A Data-Driven Approach to Identify Potential AI Chatbots for Electrical Construction Contractors

Abstract

The electrical construction industry faces growing complexity in managing technical information and regulatory compliance. While AI chatbots have shown promise in various sectors, their application in electrical construction remains largely unexplored. This study uses a mixed-methods approach to identify and assess potential applications of custom-built AI chatbots for the industry. We analyzed 542 YouTube videos, yielding over 1,388 transcript pages, and conducted a pilot interview with a service electrician to uncover key field challenges. Leveraging GPT-4o and Claude 3.5 Sonnet, we identified eight thematic categories from the transcripts, with quality control and safety emerging as dominant concerns. The interview confirmed these findings and revealed additional challenges in code reference and communication. A prototype chatbot was developed to support electrical code inquiries and demonstrated improved jurisdiction-specific accuracy compared to general-purpose AI tools. This research contributes both theoretical understanding and a practical framework for implementing targeted AI solutions in specialized construction domains.

Ähnliche Arbeiten