Methodology
GSP assesses whether sustainable software and AI practices are systematically adopted within an organisation. It looks at practices, not marketing claims.
Evidence-based
Each practice states what an organisation shall do and gives examples of evidence that can show it is in place. Depending on the certification status, the evidence includes:
- documented processes
- technical artefacts
- measurement results
Self-assessed certification relies on what the organisation declares. Verified certification adds an independent review of the evidence by GreenSeal.dev. See Certification.
Research foundation
The practices are derived from peer-reviewed research in Sustainable Software Engineering and Green AI, not from proprietary frameworks or internal opinion. The work behind them covers:
- Sustainable Software Engineering
- Green AI
- Performance and energy profiling
- Organisational knowledge transfer
The methodology is informed by work published at venues including ICSE, CAIN, WIREs and ACM SIGSOFT. Read about the research behind GreenSeal.dev (external site).
The Level 2 practices have been reviewed against the six lifecycle phases of ISO/IEC TS 20125-1:2026 (external site). The practices that apply at Level 2 only are the ones added in that review.
Principles
- Engineering first, not offsets. GSP is about how software is designed, built and run, not about compensating for its emissions.
- Evidence-driven and auditable. Every practice says what evidence can show it is in place.
- Open tooling. Energy measurements use open-source tools such as EnergiBridge (external site), not a proprietary black box.
- Software and AI. The practices apply to software-intensive systems, cloud-native applications and AI/ML workloads.
- Built for real teams. GSP is designed for engineering organisations of any size, from scale-ups to enterprises.