Member Story
"Standards move too slowly for our regulatory timeline"
How GSF used an AI-assisted assembly process to bring 14 experts from 15 organisations to consensus in ten weeks — a process that previously took years — proving a new model for accelerating green software standards development as regulations tighten.
Organisations involved
10 weeks
From blank page to consensus design document — compared to multi-year traditional standards timelines
14
Expert participants from 15 organisations reaching consensus through the AI-assisted process
2
Standards developed using AI-facilitated consensus — SCI for AI and SCI for Web
The Problem
The bottleneck isn't technical complexity — it's coordinating dozens of stakeholders at the speed regulations require
Traditional standards development takes years. The SCI specification — widely celebrated as moving at unprecedented speed — still took over three years from inception to ISO certification. By the time specifications reach final form, regulations have evolved, technology has shifted, and the window for influence has narrowed. The bottleneck is not technical complexity. It is the challenge of coordinating dozens of stakeholders with competing priorities, different terminologies, and divergent views on scope, accuracy, and implementation.
The problem was becoming urgent. The EU Green Digital Coalition, the NY Corporate Climate Accountability Act, and evolving GHG Protocol guidance were all creating pressure for organisations to report on digital sustainability. A standard that arrived two years too late would be a standard nobody adopted.
The human-in-the-loop approach made objections visible and resolvable — a deliberate decision gate rather than the traditional model of circulating documents and hoping for feedback. What had been an opaque social process became an explicit technical one.
The Journey
From a blank page to consensus in ten weeks
ISO standard proves consensus is possible — but slow
April 2024
When the SCI Specification achieved ISO 21031:2024 status, it validated that a multi-stakeholder, consensus-driven process could produce a globally recognised standard. But even this celebrated achievement had taken over three years from inception. The question became: how do you keep the rigour of consensus while compressing the timeline? GSF had 17 active projects. If each required 3+ years of traditional consensus-building, the pipeline would be unmanageable.
Read about the SCI achieving ISO standard status →SCI for AI proves the concept
Early 2025
The AI industry was moving so fast that GSF needed to reach consensus on standardising AI emissions measurement quickly. Rather than follow the traditional multi-year process, the team experimented with a rudimentary version of AI-facilitated consensus — structured workshops where expert responses were synthesised and refined collaboratively. The SCI for AI specification was ratified in late 2025, and the experience revealed that AI-facilitated consensus was a vibrant area with significant potential to compress standards timelines further.
Read the SCI for AI workshop report →W3C and GSF announce formal collaboration
September 2025
The Green Web Foundation, active in both communities, brokered a strategic collaboration between GSF and W3C to advance sustainable web development. The collaboration's first major initiative was SCI for Web, and GSF chose it as the proving ground for a fully developed AI-facilitated consensus process — building on the lessons learned from the rudimentary approach used for SCI for AI.
Read about the GSF and W3C collaboration →The AI-assisted assembly
September–November 2025
Fourteen members with web-specific expertise were assembled: Alekh Gupta (Google), Alexander Dawson (ClimateAction.tech), Asim Hussain (GSF), Camille Fassett (WattTime), Chris Adams (Green Web Foundation), Daniel Schien (University of Bristol), Facundo Armas (Globant), Florent Morel (Amadeus), Francesco Fullone (GrUSP), Mathias Uhlitzsch (Evosoft), Nisha Ramachandra (Accenture), Raghava Rao Battina (HSBC), Riccardo Pomato (Microsoft), Ryan Sholin (Electricity Maps), and Thiago Falcao Silva (NTT DATA). Participants answered structured questions about measurement accuracy and adoption complexity. An LLM synthesised their responses into draft content. The group reviewed and refined through multiple rounds. In contentious areas, participants revised until all objections were resolved.
About the SCI for Web assembly →Consensus reached in ten weeks
Late 2025
The assembly resolved core tensions: a specification that is technically accurate but unused serves no purpose, and a widely used metric that lacks credibility also fails. Three key principles were agreed: accuracy proportionate to control, mandatory disclosure of boundaries and assumptions, and integration into existing workflows over added complexity. The scope was defined — web applications delivering value through browser interfaces via HTTP/HTTPS — with third-party services explicitly required within the measurement boundary.
Read what the SCI for Web assembly agreed →Specification development begins
Q1 2026
With the design foundation established in weeks rather than months, SCI for Web entered active specification development. The process is now proven and replicable — the AI-assisted assembly model can be applied to future GSF standards work, compressing consensus timelines across all new specifications.
Learn about the SCI for Web standard →Traditional standards processes take years. But regulations aren't waiting, and neither is the technology. We needed a way to keep the rigour of multi-stakeholder consensus while compressing the timeline from years to weeks. AI-facilitated consensus gave us that — and now we have a replicable model for every standard we develop.
Asim Hussain
Executive Director, Green Software Foundation
Who came together
The people who made it happen
Asim Hussain
Executive Director
Green Software Foundation
Conceived and led the AI-assisted assembly process, participated as an assembly member, and championed the web sustainability agenda.
Chris Adams
Director of Technology and Policy
Green Web Foundation
Proposed SCI for Web, brokered the W3C collaboration as an active member of both communities, and leads the SCI for Web project.
Joseph Cook
Head of Research
Green Software Foundation
Designed the AI-facilitated consensus mechanism and facilitated the assembly process, enabling 14 experts to move from a blank page to consensus in ten weeks.
In their words
"Traditional standards processes take years. But regulations aren't waiting, and neither is the technology. We needed a way to keep the rigour of multi-stakeholder consensus while compressing the timeline from years to weeks. "
Asim Hussain
Executive Director, Green Software Foundation
"A specification that is technically accurate but unused won't serve its purpose, and a widely used metric that lacks credibility also won't serve the purpose. "
SCI for Web design document
Consensus position from the assembly
Related articles
Designing SCI for Web: What We Agreed and What Comes Next
A consensus-built methodology for measuring the carbon intensity of web applications
The Green Software Foundation and World Wide Web Consortium (W3C) Collaborate to Advance Adoption of Web Sustainability Measurement
Through this agreement, the GSF and W3C aim to promote the adoption of sustainable web development best practices and standards.
Software Carbon Intensity (SCI) Specification Achieves ISO Standard Status, Advancing Green Software Development
We've achieved a big win for green tech: The Software Carbon Intensity (SCI) Specification v1.0 is now an ISO standard.
SCI for AI Specification Ratified: Standard for Measuring AI Emissions Across the Lifecycle
A standardized specification extending the ISO Software Carbon Intensity methodology to measure the carbon emissions of AI systems throughout their lifecycle.