Page 859
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue II, February 2026
AI-Driven Marketing Communications and Herbal Drugs Brand
Awareness Among Retirees in Southwest Nigeria
Sadamoro, F. (Ph. D)1, Olukorede, B. B. (Ph. D)2, Ayodele, O. O3
Department of Business Administration, Faculty of Management Sciences, Ekiti State University, Ado-
Ekiti
DOI: https://doi.org/10.51583/IJLTEMAS.2026.15020000075
Received: 24 February 2026; Accepted: 02 March 2026; Published: 17 March 2026
ABSTRACT
This study investigated the effects of AI-driven marketing communication messages on herbal drugs brand
awareness among retirees in Southwest Nigeria. Specifically, AI-personalized advertising, AI-optimized
promotional messaging, and AI-assisted interactive engagement were tested as the explanatory variables of brand
awareness among retirees in Southwest, Nigeria. Using a survey research design and Multiple Regression
analysis, the three proposed hypotheses were tested.
The findings revealed that AI-driven communication variables exert positive and statistically significant
influence on herbal drugs brand awareness among retirees. AI-assisted interactive engagement showed the
strongest predictive strength. The study concludes that AI-Driven marketing communication significantly affects
brand recall and recognition among retirees. It recommends among others that herbal drug firms adopt AI-based
personalization tools to improve message clarity, credibility, and audience targeting.
Keywords: Marketing Communications, Brand Awareness, Personalized Adverting, Optimized Promotion,
Interactive Engagement
INTRODUCTION
Traditional advertising and promotional methods are no longer the only tools used in marketing communication.
Businesses may now create data driven, customized messages that are sent via automated digital channels
through artificial intelligence. To increase message accuracy and efficacy, AI-driven marketing communication
combines machine learning algorithms, customer data systems, predictive analytics, and automated interaction
platforms. A separate demographic group, retirees have particular health concerns, which call for reasons for
making purchases, and ways of processing information.
Herbal medicine is a very common practice in Southwest Nigerian culture and health sector. However,
inappropriate product communication, mistrust, ignorance, and brand confusion continue to pavade the
populace. Therefore, platforms like automated health chatbots, WhatsApp messaging systems, SMS
notifications, and targeted social media ads, AI-enabled technologies offer the chance to provide retirees with
consistent, reliable, and personalized information.
There is dearth of empirical data on the efficacy of AI-enabled marketing in Nigeria's herbal drug sector,
especially among retirees in Southwest Nigeria, despite the technology's explosive global growth.
The majority of earlier research ignored intelligent communication technologies in favour of traditional
advertising, sales promotion, and human selling. This study fills that knowledge vacuum by examining how
retirees' brand awareness of herbal drugs is influenced by AI-driven promotional platforms.
Research Objectives
The broad objective of this study is to examine the effects of AI-driven marketing communication messages on
herbal drugs brand awareness among retirees in Southwest Nigeria.
Page 860
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue II, February 2026
Specific objectives are to:
i. Determine the effect of AI-personalized advertising on herbal drugs brand awareness among retirees.
ii. Assess the influence of AI-optimized promotional messaging on herbal drugs brand awareness. iii.
Evaluate the impact of AI-assisted interactive engagement on herbal drugs brand awareness
Research Hypotheses
H01: AI-personalized advertising does not significantly affect herbal drugs brand awareness among retirees in
Southwest Nigeria.
H02: AI-optimized promotional messaging does not significantly affect herbal drugs brand awareness among
retirees in Southwest Nigeria.
H03: AI-assisted interactive engagement does not significantly affect herbal drugs brand awareness among
retirees in Southwest Nigeria.
LITERATURE REVIEW
Conceptual Review
AI‑Personalized Advertising
Conventional advertising does not customize information for each receiver; instead, it conveys consistent
messages to large audiences. Segmentation, such as tailoring according to demographic categories, was the initial
step toward personalization (Kotler, 2014). But when compared to AI methods, this kind of segmentation is still
crude. Higher engagement, greater memory, stronger brand attitudes, and better conversion outcomes are the
objectives of AI-personalized advertising, which employs real-time data and computational models to deliver
personally tailored content across digital platforms (Eisenbeiss & Bleier, 2015).
According to Bleier, Harmeling and Palmatier (2020), AI is revolutionizing targeted advertising in the
automobile sector in a time when every click, search, and purchase is painstakingly documented. AI creates
customized marketing messages that directly address the requirements and interests of potential customers by
utilizing algorithms that sort through massive information. Dealerships can increase customer engagement and
boost sales by knowing the preferences and habits of each individual consumer (Bleier et al., 2020).
According to Bleier and Eisenbeiss (2015), AI is transforming personalized advertising by anticipating consumer
preferences, making tailored recommendations, and utilizing chatbots to respond to inquiries. Additionally, it
supports customized shopping experiences, real-time data analysis, targeted advertising, and raising customer
satisfaction levels. Artificial intelligence (AI) offers a number of benefits that have the potential to significantly
influence marketing campaigns.
Delivering highly specialized and targeted marketing messages to prospective purchasers is one of the most
noteworthy advantages. Businesses may learn a great deal about the tastes, actions, and historical purchase
histories of their customers by employing AI-driven algorithms to analyze large databases. This thorough
research makes it possible to create recommendations and offers that are unique to each customer (Bleier &
Eisenbeiss, 2015).
AI-Optimized Promotional Messaging
AI optimized promotional messaging refers to the use of artificial intelligence systems in the creation,
customization, and distribution of marketing communications. AI-driven solutions, in contrast to traditional
messaging, create content automatically, customize messages for each user profile, and maximize distribution
for engagement and conversion.
Page 861
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue II, February 2026
AI-Optimzed promotion focuses on continual optimization via computational models, automation, and
personalization. (Li, Monroe & Jurafsky, 2020). Li et al., (2020) state that AI-generated promotional content is
based on Natural Language Generation (NLG).
While adjusting to contextual elements like user preferences, platform, or campaign objectives, NLG systems
generate messages that represent brand voice. While contemporary methods use deep learning and transformer
structures to produce more cohesive and contextually relevant material, early frameworks concentrated on
structured text production (Reiter & Dale, 2000)
According to Radford, Wu, and Child (2019), Predictive models are used in AI-optimized messaging to foresee
engagement and conversion results prior to message delivery. Machine learning is used by optimization
frameworks to choose the best time, channels, and message variations. To study interaction patterns and optimize
promotional content repeatedly, gradient boosting, ensemble models, and neural networks are commonly used
(Nguyen, Li, & Zhao, 2021).
AI-Assisted Interactive Engagement
AI assisted interactive engagement, according to Zhao, Kumar, and Yang (2019), is the use of artificial
intelligence systems to support, enhance, and sustain real-time or nearly real-time interactions between humans
and digital systems, or between humans mediated by digital systems.
AI promotes engagement by deciphering user input, producing response actions, and modifying interactions
according to context. Conversational agents, such as chatbots and virtual assistants, are at the heart of AI-assisted
interactive engagement. They leverage user input to produce contextually relevant responses. These agents can
comprehend, interpret, and generate human-like language at scale through Natural Language Processing (NLP)
(Radford et al., 2019).
Radford, Wu, and Child (2019) added that adaptive content, targeted prompts, and interface modifications are
examples of personalization in interactive engagement that match system outputs to user requirements and
expectations. AI's capacity to comprehend changing user context is essential for real-time interactive interaction.
Systems can instantly modify their interaction techniques through methods like intent identification, sentiment
analysis, and multimodal input processing.
According to Mehrabi, Morstatter, Saxena (2021), sentiment analysis views user emotion as a signal that shapes
the course of interactions. Supportive reactions may be elicited by negative signals, while promotional or
reinforcing engagement may be guided by positive cues. AI-assisted interactive engagement differs from static
scripted interaction systems in that it is always adapting (Nguyen et al., 2021).
Brand Awareness
Branding is a notable topic that has been extensively studied by researchers studying brand identification and
the introduction of new products. Additionally, brands are more successful at creating profitable and enduring
relationships with consumers than regular unbranded products (Heath, 2016).
Presenting brands to customers can increase brand awareness by causing a stimulus-like reaction in them that
enables people to relate to, recognize, recall, and be generally aware of the brands, according to the reviewed
literature.
Established brands often use brand reinforcement tactics to bolster their brand awareness campaigns. Conversely,
the new products use advertising and promotion to increase product awareness among present and potential
consumers. Companies can use brand image management and attitude advertising as tactics to increase brand
recognition, claim (Heath, 2016).
The elements of consumers' value framework largely dictate how they behave in the marketplace. The value
framework for customers includes factors such as pricing, class connection, brand image, and overall market
awareness in relation to others (Heath, 2016).
Page 862
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue II, February 2026
Theoretical Review
Information Processing Theory
Information processing theory states that people selectively pay attention to, interpret, and retain messages. By
tailoring material to customers' interests and requirements, personalized advertising lowers cognitive load and
improves information retrieval and comprehension (Petty & Cacioppo, 1986).
Technology Acceptance Model (TAM)
In Technology Acceptance Model (TAM), new technology adoption is influenced by perceived utility and
usability (Davis, 1989). Relevant content sent by AI-personalized systems is more likely to be viewed as helpful
and less invasive, which increases engagement.
Elaboration Likelihood Model (ELM)
Elaboration Likelihood Model (ELM) stipulates that because customized messages are more likely to follow the
core route of persuasion because they are seen as more credible and relevant, resulting in stronger attitude
formation and deeper cognitive processing (Petty & Cacioppo, 1986).
Empirical Review of Related Literature
Bleier and Eisenbeiss (2015) conducted controlled experiments examining online display advertisements and
found that personalized ad content significantly improved click-through rates and visual attention. Their results
indicated that perceived relevance mediated the relationship between personalization and engagement. However,
they also observed that overly intrusive personalization reduced trust, suggesting a non-linear relationship which
raises concerns about generalizability, as the optimal level of personalization may vary across different
populations and cultural contexts.
Similarly, Wedel and Kannan (2016) demonstrated that algorithmic targeting improves message-context
congruence, thereby increasing engagement time and interaction depth. These findings underscore that AI-driven
personalization operates through attention capture mechanisms rooted in perceived relevance. However, their
study primarily emphasizes interaction metrics such as engagement time and depth, without examining
downstream outcomes like brand awareness, recall, or purchase intention, which are crucial for assessing
marketing effectiveness.
More recent large-scale digital experiments by Kumar et al. (2023) reported that machine learning–based ad
optimization significantly improved consumer interaction metrics compared to rule-based targeting systems.
Their study confirmed that predictive personalization yields measurable improvements in engagement quality,
not merely quantity. However, the research relied on digitally savvy populations, leaving uncertainty about how
predictive personalization performs among older adults or retirees, especially in contexts with lower digital
literacy or differing cultural attitudes toward technology
Arora et al. (2021) observed that individualized product recommendations increased brand recall and recognition
scores in experimental settings. Participants exposed to AI-tailored messages demonstrated stronger unaided
recall than those exposed to generic advertising. However, much of this evidence comes from Western consumer
samples, leaving open the question of whether similar cognitive responses occur among older adults or retirees
in developing markets.
Huang and Rust (2021) argued that AI enhances persuasion effectiveness by dynamically adapting message
framing to individual preferences. Empirical tests confirmed that adaptive framing increases purchase intentions
and favourable brand attitudes. However, their study primarily measures intentions and attitudes, rather than
actual behavioral outcomes such as brand awareness, product adoption, or long-term engagement, which limits
the practical applicability of the findings.
Page 863
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue II, February 2026
Mariani, Borghi and Gretzel (2023) found that AI chatbots in healthcare settings increased trust and engagement
when compared with static website information. The interactivity element contributed significantly to user
satisfaction. However, older populations exhibit mixed responses to digital health advertising. Studies suggest
that retirees may require simplified language, visual clarity, and explicit trust cues to respond positively to
AIpersonalized messages. This indicates the importance of contextual adaptation in markets such as Southwest
Nigeria.
Martin and Murphy (2017) demonstrated that perceived data misuse significantly reduces trust in personalized
advertising. When consumers believe their information is exploited without transparency, positive
personalization effects diminish.
METHODOLOGY
Research Design
The study used a survey methodology known as the descriptive technique, which uses a questionnaire to gather
data from participants about all of the explanatory variables being examined. It is generally accepted that this
approach works best for gathering information about experiences, emotions, motivations, and thoughts that are
hard to witness directly. Adoption of surveys is also seen to lessen the tendency toward manipulation.
The Population of the Study
According to Nigerian Union of Pensioners (2024), there are 145,232 retirees in the study's population. Taking
into account a five-year life expectancy, these are the current retirees in Southwest Nigeria as disclosed by the
pension boards and the Nigerian Union of Pensioners of several states. In particular, the population consists of
145,232 pensioners, including 19,080 from Ogun State, 18,600 from Ekiti State, 19,480 from Osun State, 24,740
from Oyo State, 39,398 from Lagos State, and 23,934 from Ondo State. This was based on the data the researcher
collected from the Nigeria Union of Pensioners.
Table 1: Population of the Study
States
Number of Retirees
Ekiti state:
18,600
Ogun state:
19,080
Osun state:
19,480
Oyo state:
24,740
Ondo
23,934
Lagos
39,398
Total
145,232
Source: Nigerian Union of Pensioners 2026
Sample and Sampling Techniques
Arising from the population, the sample for the study, using Yamane (1967) is 398 respondents. This is
considered to be the lowest level of acceptable responses to maintain a confidence level of 95% and a 5% error
level. The sample size is arrived at through a formula developed by Yamane (1967) as stated below.
Page 864
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue II, February 2026
𝑁
n = 1+ 𝑁(𝑒)2
For the population of 81,900, the sample size is computed below
n =
n =
n =
n =
n = 398
In order to ensure appropriate administration of research questionnaire for this study, heterogeneous
proportionate sampling technique was used. This is because the respondents are not in the same state. The
respondents in the Southwest States are considered to be influenced by different factors. This means their
experiences are influenced by heterogeneous factors. The population was grouped as:
𝑁!𝑛!
n =
𝑁
Table 2: Sample Size for Each State
States
Computations
Sample Size
Ekiti state:
50
Ogun state:
53
Osun state:
54
Oyo state:
=
68
Ondo
=
65
Lagos
=
108
Total
398
Source: Author’s computation, 2026
Research Instrument
Data were this study shall be gathered from primary sources by the use of a well structured questionnaire of four
(4)-point Likert scale, adapted from Onyiengo (2014) which shall be divided into six sections of A – F. Section
Page 865
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue II, February 2026
A gathered information on the demographic characteristics of the respondents. Sections B,C and D focused on
marketing communications and herbal drugs brand awareness variables.
Method of Data Analysis and Model Specifications
Analysing data for this study, a combination of descriptive and inferential statistics was employed. The
descriptive statistics were employed include Frequency Table, Charts and Percentages while the inferential
statistics employed was Multiple Regression Technique
Model Specification
BRAWA = β0+ β1AIPA + β2AIPM + β3AIIE + ε
Where:
BRAWA = Brand Awareness
AIPA = AI-Personalized Advertising
AIPM = AI-Optimized Promotional Messaging AIIE
= AI-Assisted Interactive Engagement
β0 = Constant
β1, β2, β3 are the coefficients of the explanatory variables
RESULTS AND DISCUSSION
Analysis of Administered Questionnaire
Table 3: Distribution of Questionnaire by States
S/N
States
Nos Distributed
Nos Returned
Return Rate
1
Lagos
108
103
25.9
2
Ogun
53
52
13.1
3
Oyo
68
66
16.6
4
Osun
54
51
12.8
5
Ondo
65
62
15.6
6
Ekiti
50
50
12.6
TOTAL
398
384
96.6
Source: Researcher’s Data Output (2026).
A total of 398 questionnaires were distributed across six states, resulting in 384 completed and returned
responses. This impressive return rate of 96.48% signifies a strong level of engagement from the retiree
population, which is critical for ensuring the validity and reliability of the study's findings. Such a high response
rate suggests that retirees are not only interested in the topic but also actively engaged in discussions surrounding
herbal drug usage influenced by AI-Driven marketing communications.
Examining the individual states reveals noteworthy variations in return rates that may reflect the effectiveness
of AI-Driven MC in influencing purchasing behaviour. In Lagos, 108 questionnaires were distributed, with 103
Page 866
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue II, February 2026
returned, resulting in a return rate of 95.37%. This significant level of participation from retirees in this urban
center indicates that integrated marketing communications may resonate well with this demographic,
highlighting their awareness and interest in herbal drugs. Similarly, Ogun state exhibited a remarkable return
rate of 98.11%, with 53 distributed and 52 returned. Such a high response rate suggests that the IMC strategies
implemented in Ogun may be particularly effective in capturing the attention of retirees, potentially encouraging
them to explore herbal drugs.
The state of Oyo also demonstrated strong engagement, with a return rate of 97.06% from 68 distributed
questionnaires, of which 66 were returned. This high participation rate implies that the integrated marketing
communications have likely made a positive impact on the retirees' perceptions and behaviours regarding herbal
drug purchases.
In Osun, the return rate was slightly lower at 94.44%, with 54 distributed and 51 returned. While this rate is still
commendable, it suggests that there may be specific factors at play influencing the level of interest among
retirees in this state, warranting further investigation. Ondo showed a similar trend with a return rate of 95.38%
from 65 distributed and 62 returned, reinforcing the overall pattern of high engagement across the states.
Meanwhile, Ekiti achieved a notable 100% return rate, with all 50 distributed questionnaires returned. This
complete participation indicates a particularly strong inclination towards herbal drugs among retirees in Ekiti,
which may be indicative of effective marketing strategies that resonate deeply with this demographic.
Test of Research Hypotheses
Table 4: Linear Regression AI-Driven Marketing Communication on Herbal Drugs Brand Awareness
among Retirees in Southwest, Nigeria.
Independent Variable
Unstandardised Coefficient
Standardised
Coefficients
T
P-Value
B
Std. Error
Beta
(Constant)
2.291
.641
3.575
.000
AIPA
.361
.043
.145
3.183
.002
AIPM
.471
.044
.154
3.363
.001
AIIE
.601
.062
.479
9.765
.000
R = 0.721 Adj. R2 = 0.514
R2 = 0.520 F = 134.662 (0.000)
Std Err. = 1.50679
Dependent Variable: Brand Awareness
Source: Data Output, 2026
R
R²
Adjusted R²
F
Sig.
.721
.520
.514
134.662
.000
communications on herbal drugs brand awareness among retirees in Southwest Nigeria with coefficient value
(R) = 0.721 showed a positive and strong linear relationship among the dependent and independent variables of
the model.
Page 867
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue II, February 2026
The value of the coefficient of determination R2 is 0.514 signifying that the regression model explained 52%
variance in herbal drugs brand awareness among retirees in Southwest, Nigeria while the remaining changes
were accounted for by other extraneous dimensions beyond the scope of the model.
The regression coefficients of the independent variables show that one unit change in AI-Personalized advertising
will lead to 0.361 change in brand awareness; a unit change in AI-Optimized promotional messaging will lead
to 0.471 change in brand awareness while a unit change in AI-Assisted interactive engagement will lead to 0.601
change in herbal drugs brand awareness among retirees in Southwest Nigeria. All coefficients revealed positive
relationship with herbal drugs brand awareness among retirees in Southwest, Nigeria
Testing the hypotheses with consideration of the P-values of the independent variables, AI-Personalized
advertising has a P-value of 0.002 which is less than 0.05 level of significance, the null hypothesis that
AIPersonalized advertising does not significantly affect brand awareness is therefore rejected.
The P-value of AI-Optimized promotional messaging is 0.001 which is less than 0.05 level of significant. As a
result of this, the null hypothesis that AI-Optimized promotional messaging does not significantly affect brand
awareness is rejected
Finally, the P-value of AI-Assisted interactive engagement is 0.000 which is less than 0.05 level of significance.
Thus, the null hypothesis that AI-Assisted interactive engagement does not significantly affect brand awareness
id rejected.
RECOMMENDATIONS
i. Herbal drug companies should adopt AI-enabled personalization systems to tailor messages to retirees’
health concerns.
ii. Firms should implement AI-powered interactive platforms such as chatbots to provide structured product
explanations and build trust.
iii. Promotional campaigns should utilize predictive analytics to determine optimal timing and frequency of
communication.
iv. Companies should ensure simplicity and clarity in AI-generated content to accommodate varying digital
literacy levels among retirees.
CONCLUSION
The study concludes that AI-driven marketing communication significantly improves retirees’ brand awareness
of herbal drugs in Southwest Nigeria.
REFERENCES
1. Arora, N., Dreze, X., Ghose, A., Hess, J. D., Iyengar, R., Jing, B., … Shankar, V. (2021). Putting one-
to-one marketing to work. Journal of Marketing, 85(1), 109–127.
2. Bleier, A., & Eisenbeiss, M. (2015). The importance of trust in personalized advertising. Journal of
Advertising Research, 55(1), 24–38.
3. Bleier, A., Harmeling, C. M., & Palmatier, R. W. (2020). Creating effective online customer
experiences. Journal of Marketing, 84(2), 98–119.
4. Davis, F. D. (1989). Perceived usefulness and user acceptance of IT. MIS Quarterly, 13(3), 319–340.
5. Heath, R. (2016). The hidden power of advertising. What really affects the choice of the brand? GWP
6. Huang, M. H., & Rust, R. T. (2021). A strategic framework for artificial intelligence in marketing.
Journal of the Academy of Marketing Science, 49(1), 30–50.
7. Jachnis, A. (2022). Consumer psychology. Psychological and sociological determinants of consumer
behaviour. Oficyna Wydawnicza Branta.
Page 868
www.rsisinternational.org
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue II, February 2026
8. Koren, Y., Bell, R., & Volinsky, C. (2009). Matrix factorization techniques for recommender systems.
Computer, 42(8), 30-37.
9. Kotler, P. (2014). Marketing management: Analysis, planning, implementation and control (9th ed.).
Prentice-Hall.
10. Kumar, V., Ramachandran, D., & Kumar, B. (2023). Influence of AI-driven personalization on
consumer engagement. Journal of Business Research, 156, 113–124.
11. Li, J., Monroe, W., & Jurafsky, D. (2020). Deep reinforcement learning for neural dialogue generation.
Transactions of the Association for Computational Linguistics, 8, 663-676.
12. Mariani, M., Borghi, M., & Gretzel, U. (2023). AI chatbots in healthcare marketing. Technological
Forecasting and Social Change, 188, 122–134.
13. Martin, K. D., & Murphy, P. E. (2017). The role of data privacy in personalization. Journal of Business
Ethics, 152(1), 1–14.
14. Mehrabi, N., Morstatter, F., Saxena, N. (2021). A survey on bias and fairness in machine learning
systems. ACM Computing Surveys, 54(6), 1-35.
15. Nguyen, T., Li, C., & Zhao, K. (2021). Predictive analytics for adaptive digital engagement. Journal of
Marketing Analytics, 9(1), 45-60.
16. Petty, R. E., & Cacioppo, J. T. (1986). Communication and persuasion. Springer.
17. Radford, A., Wu, J., & Child, R.. (2019). Language models are unsupervised multitask learners.
OpenAI Technical Report.
18. Reiter, E., & Dale, R. (2000). Building natural language generation systems. Cambridge University
Press.
19. Wedel, M., & Kannan, P. K. (2016). Marketing analytics for data-rich environments. Journal of
Marketing, 80(6), 97–121.
20. Zhao, K., Kumar, A., & Yang, Q. (2019). Deep learning for user modeling and recommendation.
Information Retrieval Journal, 22(1-2), 1-28.