AI and ML in Real Estate Underwriting: Transforming Financial Decision-Making and Operational Efficiency
Name
jaklis-cjaklis-msms-mgt-2025-thesis.pdf
Description
Thesis PDF
Size
3.32 MB
Format
Adobe PDF
Checksum (MD5)
2935130c04077fdf0a32365858513121
Author(s)
Jaklis, Cyril
Advisor(s)
Cohen, Jake
Date Issued
May 2025
Publisher
Massachusetts Institute of Technology
Abstract
Real estate is the world's largest untapped market, at $650 trillion (Statista, 2023), yet technological innovation, particularly in financial underwriting, is underrepresented. Excel spreadsheets, broker-driven data collection, and expensive public database subscriptions are still used by most institutional players and family offices. These outdated approaches result in inefficiencies and higher operational expenses. Firms are now waiting for more innovative tools to improve their workflows and predict their Net Operating Income (NOI). Development and maintenance costs are often underestimated due to optimistic estimates and unplanned material or labor cost price escalations. This paper examines how to increase the accuracy of underwriting by examining the full underwriting process, identifying operational inefficiencies, and analyzing how new technologies like Artificial Intelligence (AI) and Machine Learning (ML) are currently being utilized to better value properties and reduce error margins. The analysis covers the entire underwriting process, from data sourcing, collection, structuring, and analysis. It also reviews the platforms and software tools utilized to connect these phases, from initial appraisal to investment memo and investment committee (IC) decision-making. The objective is to understand practical constraints, recognize opportunities for optimization, and explore where investors can strategically position themselves to leverage these technologies while also providing a forward-looking outlook on the changing function of AI/ML in the sector over the next decade.
MIT Department
Sloan School of Management
Terms of Use
In Copyright - Educational Use Permitted
Copyright retained by author(s)
Persistent DSpace Link
