Repository logo

TailorAlert: Large Language Model-Based Personalized Alert Generation System for Road Infrastructure Management with Digital Twins

Accepted version


Conference Object

Change log


Parn, Erika 
Kookalani, Soheila 
Brilakis, Ioannis 


This paper presents a proof-of-concept study on enhancing road maintenance information dissemination using a Large Language Model (LLM)-based system. Traditional digital twin systems often generate general alerts without considering the diverse information needs of different personnel from senior executives to operational staff, due to an inability to disseminate role and skill-level-specific information. In this paper, we propose TailorAlert, a system designed to convert general alerts into user-friendly natural language notifications tailored to each staff member's role and expertise level. The system can provide high-level strategic overviews for senior executives, focusing on decision-making and overall maintenance status, and detailed, actionable, data-centric alerts for operational staff, aimed at guiding specific issue resolution. The system can also differentiate between senior and junior personnel, offering tailored information to suit their roles and skill levels. Senior staff are provided with in-depth data, facilitating advanced decision-making or diagnostics, while junior personnel receive more straightforward, practical, task-specific instructions and data focused on immediate issue resolution. To validate the effectiveness of the system, tests were conducted on simulated maintenance scenarios. These tests demonstrate the system's effectiveness in simulated scenarios typical of real-world maintenance operations. The system shows the potential to improve operational response time, decision-making efficiency, and communication efficiency by providing relevant personnel with tailored data. This research demonstrates LLMs’ potential in personalized information dissemination, thus streamlining the maintenance workflow and improving the overall efficiency of road infrastructure management. The findings also suggest its broader applicability in other infrastructure management sectors and beyond where tailored information dissemination is crucial.



Journal Title

Conference Name

2024 ASCE International Conference on Computing in Civil Engineering

Journal ISSN

Volume Title


Publisher DOI

Publisher URL

Engineering and Physical Sciences Research Council (EP/S02302X/1)
Engineering and Physical Sciences Research Council (2728220)
EPSRC, Bentley Systems, and the University of Cambridge through the EPSRC Centre for Doctoral Training in Future Infrastructure and Built Environment: Resilience in a Changing World (EPSRC grant reference number EP/S023 02X/1)