AI-driven engagement in public services : enhancing customer journey with a chatbot
Barros, Diego (2025)
2025
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2025052113939
https://urn.fi/URN:NBN:fi:amk-2025052113939
Tiivistelmä
This thesis developed a chatbot framework to be implemented in the case organization, a public institution dedicated to teaching Portuguese Language in Helsinki. The study addressed a specific business challenge of the instructors spending excessive time to answer repetitive administrative questions, in addition to their pedagogical duties. The objective was to create an automated solution able to provide immediate answers to frequently asked questions, allowing teachers to prioritize their educational responsibilities.
The study employed an action research approach using the newly created Dynamic Implementation in the Groundwork — DIG Framework. Data was collected through analyses of student emails, interviews with instructors, and user surveys. The study revealed that student inquiries consistently fell into three categories: course duration, content, and subscription practicalities. Based on this assessment, the in-house development of a closed-domain, retrieval-based chatbot was identified as the most suitable solution, considering the case organization's particularities.
Following the DIG Framework's iterative cycles, RosaBOT was developed using the IBM Watsonx Assistant's Lite plan. The final version features four conversation flows addressing common inquiries and facilitating course subscriptions. The solution was implemented and demonstrated high satisfaction levels (4.56/5 average rating) and a 100% subscription completion rate. The solution helped reduce repetitive email inquiries by approximately 73% during the comparable period after implementation.
The implementation shows how public educational institutions can employ AI solutions to enhance service delivery without significant investment. By automating routine inquiries, RosaBOT allows instructors to focus on delivering quality education. The solution is currently deployed on the case organization's website. The ultimate purpose of this study, therefore, is to illustrate how technology can be used in educational settings to effectively support creative, human work.
The study employed an action research approach using the newly created Dynamic Implementation in the Groundwork — DIG Framework. Data was collected through analyses of student emails, interviews with instructors, and user surveys. The study revealed that student inquiries consistently fell into three categories: course duration, content, and subscription practicalities. Based on this assessment, the in-house development of a closed-domain, retrieval-based chatbot was identified as the most suitable solution, considering the case organization's particularities.
Following the DIG Framework's iterative cycles, RosaBOT was developed using the IBM Watsonx Assistant's Lite plan. The final version features four conversation flows addressing common inquiries and facilitating course subscriptions. The solution was implemented and demonstrated high satisfaction levels (4.56/5 average rating) and a 100% subscription completion rate. The solution helped reduce repetitive email inquiries by approximately 73% during the comparable period after implementation.
The implementation shows how public educational institutions can employ AI solutions to enhance service delivery without significant investment. By automating routine inquiries, RosaBOT allows instructors to focus on delivering quality education. The solution is currently deployed on the case organization's website. The ultimate purpose of this study, therefore, is to illustrate how technology can be used in educational settings to effectively support creative, human work.