Challenge

The old way stalled at 10 suppliers in 6 months

Primary carbon data matters to Western Digital for two reasons. Internally, it improves tracking toward the company's public Scope 3 direct material goal and shifts reporting away from spend-based modeling and EEIO emission factors toward primary data in its LCAs and PCFs. Externally, its customers, the hyperscalers, are asking for a shift away from spend-based models.

The traditional approach did not scale. Western Digital reached out to 100-plus suppliers and collected data manually through emails, which produced long cycles and low response rates: 5 to 6 months of turnaround with only 10 suppliers covered. For suppliers, each request meant forms and platform logins against an expertise gap and a lack of resources. Slow. Incomplete. Repeats every cycle.

Western Digital was not alone. Nearly 10,000 companies now need supplier carbon data, yet only 2% of suppliers have ever produced a product-level footprint.

Solution

We made it easy for the supplier

Working with Sluicebox, the team introduced an agentic workflow called Lucy, the supplier AI agent. Lucy runs the whole exchange inside the email chain, and every supplier walks away with its own ISO 14067 PCF.

A human stays in the loop. The agent organizes, drafts, and collects, and whenever judgment or a decision is involved, a person steps in. That is how Western Digital maintains trust and transparency with its suppliers. The rollout went in two phases: a pilot with a few suppliers first, then scaling across 60 to 70 suppliers.

Impact

Coverage tripled, cycle time cut 6x

Measured against Western Digital's own traditional approach, the change is stark. The old way covered 10 suppliers in a 5-to-6-month cycle at low response rates. With Lucy, primary-data coverage of direct material spend rose from 30% to 90%, the cycle closed in 4 weeks, and 99% of supplier emails ran agent-assisted with a human in the loop.

The incentive ran the other way too: suppliers that shared their data got back a free, ISO-aligned PCF of their own.

Why it matters

Primary data revealed lower emissions

Primary data also changed what the numbers said. Spend-based EEIO factors tend to overestimate emissions. When Western Digital replaced them with actual primary data and the real decarbonization work suppliers had done, the footprint came in lower than what had been reported before.

What's next

PCFs that hold up in audit, even for non-responders

Comprehensive checks and smarter follow-up across the long tail of suppliers.

Verify every PCF with CDP. Every submission is cross-checked against the supplier's CDP disclosure and industry baselines.

Audit the outliers. Flagged numbers trigger the agent to request the supplier's calculation and surface where assumptions break.

Estimate the gaps. For non-responders, use Sluicebox to build a defensible PCF estimate from the bill of materials, commodity, and manufacturing location.