������

Title of the thesis: "Data-driven transformation in construction management  - From artificial intelligence to network modeling”

Thesis defender: Roope Nyqvist
Opponent: Prof. Daniel Hall, Delft University of Technology, Holland
Custos: Prof. Antti Peltokorpi, Aalto University School of Engineering

The construction industry struggles with inefficiency, delays, and cost overruns, making a digital transformation necessary. This dissertation explores how data-driven innovations, such as AI-driven platform business models, Generative AI (GenAI), and new network modeling tools, could address these persistent challenges.

The study identifies key barriers and drivers for digitalization, exploring how companies can shift towards AI-driven platform business models. It studies the integration of GenAI in management tasks and develops a novel method, Uncertainty Network Modeling (UNM), to structure the critical human knowledge, also needed to guide the advanced digital tools.

A blind, expert-reviewed test pitting GenAI against human professionals yielded a striking finding: an AI model significantly outperformed experts in a risk management scenario. However, the AI's analysis, while comprehensive, was generic. Some human experts, in contrast, provided more practical, context-specific insights. This highlights the current limits of AI and the continued need for deep human expertise, which the UNM method helps make explicit.

The findings offer a practical roadmap for construction companies to shape their digital strategy, moving beyond software adoption to developing new AI-driven services and platform business models. It shows how managers can integrate GenAI as a powerful tool in their actions while using UNM to ensure human expertise guides critical decisions.

The main conclusion is that a successful data-driven transformation requires an integrated, multi-layer framework. This dissertation proposes a model that links high-level digital transformation strategy to the practical implementation of digital solutions, including AI-driven platforms and GenAI. The framework's foundation is the integration of these tools with human expertise. It accomplishes this by introducing a knowledge structuring layer that uses methods such as UNM to make tacit expert knowledge explicit and machine-readable. This structured data is then fed into digital tools, creating a human-machine synthesis. This integrated approach ensures that the analytical power of AI is guided by practical human insights, forming the foundation for the successful adoption of GenAI and unlocking new value also through AI-driven platform

Key words: Artificial intelligence, digital transformation, business model innovation, platform economics, construction management, risk management, network-based methods

Thesis available for public display 7 days prior to the defence at . 

Contact information: roope.nyqvist@aalto.fi