ProcurementVerified Real deployment at a named or precisely described company, with a measured result from a primary source.

AI collects supplier emissions data and builds product carbon footprints for Scope 3 reporting (Western Digital, PALFINGER, Knorr-Bremse, Vishay)

Added 2026-11-09 · Last updated 2026-11-09

Manufacturers in this case: Knorr-Bremse · PALFINGER · Vishay Intertechnology · Western Digital · all manufacturers using AI

Role

Sustainability and procurement (supplier data collection for Scope 3 emissions and product carbon footprints) · Typical users: sustainability or ESG manager, purchasing and supplier quality manager, category manager, product compliance engineer

The problem

Customers and regulations (CSRD, CBAM, EUDR in Europe) increasingly ask manufacturers for emissions data on what they buy and what they sell, and most of it sits at suppliers. Collecting it by email is slow for both sides. At Western Digital, collecting emissions data from component suppliers by email was time-consuming for the sustainability team and for suppliers. Knorr-Bremse managed Scope 3 emissions in spreadsheets with a spend-based method that lacked product and supplier detail. PALFINGER's emissions data was spread over several systems with limited supplier transparency, about 10% of it activity-based. At Vishay, manual life-cycle assessment sometimes double-counted emissions or modelled similar products inconsistently and slowed answers to customer requests.

How AI helps

AI does two jobs. First, it handles the supplier conversation and the data clean-up: at Western Digital, Sluicebox's AI agent makes the data request by email (suppliers do not have to log into a portal or fill in a form), answers supplier questions and turns the raw data into a product carbon footprint aligned with ISO rules, while sustainability staff review the output. Carbmee's AI at Knorr-Bremse and PALFINGER imports procurement and product data, maps transactions, materials and suppliers to emission factors, and then runs a structured supplier programme so that suppliers replace estimates with their own product-level data. Second, it speeds up the calculation: Sluicebox's generative life-cycle assessment models each component from its manufacturing inputs and flags inconsistencies, which at Vishay replaced about two hours of analysis per component. In all cases people review exceptions and own the reported numbers.

Step by step

  1. List your suppliers by spend and by category (steel, aluminium, resins, electronics) and pick the 20 to 50 that cover most of your purchased emissions.
  2. Decide what you need from them (product carbon footprint per part, plant-level energy data, recycled content) and in which standard, with your reporting or auditor team.
  3. Prepare a short, plain request in the supplier's language that explains why you ask, what to send and by when, and accepts a reply by email, since many suppliers will not use a portal.
  4. Let an AI agent send, follow up and answer routine questions, and have your sustainability team review every number that comes back before it is used.
  5. Cross-check the results against bills of materials, industry baselines and other suppliers, and flag outliers for follow-up questions.
  6. Fill gaps for non-responders with transparent estimates, labelled as such, and track the share of spend covered by primary data over time.

Tools

  • Sluicebox (AI email agent for supplier carbon data collection and generative life-cycle assessment): used at Western Digital and Vishay
  • Carbmee Environmental Intelligence System (AI import, emission-factor matching, supplier engagement and product carbon footprints): used at PALFINGER and Knorr-Bremse
  • Jaggaer (procurement platform, integrated with Carbmee): used at Knorr-Bremse
  • Other tools to evaluate: the sustainability module of your ERP or procurement platform, supplier portals that collect product carbon footprints, the emission-factor databases your auditors accept
  • To prepare a request without new software: your top suppliers by spend and a supplier questionnaire, plus any LLM (ChatGPT, Claude, Copilot, Gemini) with the prompt below

In the AI tools index: Carbmee · Sluicebox · all AI tools used by manufacturers

Result

Western Digital (hard drives), Sluicebox, pilot from late 2025: primary carbon data for its biggest suppliers went from 30% to 90% of spend and the time to collect it fell from five to six months to four weeks; suppliers in the pilot made the device housing, baseplate and motor, and staff keep reviewing the outputs (Trellis, 17 July 2026, quoting Western Digital's sustainability programme manager). PALFINGER (cranes and lifting solutions, about 30 production sites), Carbmee: ERP systems of 79 legal entities and about 5,000 suppliers connected in four weeks; activity-based coverage up to 75% from about 10%; manual carbon-reporting effort down about 85%; first certified product carbon footprints, used in tenders; primary supplier data covering 33% of raw-material emissions (Carbmee case study, a vendor publication). Knorr-Bremse (rail and commercial-vehicle braking systems), Carbmee: 300,000 SKUs and 1,000 suppliers loaded in the first week, 840 suppliers engaged to provide product-level data, integration with Jaggaer; the case study describes a data foundation and states no measured saving (Carbmee case study). Vishay Intertechnology (electronic components), Sluicebox: analysis time per component cut from about 2 hours to seconds (7,200x faster) and results within 8 to 12% of manual life-cycle assessments by experts (Sluicebox customer story, a vendor publication).

As reported by the source; most figures are vendor or self-reported and not independently verified.

Human review: cautions

  • The figures come from the vendors' own case studies and from one press article; Knorr-Bremse's page states no measured result, and Western Digital's 90% covers its biggest suppliers in a pilot.
  • Speed is not accuracy: a footprint computed quickly still needs review by someone who understands the method and boundaries, and an auditor may not accept AI-estimated values.
  • Suppliers can be wary of sharing data with an AI agent or of how it will be used: say who sees it, and keep a named contact at your company.
  • Rules such as CSRD, CBAM and EUDR change and depend on your size and sector: check the current requirements with your sustainability or legal advisers.

Try it yourself

You are a sustainability and procurement analyst at a manufacturer preparing to collect Scope 3 emissions data from suppliers.

I will paste: (1) our top 30 suppliers by spend [name, country, what we buy, annual spend]; (2) the standard or customer request we must answer [for example CSRD, a customer questionnaire]; (3) the product carbon footprint data we already have, if any.

Do this:
1. Group the suppliers by category and tell me which ones probably carry most of our purchased-goods emissions, and why, without giving any emission figures.
2. List the exact data points we should ask each category for, and what a good answer looks like (unit, boundary, year, method).
3. Draft one short supplier request email in plain English, and say what should be adapted for suppliers who do not read English.
4. Write a checklist for reviewing the answers (unit errors, missing boundaries, values far from similar suppliers, double counting).
5. Propose a way to fill the gaps for suppliers who do not answer, clearly labelled as estimates.

This is not a certified carbon calculation: do not make up emission values, and flag every question where a specialist or auditor must decide. Use only the data I paste in; if something is missing, write 'not found' or list it as a question instead of guessing. Do not send, order, post or change anything in any system: produce drafts for my review and approval.

Want more like this? Join the newsletter

Related cases

All 15 Procurement use cases →

Sources

Weekly: 1–2 practical AI use cases for your plant

Newsletter in

Free, one email a week. By subscribing you agree to receive the weekly newsletter (and its sponsor message) from AI Agents for Manufacturing. Every email includes an unsubscribe link, and you can unsubscribe anytime. See our Privacy Policy.