Artificial Intelligence Appliances in Customer Communications Management solutions
Heljaste, Juuso (2023)
2023
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-2023121135980
https://urn.fi/URN:NBN:fi:amk-2023121135980
Tiivistelmä
The main objective of this constructive research study was to find suitable use cases of Artificial Intelligence (AI) appliances in Customer Communication Management (CCM) business domain. The second main objective was to create a prototype for the most feasible identified use case while describing best practices guidelines on how to implement such modules utilizing open-source tools. CCM domain is still in its early phases of adoption of AI features. Although many AI branches seem prominent to CCM, this study was limited to Machine Learning (ML) and Natural Language Processing (NLP) AI branches.
The theoretical framework part of the study started with an overview of AI, its history, and de-scription of different AI branches. AI section was followed by ML overview and practical implementation guidelines for ML systems and then by similar structure with NLP. Theoretical frame-work was concluded with an overview of CCM domain and description of typical business cases identified by the commissioner of the study, ENIT AB.
The main data collection method was a brainstorming session held among ENIT subject matter experts in March 2023. Each topic identified and discussed during the brainstorming session was analysed using applicable content analysis methods and evaluated within the concept of ML and NLP. Feasibility as an AI use case for ENIT and CCM domain in general was also analysed. The outcome of the analysis was an evaluation table of each topic in terms of complexity of implementation and possible value to CCM. The use case identified as the low-hanging-fruit was selected as the foundation for constructing the prototype.
For the prototype, a Machine Learning classifier for important log entries was successfully developed. The implementation phases were described in detail providing a solid framework for ENIT and other CCM providers to create similar features in their product and service offering. The concept of the prototype classifier is applicable throughout CCM industry.
The planning of the study started in January 2023 and the study was concluded in November 2023. The main challenges encountered revolve around the scarcity of scientific research combining AI and CCM. However, many feasible AI use cases in CCM were able to be identified and many of them can be developed with moderate effort. Further research on other AI branches should follow this study to fully unlock AI’s potential in CCM outside of ML and NLP branches.
The theoretical framework part of the study started with an overview of AI, its history, and de-scription of different AI branches. AI section was followed by ML overview and practical implementation guidelines for ML systems and then by similar structure with NLP. Theoretical frame-work was concluded with an overview of CCM domain and description of typical business cases identified by the commissioner of the study, ENIT AB.
The main data collection method was a brainstorming session held among ENIT subject matter experts in March 2023. Each topic identified and discussed during the brainstorming session was analysed using applicable content analysis methods and evaluated within the concept of ML and NLP. Feasibility as an AI use case for ENIT and CCM domain in general was also analysed. The outcome of the analysis was an evaluation table of each topic in terms of complexity of implementation and possible value to CCM. The use case identified as the low-hanging-fruit was selected as the foundation for constructing the prototype.
For the prototype, a Machine Learning classifier for important log entries was successfully developed. The implementation phases were described in detail providing a solid framework for ENIT and other CCM providers to create similar features in their product and service offering. The concept of the prototype classifier is applicable throughout CCM industry.
The planning of the study started in January 2023 and the study was concluded in November 2023. The main challenges encountered revolve around the scarcity of scientific research combining AI and CCM. However, many feasible AI use cases in CCM were able to be identified and many of them can be developed with moderate effort. Further research on other AI branches should follow this study to fully unlock AI’s potential in CCM outside of ML and NLP branches.