Intelligent conversational bot for interactive marketing

• Main Author: Pradana, Aditya
• Format: Thesis
• Language: English; English
• Published: 2017
• Subjects: T Technology (General); TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://eprints.utem.edu.my/id/eprint/20754/1/Intelligent%20Conversational%20Bot%20For%20Interactive%20Marketing%20-%20Aditya%20Pradana%20-%2024%20Pages.pdf; http://eprints.utem.edu.my/id/eprint/20754/2/Intelligent%20conversational%20bot%20for%20interactive%20marketing.pdf; http://eprints.utem.edu.my/id/eprint/20754/; https://plh.utem.edu.my/cgi-bin/koha/opac-detail.pl?biblionumber=106009
• Institution: Universiti Teknikal Malaysia Melaka

Interactivity of corporate website is important as marketing products through it become popular. Moreover, the usage of Artificial Intelligent application is growing. In order to improve the interactivity and effectiveness of corporate website in providing information, a conversational bot is developed. As a part of Artificial Intelligent application, it can respond to users’ question and prolong the conversation with its intelligence. The conversational bot in this research is called as SamBot. It is integrated to Samsung IoT website as the corporate website. SamBot contains the knowledge of Samsung marketing domain which is useful to deliver appropriate answer to the users’ questions. The knowledge base includes Samsung promotion, Samsung product Frequently Asked Question, and general knowledge. In default, this bot will generate random answers if matching knowledge cannot be found. It usually happens if novel issue or information is asked. In this research author would like to overcome this problem by improving its knowledge maintenance capability either performed by botmaster or users. Some functions are added to support the knowledge maintenance process such as random answer enhancement, conversation log-based knowledge maintenance module, and user-centric learning capability. The latter one allows user to directly teach the bot a new knowledge. In addition, enhancement to reuse recently asked question as questions recommendation is developed in order to solve a problem where users need to be updated to current issues. With this enhancement, the bot can give recommendation of questions for users to ask. After the development, the system is deployed online and some users are invited to perform evaluation. From the evaluation, the results show that all of the enhancements comply the objectives of the research in which 62% of the questions posed are responded properly and 80% of users agree that SamBot has raised the interactivity and effectiveness of Samsung IoT website. Moreover, there is a positive change on how SamBot responds to questions which were not answered properly before knowledge maintenance. There is also a noticeable usage of pre-defined question recommendation on questions such as “What is Samsung?”, “Who is the founder of Samsung?”, and “What is Galaxy Gear VR?”. In simple words, SamBot is now more intelligent with all of those enhancements.