Western Digital cuts supplier emissions data collection from 6 months to 4 weeks with AI

Western Digital, a data storage company that manufactures drives for personal computers and data centers, partnered with Sluicebox—a startup focused on collecting carbon emissions data from the electronics industry—to automate and accelerate the collection of supplier emissions data. The challenge was familiar: gathering carbon footprint information from suppliers via email was time-consuming for both Western Digital's sustainability team and the suppliers themselves, resulting in low response rates and lengthy cycles.

To address this, Sluicebox built an AI agent that handles the entire email workflow. Rather than asking suppliers to fill out forms or log into a platform, the agent sends the data request, responds to supplier questions in real time, and synthesizes the raw responses into a product carbon footprint aligned with International Organization for Standardization standards. The pilot, rolled out late last year, covered major suppliers of three key components: device housing, baseplate, and motor.

The results were substantial. Western Digital expanded primary carbon data coverage from its biggest suppliers from 30 to 90 percent—a three-fold increase in coverage. Equally important, the time required to collect and synthesize this data fell from between five and six months to four weeks. Mrinalini Iyer, Western Digital's program manager for sustainability operations, presented these findings at the AI x Sustainability Showcase at the Trellis Impact 26 event in San Francisco last month.

Iyer was explicit that the deployment was not about replacing sustainability staff. "Keeping people at the center is an important part of Western Digital's AI approach," she said. Team members continue to review all outputs and maintain trust in the process. She emphasized that in sustainability data collection, "the goal is not only speed — it is usable, traceable and defensible data." The AI agent's role is to eliminate the labor-intensive manual chasing and data wrangling, freeing the sustainability team to focus on quality review, supplier engagement, validation of methodology, and decision-making.

Looking ahead, Western Digital is expanding the system's capabilities. The company is now using the AI to estimate missing data for suppliers who don't respond by analyzing available inputs such as bills of materials, cross-checking the resulting carbon footprints against supplier disclosures and industry baselines, and flagging outliers that warrant human follow-up. Iyer noted that the company expects the project to evolve from a pure data collection workflow into a broader system of engagement, validation, and exception management.

Western Digital's pilot addresses a widespread operational bottleneck in corporate sustainability: the time and effort required to gather emissions data from supply chains. Traditional approaches—where sustainability teams manually chase suppliers via email, navigate form responses, and consolidate raw data—do not scale efficiently as reporting and regulatory demands grow. By deploying an AI agent that handles the repetitive communication and initial data synthesis, the company achieved an 8.7× reduction in collection time (from ~5.5 months to 4 weeks) and tripled coverage among major suppliers.

The framing of the pilot is deliberate: Iyer presented it as a redistribution of labor rather than a reduction in headcount. The sustainability team's work shifts from manual, low-value tasks (chasing suppliers, wrangling spreadsheets) to higher-value activities (validating assumptions, engaging suppliers on methodology, reviewing exceptions). This reframing may be especially important to address concerns that automation will simply shrink sustainability functions rather than deepen their impact. The insistence on human review and the alignment of data synthesis with International Organization for Standardization rules signals that speed alone is not the goal—the data must remain "usable, traceable and defensible," as Iyer noted.

Looking forward, Western Digital plans to extend the system beyond collection into validation and exception management, including automated gap-filling for non-responding suppliers using bills of materials and baseline comparisons. This evolution suggests the company sees the AI agent not as a one-time efficiency gain but as the foundation of a broader supplier engagement and compliance infrastructure.

No comments yet. Be the first to share your thoughts!