. In this section, we will continue to detail the specific definitions and scope of application of these four segments and will analyze the advantages and enhancements of ML in marketing with case studies.

3.1. Predictive Analysis

3.1.1. Defining Analytics

Predictive analytics is a process that utilizes statistical techniques, mining data, machine learning, and other methods to obtain information from historical information, build models, and predict future outcomes [2]. Eckerson et al. found that although the process applies a lot of cutting-edge knowledge in statistics and advanced mathematics is intimidating. But its role can make decision makers more instinctive and intuitive to react to customers and business, improve marketing efficiency and increase the competitiveness of enterprises [2].

3.1.2. Case Study on Predictive Modeling

In Cheng's article, a machine learning classification algorithm is used to build a purchase prediction model to grasp customers' purchase intentions for coupons on the Palm Life App [3]. Zhou, on the other hand, first classified customers using their historical data with the K-means algorithm and finally predicted customer churn with the Random Forest algorithm [4].Weirong Wang et al., in their article, similarly first used the k-nearest neighbor algorithm to categorize the customers and applied an online extreme learning machine to predict the customer electricity load forecast [5]. These cases have shown that machine learning can make good use of historical customer data and segment customers to predict future customer behavior, so that decision makers can take into account the weight of the impact of changes in customer behavior on the development of the company to formulate appropriate countermeasures to ensure that the company's interests can be maximized in this prediction model.

3.2. Personalization

3.2.1. Definitional Analysis

Personalization in this review specifically refers to the use of machine learning algorithms in personalized marketing, which can help companies provide customized services and recommendations based on customers' historical behaviors and preferences, effectively improving customer satisfaction and conversion rates.

3.2.2. Examples of Improving Customer Engagement

Vinay Pitchika et al. in their article used ANN and CNN models to diagnose periodontal cases after and help decision makers to integrate personalized cases [6].Yaghtin et al. for the specific case of a B2B company with the help of the fusion of machine learning and ML, used CRISP-DM and SEMMA to perform data mining followed by LASSO logistic regression to classify customer preferences and create a PIS system of customers to provide personalized services [7].Messaoudi Fayça et al. in their article built a deep learning DNCF model and found that it was effective in increasing the recall rate to a certain extent reflecting an increase in the probability of recommending content of interest to customers after content filtering [8]. Alabdulrahman, R., & Viktor, H. in 2021 distinguished a single classifier (DT) for grey sheep users and designed personalized recommendations for them [9]. It can be seen that in recent years, the application of machine learning in marketing has greatly improved the development of customer personalization, which can lead to a better e-commerce experience for the customer while increasing the customer conversion rate on the business side and increasing revenue.

3.3. Advertising Optimization

3.3.1. Definition Analysis

Advertising optimization refers to optimizing the effect of advertising through various technical means and strategies to maximize the return on investment (ROI) of advertising. This process includes the adjustment and improvement of various aspects such as ad content, audience, delivery time and delivery channels. In this review, we focus on ad optimization after incorporating machine learning.

3.3.2. Results from Recent Implementations

Ava Hajian, Russell Sadeghi, Victor R. Prybutok & Chang E. Koh. in their article Improving the Accuracy of Topics for Ad Push with Topic Modeling in NLP [10].Qin Wu. found that collaborative filtering algorithms can be utilized to improve ads by augmenting the modeling of user preference precision and quality, which greatly improves the yield of advertisements [11].Li Chunhui., on the other hand, well improves the recall and return of the advertisement recommendation algorithm by dividing the customers and advertisements into two networks for learning, and then finally utilizes CNNs to capture the higher-order interaction information [12]. From the above case, it is easy to see that the use of machine learning can well mine the historical preferences and likes of customers, so as to push the corresponding types of advertisements for the target customers thus improving the return rate of advertisements and increasing the revenue.

3.4. Customer Experience Enhancement

3.4.1. Definition Analysis

The core objective of customer experience is to ensure that customers are delighted and satisfied at every touch point throughout the consumption process. This includes every step of the process from initial brand awareness, through the purchase process to after-sales service. A good customer experience improves customer loyalty, increases customer lifetime value, and promotes positive word-of-mouth. In the digital era, companies are increasingly focusing on optimizing customer experience through data analytics and technology. In this paper, we focus on techniques such as virtual robots that utilize ML and AI technologies to achieve user experience enhancement.

3.4.2. Impact on Customer Satisfaction and Customer Retention

Manju, Aman Jatain & Deepti Singh. In the article it was found that existing chatbots have been able to apply large language models to accomplish basic communication with customers with precise needs and good enhancement of user experience [13]. Samadrita Ghosh, Stephanie Ness & Shruti Salunkhe. et al. found that it is possible to make good use of ML for the mining of historical data to predict the customer needs combined with NLP techniques for multi-channel interaction with customers to add consumer satisfaction and loyalty [14].

4. Discussion

4.1. Challenges and Limitations

Personalized marketing requires large amounts of customer data, which raises serious privacy and security concerns. Customers are increasingly concerned about their data rights and demand transparency and control [7,8]. At the same time, complex machine learning models, especially deep learning models, often lack transparency and are seen as “black boxes”. This can make the decision-making process difficult to understand and trust, especially in business decisions. Integrating new AI and ML solutions with existing marketing systems can be technically challenging and resource intensive.

4.2. Applications in Practice

4.2.1. Best Practices for Implementing AI and ML in Marketing

Currently, as pointed out in Vinay Pitchika, Martha Büttner & Falk Schwendicke, combining the techniques of AI and ML with human insight and leveraging the cooperation of multiple disciplinary fields and requiring a large number of customers can minimize human bias against AI algorithms and improve the fairness of AI algorithms. algorithms' fairness [6].

4.2.2. Strategies to Overcome Current Challenges

Education and training: invest in training marketing teams to understand the concepts of AI and ML and to be able to utilize these technologies effectively. Ethical AI practices: develop and follow ethical guidelines for AI use, focusing on transparency, fairness, and accountability.

Interpretability and Transparency: investigate how to improve the interpretability and transparency of complex AI and ML models and develop new interpretive techniques.

Bias and fairness: research on how to detect, quantify and eliminate bias in machine learning models to ensure that all user groups are treated fairly.

4.3. Research Limitations and Future Directions

4.3.1. Potential Interdisciplinary Collaborations

Behavioral Economics and AI: Combining insights from behavioral economics with AI technology to develop more psychologically informed marketing strategies [15].

Ethics and AI: Collaborate with ethicists and legal experts to address ethical and regulatory challenges of AI in marketing.

Emerging trends and technologies

Edge computing: with the proliferation of IoT devices, edge computing enables real-time data processing and personalized marketing at the point of interaction.

AI-driven content creation: advances in natural language generation (NLG) and computer vision can automate and personalize content creation, making marketing campaigns more dynamic and engaging.

Future research needs to focus more on multi-dimensional data integration, real-time, and user experience in real-world applications to enable the full application and optimization of machine learning techniques in marketing. These studies not only promote the progress of academia, but also provide practical application references for the industry to promote innovation and development in the field of marketing.

5. Conclusion

5.1. Research Summary

With AI and ML technologies, marketers are able to segment customer groups more accurately and use this to provide personalized services to increase customer engagement and conversion rates. Using models to predict market trends and customer needs helps companies develop more effective marketing strategies and maximize the return on investment (ROI) of advertising. Enriched marketing theory by studying the application of AI and ML in marketing, especially in the areas of customer behavior analysis, personalized marketing and ad optimization. Introduced and promoted new data analysis methods and machine learning algorithms to improve the accuracy and practicality of marketing research. The integration and cooperation between marketing and multiple disciplines such as data science, artificial intelligence, and behavioral economics have been promoted, and the development of comprehensive research has been advanced.

5.2. Future Outlook

As AI and ML technologies continue to advance, personalized marketing will become more accurate and efficient, providing dynamic personalized experiences based on real-time data. Encourage interdisciplinary collaboration across disciplines such as marketing, data science, AI, and ethics to solve complex marketing problems. Further research to improve the interpretability and fairness of AI and ML models to ensure that their use in marketing is widely trusted and accepted. Through these researches and applications, AI and ML will play an increasingly important role in marketing, promoting the continuous improvement of marketing efficiency and effectiveness, and at the same time providing rich materials and directions for academic research.