Designing AI Assistants for Novices: Bridging Knowledge Gaps in Onboarding
Files
Date
2025-01-07
Contributor
Advisor
Department
Instructor
Depositor
Speaker
Researcher
Consultant
Interviewer
Narrator
Transcriber
Annotator
Journal Title
Journal ISSN
Volume Title
Publisher
Volume
Number/Issue
Starting Page
3780
Ending Page
Alternative Title
Abstract
The rapid advancements of artificial intelligence (AI) have led to its widespread adoption in enhancing productivity across various domains, including both personal productivity and workplace efficiency. Despite these advancements, the effective onboarding of novices, particularly in complex environments with extensive documentation, remains a significant challenge. Therefore, this paper explores the design of AI assistants to support novices. Utilizing a design science research approach, we collaborate with a leading pharmaceutical manufacturer to develop and evaluate an AI assistant to support novices during their onboarding. Grounded in scaffolding theory, we identify two design requirements and propose three design principles: Metadata Filtering, Graduated Complexity, and Sequential Query Generation. The evaluation of these principles demonstrates the AI assistant's effectiveness in generating accurate and contextually relevant responses, facilitating the onboarding of novices. This study provides valuable insights into the design of AI assistants, contributing to the theoretical understanding of AI-driven scaffolding and practical applications in complex industrial settings.
Description
Keywords
Artificial Intelligence-based Assistants and Platforms, ai assistants, human-ai collaboration, large language model, onboarding
Citation
Extent
10
Format
Geographic Location
Time Period
Related To
Proceedings of the 58th Hawaii International Conference on System Sciences
Related To (URI)
Table of Contents
Rights
Attribution-NonCommercial-NoDerivatives 4.0 International
Rights Holder
Local Contexts
Email libraryada-l@lists.hawaii.edu if you need this content in ADA-compliant format.