Introduction
Digital operations have become a critical part of modern organisations, supporting the delivery of digital services across sectors such as healthcare, finance, and public services.
hese operations depend on software-intensive systems and digital infrastructure that consume significant amounts of energy and contribute to carbon emissions. As organisations seek to reduce their environmental impact, improving the sustainability of digital operations is becoming an increasingly important challenge.
Many of the environmental impacts associated with digital operations are determined by decisions made long before systems enter operation. Choices made during the design, deployment, and maintenance of software systems can influence resource consumption for many years and may be difficult and costly to reverse once embedded within runtime environments. Research in Green Software Engineering has highlighted the importance of considering sustainability much earlier in decision making (Lago 2019) and treating it as a key quality concern when developing new software systems (Lago 2015).
Organisations increasingly collect information about the performance, resource consumption, energy use, and carbon impacts (Microsoft 2023, SCI 2026) of their digital operations. While such information has the potential to inform future software and infrastructure decisions (Giussani 2025), it is often unclear how it should be used when evaluating alternatives and making design choices with long-term sustainability implications. The advent of Generative Artificial Intelligence (GenAI) provides opportunities to support such complex decision making. Recent research suggests considerable potential for using GenAI in software design and architecture, while also highlighting challenges around explainability and trust (Esposito 2025).
Research Aims
This project will investigate how AI can help organisations understand the long-term sustainability implications of decisions relating to software systems and digital infrastructure before those decisions become operationalised and difficult to change.
The research will explore how sustainability knowledge can be represented and combined with AI techniques to support the evaluation of alternatives, sustainability trade-off analysis, and decision making. Research directions may include sustainability knowledge representation, AI-assisted decision support, explainable and trustworthy AI, and techniques for assessing the long-term sustainability implications of digital operations. Ultimately, the project aims to help organisations make better-informed decisions that reduce resource consumption and carbon emissions across the lifetime of digital systems.
References:
Esposito, M., Li, X., Moreschini, S., Ahmad, N., Cerny, T., Vaidhyanathan, K., Lenarduzzi, V. & Taibi, D. (2025). Generative AI for software architecture. Applications, challenges, and future directions. Journal of Systems and Software, 112607.
Giussani, S., Caporuscio, M., & Perez-Palacin, D. (2025, September). A Reasoning Framework for Architecting Carbon-Aware Software-as-a-Service Applications. In Euromicro Conference on Software Engineering and Advanced Applications (pp. 231-241). Cham: Springer Nature Switzerland.
Lago, P., Koçak, S. A., Crnkovic, I., & Penzenstadler, B. (2015). Framing sustainability as a property of software quality. Communications of the ACM, 58(10), 70-78.
Lago, P. (2019, May). Architecture design decision maps for software sustainability. In 2019 IEEE/ACM 41st International Conference on Software Engineering: Software Engineering in Society (ICSE-SEIS) (pp. 61-64). IEEE.
Green Software Foundation (GSF) https://greensoftware.foundation/
Software Carbon Intensity (SCI) Specification. 2026. https://greensoftware.foundation/standards/sci/
Microsoft and UBS, Carbon-aware computing 2023 https://news.microsoft.com/de-ch/2023/01/10/carbon-aware-computing-whitepaper/