Lecture note on literature review for my MBA class: Aug 2026

Lecture note on literature review for my MBA class: Aug 2026

Related readings: 

1. Chapter 6 discussion on making recommendations and some related ideas related to literature review

An exploratory exercise with generative AI to produce a preliminary MBA dissertation proposal prototype: on AI project evaluation for customer service innovation in the HK insurance sector

An exploratory exercise with generative AI to produce a preliminary MBA dissertation proposal prototype: on AI project evaluation for customer service innovation in the HK insurance sector

 

 

I am a part-time MBA student, currently learning the subject of research methods. I am interested in the crude research theme of "examining the business value of AI projects in the domain of customer service in the Hong Kong Insurance sector". I am also interested in considering the following 4 key words to refine my research theme: benefit realization, pragmatism, exploratory research, and service innovation. Please suggest 4 concrete research themes for me to review based on this crude research theme and the 4 key words.

 

Here are 4 concrete research themes you can review, each built from your crude topic and the four keywords: benefit realization, pragmatism, exploratory research, and service innovation.

1. Benefit realization from AI customer service projects

Theme: How AI-enabled customer service projects create business value in Hong Kong insurance firms through benefit realization mechanisms.

This theme is strong if you want to focus on how firms measure and capture value after implementing AI, such as cost reduction, faster response time, improved customer satisfaction, or higher retention. It fits well with benefit realization and gives you a clear business-value lens.

2. Pragmatic evaluation of AI in insurance service delivery

Theme: A pragmatic assessment of AI adoption in customer service operations in the Hong Kong insurance sector.

This theme suits a pragmatism-based research philosophy, where you care less about abstract theory and more about what actually works in practice. You could examine how insurers decide whether AI tools are effective, useful, scalable, and worth continuing.

3. Exploratory study of AI value creation in insurance service

Theme: An exploratory study of how AI projects reshape customer service value creation in Hong Kong insurance companies.

This is a good option if the topic area is still underdeveloped and you want to discover patterns, perceptions, and emerging practices rather than test a fixed hypothesis. It aligns naturally with exploratory research and allows interviews or case studies.

4. AI-driven service innovation in Hong Kong insurance

Theme: The role of AI in service innovation and business value enhancement in the customer service function of Hong Kong insurance firms.

This theme places more emphasis on service innovation, so you can study how AI changes the way service is designed, delivered, and experienced. It is useful if you want to link technology adoption to innovation outcomes, not just efficiency outcomes.

Best fit by keyword

Keyword	Best-matching theme
Benefit realization	Theme 1
Pragmatism	Theme 2
Exploratory research	Theme 3
Service innovation	Theme 4

 

 

I am interested in concrete research theme 3: "Exploratory study of AI value creation in insurance service". Could you provide some relevant information on (1) the research issues involved, (2) some public media reported issues and (3) three relevant academic theories to use (with citations and references in Harvard reference format.

 

Below is a focused guide for your topic: “Exploratory study of AI value creation in insurance service” in the Hong Kong insurance context. I have grouped it into research issues, public-media issues, and three useful theories, with Harvard-style references at the end.

1. Research issues

For an exploratory dissertation, your main task is to identify how AI creates value in insurance customer service, rather than to test a fully mature model. A good starting point is to examine what “value” means in practice: faster service, better customer experience, lower service cost, improved staff productivity, better consistency, or new service offerings.emerald+1

Possible research issues include:

·        How AI value is defined and recognised by insurers, customer-service staff, and customers.

·        Which AI use cases in customer service actually create value, such as chatbots, agent-assist tools, claims triage, or automated information retrieval.info.gov+1

·        What conditions enable value creation, such as data quality, legacy-system readiness, management support, skills, and regulatory clarity.info.gov+1

·        What risks or trade-offs exist, including hallucinations, privacy concerns, discrimination, lack of transparency, and weak customer trust.hkust+1

·        How value is realised over time, because pilots do not always become operational benefits; some projects remain isolated experiments without redesigning work processes.org+1

A practical exploratory research question could be:
How do Hong Kong insurance firms create and realise business value from AI in customer service, and what factors shape that process?

2. Public media issues

Public reporting suggests that Hong Kong insurers are moving into AI, but adoption is still uneven and often exploratory. The Insurance Authority-related material indicates that many insurers are still exploring AI or have not fully implemented it, while customer-facing applications are expected to grow.info.gov+1

Key public-media issues you can cite:

·        Low maturity and uneven adoption. A large share of insurers are still exploring AI, with only a small minority fully implementing it.org+1

·        Legacy systems and integration problems. Older core systems make it difficult to scale AI beyond pilot projects.info.gov+1

·        Data privacy and compliance concerns. These are repeatedly highlighted as major barriers, especially in regulated financial services.soa+2

·        Model reliability and hallucination risk. Generative AI can generate incorrect or misleading responses, which is particularly serious in insurance customer service.hkust+1

·        Fairness, transparency, and trust. Media and industry commentary warn that AI can amplify exclusion or opaque decision-making if not carefully governed.hkust

·        Skills and organisational readiness. Reports mention shortages in AI talent, budget constraints, and the need for upskilling and process redesign.info.gov+1

For your dissertation, these issues can become the basis for exploratory interview prompts or case-study categories.

3. Three relevant theories

Here are three theories that fit your theme well.

a) Service-Dominant Logic

This is the best theory if your interest is value creation in service. Service-Dominant Logic argues that value is co-created through interaction between the firm and the customer, rather than simply produced by the firm and delivered passively. In AI-enabled customer service, this helps you study how chatbots, digital assistants, and agent-assist systems support customer value creation.assets.cambridge+2

Why it fits:
It directly supports the idea of AI as a resource in service exchange and helps you analyse value co-creation in customer service.

b) Technology–Organization–Environment (TOE) framework

TOE is useful if you want to explore why some insurers adopt AI more effectively than others. It explains adoption through three contexts: the technology itself, the organisation, and the external environment. In your topic, that could include system compatibility, leadership support, internal capabilities, regulation, and competitive pressure.diva-portal+2

Why it fits:
It gives you a strong structure for exploring organisational readiness and implementation barriers.

c) Practice Theory

Practice Theory is useful if you want to study how AI changes the actual work of customer service. Instead of focusing only on adoption decisions, it looks at everyday practices, routines, and how technology becomes embedded in work. This is especially relevant if you want to understand how employees use AI tools, how workflows change, and how value is produced in real service situations.socialecology.uci

Why it fits:
It helps you examine the lived reality of AI in service operations, not just policy or strategy.

Suggested way to position them

A strong theoretical combination for your dissertation could be:

·        Service-Dominant Logic as the main value-creation lens.

·        TOE as the adoption/implementation lens.

·        Practice Theory as the work-process lens.

That combination gives you a good exploratory structure: why AI is adopted, how it is used, and how value is actually created.

Harvard-style references

Artificial Intelligence Association of Hong Kong (2025) Annual Conference 2025. Available at: https://www.ia.org.hk/en/infocenter/files/AIHK_Annual_Conference_2025_EN.pdf (Accessed: 3 August 2026).

Orlikowski, W.J. (2007) ‘Sociomaterial practices: exploring technology at work’, Organization Studies, 28(9), pp. 1435–1448.

Tornatzky, L.G. and Fleischer, M. (1990) The Processes of Technological Innovation. Lexington, MA: Lexington Books.

Vargo, S.L. and Lusch, R.F. (2004) ‘Evolving to a new dominant logic for marketing’, Journal of Marketing, 68(1), pp. 1–17.

Vargo, S.L. and Lusch, R.F. (2008) ‘Service-dominant logic: continuing the evolution’, Journal of the Academy of Marketing Science, 36(1), pp. 1–10.

Vargo, S.L. and Lusch, R.F. (2016) ‘Institutions and axioms: an extension and update of service-dominant logic’, Journal of the Academy of Marketing Science, 44(1), pp. 5–23.

 

 

I would like to offer the following additional information and would like you to produce a concrete dissertation proposal for a 4-month part-time MBA project to me: I m currently the IT manager of the Hong Kong office of an Insurance company; I will use two primary research methods: (1) focus group with 3 of my office colleagues and 1 university lecturer on IT management and (2) opinion survey by questionnaire with my friends (about 60 respondents). I will also conduct secondary data analysis on my research theme. The overall research approach would be mixed methods research. Based on these information, please come up with a preliminary dissertation proposal with a project title, an overall research theme, 2 relevant research issues and 2 relevant issues reported in the public domain, the researcher's positionality, an overall research methodology, 4 research objectives, 3 research questions, 3 academic theories to use, and a preliminary project time line.

Below is a preliminary dissertation proposal tailored to your situation as an IT manager in a Hong Kong insurance office, using a mixed-methods design with a focus group, questionnaire survey, and secondary data analysis.

Proposed title

Exploring AI-Enabled Value Creation in Customer Service in the Hong Kong Insurance Sector: A Mixed-Methods Study of Benefits, Barriers, and Service Innovation

Overall research theme

The dissertation will explore how AI creates value in insurance customer service in Hong Kong, with emphasis on both business benefits and implementation challenges. The study will look at perceived value creation, service innovation, and the conditions under which AI produces practical business value. Hong Kong industry sources indicate that adoption is still uneven, with many insurers in exploratory or pilot stages, while public guidance highlights privacy, fairness, accountability, and integration risks.^{[1][2][3][4][5]}

1) Research issues

Issue 1: How AI creates business value in customer service

This issue examines what kinds of value AI delivers in insurance customer service, such as faster response, better service consistency, improved accessibility, and higher customer satisfaction. It also considers whether value is realised at the operational level, the customer experience level, or the strategic level.

Issue 2: Why AI value creation is difficult to scale

This issue explores the barriers that prevent AI from moving beyond pilots, including data quality, legacy system integration, staff capability, governance, and regulatory compliance. Hong Kong industry commentary repeatedly notes these constraints, especially for customer-facing use cases such as chatbots and claims support.^[3][6][7][1]

2) Public domain issues

Public issue 1: Privacy, security, and governance concerns

Public guidance in Hong Kong highlights risks such as personal data leakage, misuse of sensitive information, and the need for stronger governance when using generative AI. The Insurance Authority and related public materials also stress fairness, accountability, transparency, and policyholder protection.^[2][4][5][3]

Public issue 2: Uneven adoption and implementation maturity

Public reporting suggests that a significant share of insurers are still in exploratory or pilot phases, while only a smaller group has formal AI strategies or active implementation. This means AI value creation in Hong Kong insurance is still emerging rather than fully mature.^[8][1]

3) Researcher positionality

You are not a detached outsider: you are an insider-researcher because you work as the IT manager of a Hong Kong insurance office. This gives you practical access, contextual understanding, and the ability to interpret organisational realities, but it also creates risks of bias, role conflict, and social desirability in responses.

Your positionality statement can say that you will manage this by:

·       Separating your managerial role from your research role.

·       Using an external university lecturer in the focus group to broaden perspectives.

·       Triangulating focus group, survey, and secondary data findings.

·       Being explicit about your assumptions and possible organisational bias in the write-up.

4) Overall methodology

Research design

A mixed methods exploratory design is appropriate. The qualitative stage can help identify themes and refine the survey, while the quantitative stage can test whether the themes are broadly reflected in respondent opinions. Secondary data analysis will provide contextual support from public reports, regulatory guidance, and industry publications. Mixed methods is especially suitable here because your topic combines organisational meaning, practical implementation, and perceived value.^[4][1][3]

Primary methods

1.    Focus group

o   Participants: 3 office colleagues and 1 university lecturer.

o   Purpose: explore perceptions of AI value creation, implementation barriers, and service innovation.

o   Output: themes for the survey and conceptual refinement.

2.    Questionnaire survey

o   Respondents: about 60 friends.

o   Purpose: gather broader opinions on perceived value, trust, service quality, usefulness, and concerns about AI in insurance service.

o   Note: because the sample is convenience-based, you should present this as exploratory and not statistically representative.

Secondary data analysis

Use public reports, regulatory statements, media articles, and academic literature to contextualise your findings. This is important because your topic is closely connected to current Hong Kong industry developments and regulatory concerns.^{[5][1][2][4][8]}

5) Research objectives

1.    To explore how AI is perceived to create value in customer service within the Hong Kong insurance sector.

2.    To identify the main barriers and enablers affecting AI value creation in insurance service.

3.    To examine how AI may support service innovation and service improvement in insurance customer service.

4.    To develop a practical conceptual understanding of AI-enabled value creation suitable for a Hong Kong insurance context.

6) Research questions

1.    How do stakeholders perceive AI as creating value in insurance customer service in Hong Kong?

2.    What barriers and enablers influence the realisation of AI value in insurance service?

3.    How does AI contribute to service innovation and customer service improvement in the Hong Kong insurance sector?

7) Academic theories to use

1. Service-Dominant Logic

Use this as the main theory for value creation. It helps explain value as co-created through interactions among customers, employees, and technology rather than simply delivered by the firm.^[9][10][11]

2. Technology–Organization–Environment (TOE) framework

Use this to explain organisational adoption and implementation conditions. It is useful for identifying technological, organisational, and environmental factors affecting AI use in insurance.^[12][13][14]

3. Practice Theory

Use this to analyse how AI changes day-to-day customer service work, routines, and service delivery practices. It is a good fit if you want to understand how value is embedded in actual service processes.^[15]

8) Preliminary project timeline

Given your 4-month window, a realistic timeline would be:

Month	Main activities
Month 1	Finalise topic, refine literature review, develop proposal, prepare interview and survey instruments, seek supervisor feedback.
Month 2	Conduct focus group, transcribe and analyse findings, refine questionnaire based on qualitative results.
Month 3	Administer questionnaire, collect responses, conduct preliminary quantitative analysis, continue secondary data review.
Month 4	Integrate findings, write discussion and conclusion, revise draft, final proofreading and submission preparation.

 

9) A practical note

Because your questionnaire sample is a convenience sample of friends, your dissertation should clearly frame itself as exploratory and interpretive, not as a statistically generalisable survey of the whole industry. That will make the design more academically defensible and consistent with your mixed-methods approach.

A collection of blog notes on using chatgpt for research purpose.

Lecture note on “making recommendations” in an MBA dissertation report

Lecture note on “making recommendations” in an MBA dissertation report

 

In the context of doing my par-time four-month MBA dissertation project, I have a query about formulating recommendations in Chapter 6 (conclusions and recommendations) dissertation report, which is: how to formulate recommendations that are clearly supported by the dissertation report findings; that the recommendations are implementable; and that the recommendations are endorsed by the readers (e.g. managers in particular sector or managers for a particular client organization) as practical, useful, holistic and strategic. I feel that recommendations could not solely rely on the dissertation researching findings (e.g. Chapter 5 [discussions of findings]).

 

Yes — your instinct is right. In a strong MBA Chapter 6, recommendations should be grounded in findings, but not mechanically copied from Chapter 5; they should translate findings into managerial action, using judgment, theory, feasibility, and context.warwick.ac+1

Core principle

A useful way to think about it is:

• Chapter 5 answers: “What did the study find, and what does it mean?”
• Chapter 6 answers: “Given those meanings, what should managers do next, and why is that the best practical response?”

So recommendations are not just a repeat of findings; they are an evidence-informed decision proposal built from findings plus interpretation, literature, constraints, and managerial priorities.gradcoach+1

How to formulate them

Use a chain like this for each recommendation:

1.    Finding: state the key result briefly.

2.    Implication: explain the business problem/opportunity revealed.

3.    Recommendation: specify the action, owner, and scope.

4.    Justification: show why this is the best response, using evidence and relevant literature.

5.    Feasibility: mention resources, timeline, risks, or dependencies.

6.    Expected outcome: state what improvement should result.

This structure keeps the recommendation clearly linked to the report while making it managerial and actionable.premierdissertations+2

What makes it credible

To make recommendations implementable and more likely to be accepted by managers, they should be:

• Specific: say exactly what should be done.
• Prioritized: focus on the highest-impact actions first.
• Feasible: fit the organization’s resources, capability, and time.
• Strategic: connect to business goals, not just operational fixes.
• Context-sensitive: reflect the client organization or sector conditions.
• Measurable: include success indicators or outcomes where possible.

If a recommendation is theoretically attractive but unrealistic, it weakens Chapter 6. Better to offer a narrower, realistic action than a grand but unworkable one.gallowayresearch+1

How to go beyond Chapter 5

You are also correct that recommendations should not rely solely on the findings section. A strong Chapter 6 usually also draws on:

• Literature: to support why the recommendation makes sense beyond your sample.
• Theory: to explain the logic of the proposed action.
• Managerial judgment: to adapt the recommendation to real-world constraints.
• Assumptions and limitations: to signal where the recommendation is context-bound.
• Implementation considerations: resources, sequencing, risks, and accountability.

That combination makes the chapter feel practical, holistic, and strategic rather than merely descriptive.warwick.ac+2

Simple writing formula

You can use a sentence pattern like:

“Because the study found X, which suggests Y, the organization should do Z by [who/when], in order to achieve [outcome], subject to [constraint].”

Example:

“Because the study found that service inconsistency reduced customer trust, the firm should introduce a standardized frontline service checklist and monthly quality review, led by the operations manager, to improve service reliability within three months.”

That style is concise, evidence-linked, and implementable.

** a collection of lecture notes on research methods