AI research at KTH to facilitate maintenance work for older buildings
A research collaboration between Einar Mattsson and KTH explores how AI can improve the understanding of renovation needs in various types of buildings. In the project ”AI-driven Predictive Maintenance Planning in Buildings”, large volumes of property data are being analysed to enhance maintenance planning, reduce costs for property owners, and lower the environmental impact of buildings.
In a typical older building, maintenance represents the largest cost — expenses that increase with age. Properties built in the 1960s and 70s are now approaching the point where large-scale renovation projects are needed. Meeting this challenge requires new approaches, updated processes, and modern IT support.
Mikael Dimadis, Project Manager for Research and Development at Einar Mattsson, leads the new project ”AI-driven Predictive Maintenance Planning in Buildings” (AIDA-B). The project aims to deepen the understanding of maintenance needs across various property types. He describes how AI increases the capacity to process and analyze data:
“For example, several years’ worth of service requests from a building can be quickly analysed, allowing us to identify which building components are driving the highest costs. This analysis can then serve as a foundation for decisions on how to optimise and anticipate future maintenance work,” he says.
In collaboration with researchers at KTH, a large amount of property data from Einar Mattsson properties will be analysed during the spring and, with the help of AI, transformed into decision-making support for property managers.
“I hope the project will lead to improved methods for maintenance planning. That’s how we envision property owners being able to reduce both their costs and their environmental impact,” says Mikael Dimadis.
Through the collaboration with KTH’s Dig-IT Lab and KTH Live-In Lab, Einar Mattsson also sees potential for anchoring the project among other major property companies and exchanging experiences related to development and renovation work.
The project started in February 2025, and an initial report is planned for September. The research is led by Kjartan Gudmundsson, Associate Professor, and Angela Fontan, Assistant Professor, both from KTH. They are supported by master’s students who will analyze the data and train the AI models.