The Advent of Speech Based NLP QA Systems: A Refined Usability Testing Model

摘要

This paper outlines a refined usability testing model, developed for industry specific comparative usability testing of two natural language processor (NLP) based question answering (QA) systems. The systems operate over differing modalities, one through text and the other speech. The revised model combines two existing usability testing frameworks, the System Usability Metric (SUM) and the System Usability Scale (SUS). It also integrates the context-specific determined target user - financial managers working within the financial services industry. The model's metric weightings are determined through key informant interviews. The presented model will be the working framework from which a series of comparative usability tests will be carried out within the target organisation.