As generative AI continues to scale, a shift is occurring at the intersection of sustainability and technology. For years, reporting teams were bogged down by scattered data and labor-intensive documentation, leaving little space for strategic sustainability work. But for those now adopting AI, they’re spending less time trying to answer for the impact of last year, and more time on proactive resilience, gaining efficiency and speed.
Sustainability teams often need to build reports using data from multiple disparate sources such as, financial data, real estate footprints, energy usage and materials consumption. The standards for reporting are also rapidly evolving. That presents a unique challenge, where accuracy is difficult to obtain across data sources, and the stakes for getting it incorrect are high. At Google, we’ve spent the last two years testing and integrating AI into our own environmental reporting process to solve for this.
Inside how Google uses of AI for reporting
We first turned our own internal reporting processes into a testing ground for AI solutions. In developing our Environmental Report, we tested Gemini as a first line of review for validating environmental claims, automatically cross-referencing draft claims against our internal policies and best practices. This automation doesn’t replace the expert; rather, it empowers them by freeing the human reviewer to focus on validating the assessment rather than starting from scratch for every claim. We also used NotebookLM to transform our static Environmental Report into an interactive knowledge base, allowing users to query complex data and receive instant, cited answers.
Our team documented our experimentation and progress—including the prompts that worked and lessons learned from the ones that didn’t—in our open-source AI playbook. In documenting our efforts we realized that AI is more than just a tool for efficiency—it’s a catalyst for impact. By streamlining the manual, complex mechanics of reporting, we can all spend less time managing files and data and more time driving the strategy that moves the world forward.
Building a sustainability data lake at Equinix
Digital infrastructure provider Equinix undertook their own reporting transformation with the help of Google Cloud. Facing a 46% year-over-year increase in customer sustainability requests, the Equinix team realized that manual spreadsheets were no longer a viable option. They needed a tool for real-time decision-making.
Equinix built a Sustainability Data Lake on BigQuery. They ingest data from 240+ global sites automatically, transforming their reporting cycle from weeks of manual data cleaning to on-demand insights.
“I’m no longer just a data analyst. I’m a strategic advisor,” says Alexa Cotton, Senior Manager of Sustainability at Equinix. “Our sustainability data is now a strategic asset that impacts over 60% of our ARR [Annual Recurring Revenue]. We’ve moved from reactive mode to looking at automated actions that save both energy and money.”
Sustainable by design: using the well-architected framework (WAF)
Equinix demonstrated that a solid data foundation is the prerequisite for any AI-driven breakthrough. They also demonstrated that modernizing legacy processes often come with efficiency gains when they are well-architected.
Equinix moved to a serverless architecture with BigQuery, they achieved a “triple win” of price, performance, and footprint:
Equinix leveraged serverless elasticity to ensure they only use (and pay for) the exact compute resources required. No “zombie” servers, no wasted energy.
BigQuery handles the scaling and optimization of workloads programmatically. This removes the human error in resource allocation, ensuring the data lake remains lean and high-performing as it grows.
By leveraging Google’s carbon-intelligent data infrastructure, Equinix fundamentally improved their “performance per watt,” turning a reporting requirement into a showcase of operational efficiency.
These are great examples of what we refer to as the “4Ms” of our well-architected framework: Machine, Model, Mechanization, and Map.
The ambition loop
The work being done at Equinix creates what we call an “ambition loop.” When you intervene at the architectural level, you aren’t just checking a box for a sustainability report; you’re improving your economics. Which in turn improves your sustainability outcomes and ultimately the success you report.
To learn more, dive into Google’s AI Playbook for Sustainability Reporting and explore the new WAF sustainability pillar to start your own data-to-AI journey. Together, we can build the future of reporting—and a more resilient world.