Scripted or Smart: Navigating Chatbot Choices
(2025) SYSK16 20251Department of Informatics
- Abstract
- This thesis examines the adoption of LLM-powered chatbots in Nordic insurance companies, comparing them to traditional rule-based systems. While LLMs enable flexible, human-like interactions, their use in regulated industries remains cautious. Using the Technology-Organization-Environment (TOE) framework and five interviews with industry professionals, the study identifies legal accountability, predictability, and integration ease as key adoption drivers, factors more aligned with rule-based systems. Concerns over hallucinations, data privacy, and infrastructure complexity often outweigh LLM benefits. Yet, vendor pressure and shifting customer expectations are encouraging hybrid experimentation. Organizational readiness, leadership... (More)
- This thesis examines the adoption of LLM-powered chatbots in Nordic insurance companies, comparing them to traditional rule-based systems. While LLMs enable flexible, human-like interactions, their use in regulated industries remains cautious. Using the Technology-Organization-Environment (TOE) framework and five interviews with industry professionals, the study identifies legal accountability, predictability, and integration ease as key adoption drivers, factors more aligned with rule-based systems. Concerns over hallucinations, data privacy, and infrastructure complexity often outweigh LLM benefits. Yet, vendor pressure and shifting customer expectations are encouraging hybrid experimentation. Organizational readiness, leadership support, and capability development emerge as critical enablers. The findings suggest that resistance stems less from technical limitations than from efforts to balance innovation with regulation, risk, and trust. This study deepens understanding of AI adoption in customer-facing roles within highly regulated sectors. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9202253
- author
- Young, Felix LU ; Junco Hagberg, Alexander LU and Hylander, Elias LU
- supervisor
- organization
- alternative title
- A Qualitative Study on LLM and Rule-Based Chatbot Adoption in Insurance Customer Service
- course
- SYSK16 20251
- year
- 2025
- type
- M2 - Bachelor Degree
- subject
- keywords
- AI Ethics, Chatbot Adoption, Insurance Sector, Large Language Models, Rule-Based Systems, Technology-Organization-Environment Framework
- language
- English
- id
- 9202253
- date added to LUP
- 2025-06-18 13:11:36
- date last changed
- 2025-06-18 13:11:36
@misc{9202253, abstract = {{This thesis examines the adoption of LLM-powered chatbots in Nordic insurance companies, comparing them to traditional rule-based systems. While LLMs enable flexible, human-like interactions, their use in regulated industries remains cautious. Using the Technology-Organization-Environment (TOE) framework and five interviews with industry professionals, the study identifies legal accountability, predictability, and integration ease as key adoption drivers, factors more aligned with rule-based systems. Concerns over hallucinations, data privacy, and infrastructure complexity often outweigh LLM benefits. Yet, vendor pressure and shifting customer expectations are encouraging hybrid experimentation. Organizational readiness, leadership support, and capability development emerge as critical enablers. The findings suggest that resistance stems less from technical limitations than from efforts to balance innovation with regulation, risk, and trust. This study deepens understanding of AI adoption in customer-facing roles within highly regulated sectors.}}, author = {{Young, Felix and Junco Hagberg, Alexander and Hylander, Elias}}, language = {{eng}}, note = {{Student Paper}}, title = {{Scripted or Smart: Navigating Chatbot Choices}}, year = {{2025}}, }