AI-assisterad administration i byggsektorn: En studie av Microsoft 365 Copilot

Sammanfattning

This master's thesis has explored the opportunities and challenges associated with using Microsoft 365 Copilot to automate and streamline administrative processes within the Swedish construction industry. In a sector long characterized by low productivity and limited digitalization, the focus has been on how an AI-powered virtual assistant can contribute to more structured and efficient workflows. The study aimed to answer three central questions: which administrative areas Copilot can support, what effects the tool may have on workflows and what challenges may arise during implementation and use. To address these questions, a combination of quantitative and qualitative methods was applied. A digital survey was distributed to professionals across various types of construction companies to provide a broad overview of how Copilot is perceived and used. In addition, five semi-structured interviews were conducted with individuals who had varying levels of experience with AI and Copilot, including daily users. Through methodological triangulation, both general trends and individual perspectives were analyzed. The findings show that Microsoft 365 Copilot is perceived as particularly useful in document management, reporting and communication. Respondents highlighted increased productivity, improved quality in written work, easier access to information and clear time savings as key benefits. At the same time, challenges such as legal uncertainty, data security issues, poor data quality, organizational resistance and limited training were identified. The study concludes that Microsoft 365 Copilot holds significant potential to improve administrative workflows in the construction sector, especially in information-intensive environments. However, realizing this potential requires active efforts in organizational support, technical readiness, and continuous skills development. As Copilot is still in the early stages of adoption within the construction sector, further research with larger and more diverse samples is needed. Follow up studies could provide insight into how the tool’s use and impact evolve over time. The generalizability of the results should therefore be interpreted with caution.