Project – The Influence of AI-Generated Product Recommendations on Consumer Purchase Decisions and Brand Trust. A Study of Online retail consumers in Lagos State.

CHAPTER ONE

INTRODUCTION

1.1 Background to the Study

The rapid development of digital technology has transformed the way consumers search for, evaluate, compare, and purchase products. Online retailing enables consumers to access different products, compare prices, read customer reviews, and complete transactions without visiting physical stores. However, the increasing volume of information available on digital platforms can make it difficult for consumers to identify products that adequately satisfy their needs and preferences. In response to this challenge, online retailers increasingly rely on artificial intelligence (AI) to analyse consumer information and provide personalised product recommendations. These recommendations may be generated from browsing histories, previous purchases, product ratings, search behaviour, and similarities between the preferences of different consumers. According to Xiao and Benbasat (2007), product recommendation agents can support online consumers by helping them identify relevant products and make purchasing decisions. Consequently, AI-generated product recommendations have become an important component of digital retail marketing because they can influence how consumers discover products and allocate their attention within online shopping environments.

Artificial intelligence has expanded the capabilities of traditional recommendation systems by enabling platforms to process large volumes of consumer data and identify patterns that may not be easily recognised through manual analysis. Depending on the system, AI-generated recommendations may use collaborative filtering, content-based filtering, machine-learning models, or combinations of these techniques to predict which products individual consumers are likely to find useful. For example, an online retailer may recommend products similar to those previously viewed by a consumer or suggest complementary products based on earlier purchases. These systems are intended to improve product relevance, reduce search effort, and make online shopping more convenient. Adomavicius and Tuzhilin (2005) explain that recommender systems provide personalised information by predicting the utility of items for individual users, while Xiao and Benbasat (2007) discuss their role in facilitating online product search and decision-making. The influence of these systems, however, depends not only on the technical sophistication of the underlying algorithms but also on whether consumers consider the recommendations relevant, useful, and appropriate to their needs.

The relationship between product recommendations and consumer purchase decisions is particularly important because recommendations can affect the alternatives consumers consider before making a final choice. Online shoppers may encounter AI-generated suggestions on homepages, product pages, search results, shopping applications, and checkout screens. Such suggestions can expand consumers’ consideration sets, direct attention towards particular brands, and reduce the effort required to compare available products. In an empirical field experiment involving an online book retailer, Li, Grahl, and Hinz (2022) found that personalised recommendations increased consumers’ propensity to purchase and the value of their shopping baskets, with changes in the consideration set helping to explain these effects. This evidence indicates that recommendations can influence more than product discovery; they may also shape the decision process leading to a transaction. Nevertheless, the presence of a recommendation does not automatically guarantee a purchase, because consumers may reject suggested products when prices are unattractive, product information is inadequate, or the recommendations do not correspond with their preferences. It is therefore important to examine how AI-generated product recommendations relate to consumer purchase decisions among online retail consumers in Lagos State.

Beyond purchase decisions, AI-generated product recommendations raise important questions about consumer trust. Brand trust refers to consumers’ willingness to rely on a brand based on their beliefs about its reliability, honesty, and ability to deliver the expected value. In online retailing, consumers often make purchasing decisions without physically inspecting products or interacting directly with sellers, making trust an important consideration in the buying process. Recommendations that appear relevant, accurate, and consistent with consumers’ preferences may strengthen confidence in a retailer’s ability to understand their needs. Conversely, recommendations that appear misleading, repetitive, commercially biased, or unrelated to the consumer’s interests may create doubts about the retailer’s motives and the quality of its service. Research by Weidig, Kuehnl, and colleagues (2023) on attributional cues in personalised recommendations indicates that explaining the basis of recommendations can increase consumers’ self-attribution, strengthen trust in recommendations, and reduce psychological reactance under relevant conditions. This suggests that the way AI recommendations are generated and presented may influence consumer perceptions of the retailer or brand providing them.

The effect of AI-generated recommendations on trust is not necessarily positive in every situation because personalisation depends on the collection and processing of consumer information. Although consumers may appreciate recommendations that reflect their preferences, they may become uncomfortable when a platform appears to know too much about their activities or uses their information in ways they do not understand. Aguirre et al. (2015) examined the personalisation paradox and showed that the effectiveness of personalised advertising can depend on how consumer information is obtained and how personalisation is implemented. Similarly, Bleier and Eisenbeiss (2015) demonstrated that the effectiveness of personalised online advertising varies according to the relevance and context of its presentation. These studies suggest that personalisation may improve consumer responses when it is useful and appropriately implemented, but may create adverse reactions when it appears intrusive or poorly matched to the consumer’s expectations. For online retailers, therefore, the challenge is not simply to generate more recommendations but to ensure that recommendations are relevant, understandable, and consistent with responsible consumer-data practices. These issues are significant when examining whether AI-generated product recommendations strengthen or weaken trust in the brands consumers encounter online.

The Nigerian retail environment provides a relevant context for investigating these relationships because digital commerce allows consumers to access products and sellers beyond the immediate limitations of physical retail outlets. Lagos State is particularly suitable for this study because it is a major commercial centre with a large and diverse consumer market, extensive retail activity, and substantial participation in digital business. Online retail consumers in Lagos may encounter product recommendations through retailer websites, shopping applications, social-commerce platforms, and other digital channels. Their responses may vary according to their familiarity with online shopping, previous experiences with retailers, perceptions of product quality, price sensitivity, concerns about online fraud, and confidence in digital payment and delivery arrangements. The National Bureau of Statistics (2020), through its Nigerian Gross Domestic Product Report, provides evidence of the importance of trade and other service activities in the national economy, although such economic data do not directly establish the extent to which AI recommendation systems are used by Lagos retailers. This distinction makes local empirical investigation necessary rather than assuming that findings from foreign e-commerce markets apply uniformly to Nigerian consumers.

Existing international research has established that recommendation systems can influence online product search, consumer choice, purchase behaviour, and responses to personalised marketing (Xiao & Benbasat, 2007; Li et al., 2022). Other studies have drawn attention to the importance of transparency, privacy, perceived relevance, and the possible unintended consequences of personalisation (Aguirre et al., 2015; Bleier & Eisenbeiss, 2015; Weidig et al., 2023). However, the extent to which these relationships operate among online retail consumers in Lagos State requires context-specific examination. In particular, there is a need to investigate whether consumers’ exposure to AI-generated product recommendations is associated with their purchase decisions and whether their perceptions of such recommendations are related to trust in the brands offering them. The distinction between purchase decisions and brand trust is important because a consumer may buy a recommended product for convenience or price reasons without necessarily developing lasting confidence in the brand. This study therefore examines the influence of AI-generated product recommendations on consumer purchase decisions and brand trust among online retail consumers in Lagos State, with attention to the relevance of recommendations, their relationship with purchase decisions, and the trust implications of AI-supported personalisation.

1.2 Statement of the Problem

Online retailers increasingly use AI-generated product recommendations to help consumers identify products, discover alternatives, and make purchasing decisions. Although these systems are designed to improve shopping convenience and relevance, their actual influence may differ across consumers and retail contexts. Some consumers may regard recommendations as useful guidance that reduces the time and effort required to search for products, while others may ignore them because they appear repetitive, inaccurate, or commercially motivated. Where recommendations do not correspond with consumers’ preferences, they may fail to encourage purchases and may even distract consumers from more suitable alternatives. Although previous studies have demonstrated that recommendation systems can influence consumer choice and purchase behaviour, the extent of this relationship among online retail consumers in Lagos State remains an empirical question that requires systematic investigation.

A second problem concerns the relationship between AI-generated recommendations and brand trust. Online consumers frequently depend on information supplied by retailers because they cannot physically inspect products before purchase. When recommendations appear accurate and relevant, consumers may interpret them as evidence that a retailer understands their needs and can provide useful assistance. However, recommendations based on incomplete information, unexplained data collection, or apparent commercial bias may generate suspicion about the retailer’s intentions. Aguirre et al. (2015) show that personalisation can create tension between marketing effectiveness and consumers’ responses to the collection and use of personal information. This raises a practical concern for online retailers in Lagos State: the use of AI recommendations to encourage immediate purchases may not necessarily strengthen consumer confidence in the brands presenting them. The relationship between AI-generated recommendations and brand trust therefore needs to be examined rather than assumed to be automatically positive.

Another problem is that consumer purchase decisions and brand trust represent related but distinct outcomes. A consumer may purchase a recommended product because it is affordable, convenient, highly rated, or immediately available, without developing confidence in the retailer or brand. Similarly, a consumer may trust a brand but reject a particular recommendation because the product does not meet a current need or budget. Li et al. (2022) demonstrated that personalised recommendations can increase purchasing propensity partly by changing the range and depth of products consumers consider. Nevertheless, evidence of increased purchasing does not by itself establish that consumers develop stronger trust in the brand providing the recommendation. Failure to distinguish these outcomes may lead retailers to evaluate recommendation systems mainly through sales performance while overlooking the longer-term implications for consumer confidence and brand relationships. This study consequently considers purchase decisions and brand trust as separate outcomes of interest.

Furthermore, differences in consumers’ digital experiences, privacy expectations, purchasing power, and previous interactions with online retailers may affect how they respond to AI-generated recommendations. Findings from studies conducted in other countries cannot automatically be generalised to Lagos State, where consumer experiences and the practical conditions of online shopping may differ. Without context-specific evidence, online retailers may lack a clear basis for determining whether AI-generated recommendations are associated with actual purchasing decisions and stronger brand trust among their customers. The central problem addressed by this study is therefore the insufficient local empirical understanding of the relationship between AI-generated product recommendations, consumer purchase decisions, and brand trust among online retail consumers in Lagos State. The study seeks to provide evidence that can help clarify these relationships and inform the responsible use of AI-supported product personalisation in online retailing.

1.3 Aim and Objectives of the Study

The aim of this study is to examine the influence of AI-generated product recommendations on consumer purchase decisions and brand trust among online retail consumers in Lagos State.

The specific objectives are to:

1. Examine the extent to which online retail consumers in Lagos State are exposed to AI-generated product recommendations.
2. Determine the relationship between AI-generated product recommendations and consumer purchase decisions among online retail consumers in Lagos State.
3. Assess the relationship between AI-generated product recommendations and brand trust among online retail consumers in Lagos State.
4. Examine the extent to which perceived relevance and accuracy of AI-generated product recommendations are associated with consumer purchase decisions and brand trust.

1.4 Research Questions

The following research questions guide the study:

1. To what extent are online retail consumers in Lagos State exposed to AI-generated product recommendations?
2. What is the relationship between AI-generated product recommendations and consumer purchase decisions among online retail consumers in Lagos State?
3. What is the relationship between AI-generated product recommendations and brand trust among online retail consumers in Lagos State?
4. To what extent are perceived relevance and accuracy of AI-generated product recommendations associated with consumer purchase decisions and brand trust?

1.5 Research Hypothesis

The study will test the following null hypothesis at the 0.05 level of significance:

H₀: AI-generated product recommendations have no statistically significant relationship with consumer purchase decisions and brand trust among online retail consumers in Lagos State.

1.6 Significance of the Study

This study is expected to be useful to online retailers, marketing managers, consumers, technology developers, researchers, and other stakeholders interested in the use of artificial intelligence in digital commerce.

Online retailers and marketing managers: The findings may help online retailers understand whether AI-generated recommendations are associated with consumer purchasing decisions and brand trust. This understanding may support better decisions about recommendation relevance, product presentation, personalisation strategies, and customer relationship management.

AI developers and digital-platform managers: The study may provide insights into the consumer-facing characteristics of recommendation systems, particularly perceived relevance and accuracy. These insights may encourage the development of systems that offer useful recommendations while reducing irrelevant suggestions and avoidable consumer frustration.

Consumers: The study may contribute to greater awareness of how AI-generated recommendations can influence product discovery and purchasing behaviour. It may also draw attention to the importance of evaluating recommended products, comparing alternatives, and considering privacy and data-use practices before making purchasing decisions.

Marketing and business researchers: The study may contribute to academic discussions on AI-enabled marketing, consumer decision-making, personalisation, and brand trust. By focusing on online retail consumers in Lagos State, it may provide context-specific evidence that can be compared with findings from other geographical settings.

Policy and consumer-protection stakeholders: The findings may be useful in discussions about responsible digital marketing, transparency in personalised recommendations, and consumer confidence in online retailing. The study may also identify areas where further research or consumer education is needed.

1.7 Scope of the Study

The geographical scope of the study is Lagos State, Nigeria. The study will focus on consumers who purchase goods through online retail platforms and who have encountered, or can identify, AI-generated or personalised product recommendations during online shopping. The content scope covers exposure to AI-generated product recommendations, consumer purchase decisions, brand trust, and the perceived relevance and accuracy of recommendations. The study will consider consumers’ reported experiences and perceptions rather than attempting to audit the technical architecture of individual retailers’ recommendation algorithms. The empirical scope will be limited to the respondents selected for the study and the data collected during the research period. Consequently, findings should be interpreted within the study’s sampling approach and should not automatically be generalised to every online consumer in Nigeria.

1.8 Limitations of the Study

One potential limitation is that respondents may not always know whether a product recommendation was generated by artificial intelligence, a rule-based system, or another form of personalisation. To reduce this problem, the research instrument should provide a clear, accessible explanation of AI-generated product recommendations and examples of common recommendation formats. Another limitation is that self-reported purchase behaviour may be affected by memory errors or respondents’ desire to provide socially acceptable answers. The study may also face difficulties obtaining a sample that adequately represents the diversity of online retail consumers across Lagos State. Time and financial constraints may further limit the number of respondents or retail platforms covered. These limitations will be addressed, as far as practicable, through clear questionnaire wording, appropriate sampling procedures, and transparent reporting of the study’s methodology and findings.

1.9 Operational Definition of Terms

Artificial Intelligence (AI): Computer-based techniques that enable systems to perform tasks such as analysing data, identifying patterns, making predictions, and generating recommendations.

AI-generated product recommendations: Product suggestions produced or supported by AI-based systems using information such as browsing behaviour, purchase history, product attributes, or patterns in consumer preferences.

Consumer purchase decisions: The choices consumers make when deciding whether to buy a product, which product or brand to select, and whether to proceed with a purchase after evaluating available alternatives.

Brand trust: A consumer’s confidence that a brand is reliable, honest, and capable of delivering the value or performance it promises.

Online retail consumers: Individuals who search for, evaluate, or purchase goods through online retail websites, applications, or other digital shopping channels.

Perceived relevance: The extent to which consumers believe that recommended products match their needs, preferences, interests, or current shopping goals.

Perceived accuracy: The extent to which consumers believe that recommendations correctly reflect their preferences or identify products likely to satisfy their requirements.

Personalisation: The adaptation of product suggestions or shopping content to an individual consumer using available information about that consumer or patterns among similar consumers.

1.10 Organisation of the Study

The study is organised into five chapters. Chapter One presents the introduction, background to the study, statement of the problem, aim and objectives, research questions, hypothesis, significance, scope, limitations, and operational definitions. Chapter Two reviews relevant conceptual, theoretical, and empirical literature on AI-generated product recommendations, consumer purchase decisions, and brand trust. Chapter Three explains the research methodology, including the research design, study population, sample size and sampling technique, research instrument, validity and reliability procedures, data-collection process, and methods of data analysis. Chapter Four presents and analyses the data collected from respondents and discusses the findings in relation to the research questions and hypothesis. Chapter Five provides the summary of findings, conclusion, recommendations, and suggestions for further research.

Project – The Influence of AI-Generated Product Recommendations on Consumer Purchase Decisions and Brand Trust. A Study of Online retail consumers in Lagos State.