How Virtual Advisors Create Practical Employee Efficiency Boosts in the Chemical Engineering Workplace

摘要

In an effort to explore the ability of text analytics software to augment human performance within the chemical engineering domain,a virtual advisor designed for the domain was used by chemical engineering practitioners in a controlled experiment.The experiment focused on the common industry problem of inefficient and incomplete extraction of relevant knowledge and data from within the organization's digital unstructured textual (DUT) files in order to resolve issues.The experiment focused on hydrocarbon refinery content within the broader chemical engineering domain.The virtual advisor used in the experiment employs machine learning and natural language processing to interpret the intent of human queries and instantly return knowledge and data from across multiple DUT files that is relevant to the queries.The advisor has shown capacity in commercial applications within the industry,but the key details of those results remain proprietary to the corporate clients.While this study cannot name the individual participants in the experiment,the corporations they represent,or their proprietary issues explored using the virtual advisor during the experiment,the experimental results comparing the performance of the individuals with and without the virtual advisor are shared in detail.The results suggest that the virtual advisor increases the speed and accuracy of issue resolution in the chemical engineering domain over typical incumbent tools in a base case control scenario by significant margins.This in turn suggests that chemical engineering is an appropriate domain for the productive application of virtual advisors based on"artificial intelligence"methods including machine learning and natural language processing.The implications of the results are explored,and additional larger and more open studies in this area are recommended.