Bridging the Gap in AI-Driven Workflows: The Case for Domain-Specific Generative Bots

Authors:
Publisher website
Google Scholar
Abstract
The widespread adoption of generative AI tools, such as ChatGPT, has resulted in its extensive use in a broad range of situations. However, language models often generate inaccurate or misleading responses, negatively impacting its use. Developing domain-specific bots for specific work situations could enhance accuracy and robustness, enabling more effective use of Generative AI in a work context. To help explore this possibility, we developed a data science process-expert generative AI assistant (bot) and evaluated its efficacy. We observed that the bot significantly improved efficiency, guided the exploration of new concepts within data science project management, and fostered creativity. Moreover, the constant availability of the bot allowed access to expertise whenever needed. Furthermore, responses indicated people viewed the bot as a collaborative tool that enabled communication and comprehension of complex questions. In addition, a Likert-scale analysis showed that the bot has the potential to impact the data science field positively. In summary, this research underscores the value of domain-specific bots and the potential impact on data science project management, as well as in other domains.

This publication has 25 references indexed in Scilit: