Automating Chalkboard support processes using a chatbot

Date
2019-04
Authors
Brown-Pobee, Joseph-Peter Yoofi
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This project seeks to apply natural language processing techniques to automate the support processes of Chalkboard Education, a startup in Ghana. The company has many users but currently has only two operations personnel responsible for responding to customer queries. Their support processes can be made easier with the use of a chatbot. The project references work from early chatbots like ELIZA and Cleverbot as well as more recent ones like MOOCBuddy and SuperAgent. The chatbot is built on RASA Natural Language Interpreter and uses third party APIs like Twilio and Database connections to mirror human support processes. Upon implementation, the chatbot is suitably able to perform the repetitive human tasks the operations personnel were carrying out, successfully and in shorter time. In the event that the chatbot cannot handle a query, the problem is forwarded to the aforementioned operations personnel. Upon evaluation, it was discovered that the chatbot has could improve its accuracy and effectiveness with techniques such as adding more training data and using different language models for embeddings. In the future, the chatbot can be implemented via a USSD application to enable Chalkboard capitalize on the prevalence of USSD application in Africa.
Description
Applied project submitted to the Department of Computer Science, Ashesi University, in partial fulfillment of Bachelor of Science degree in Management Information Systems, April 2019
Keywords
start-up, Chalkboard Education, chatbot, customer service support, dialog systems
Citation