Carbon2Data
Sign inRequest a demo

Scope 3 · from purchase orders

Your supply chain emissions are already in your ERP. They just don’t look like emissions yet.

Carbon2Data turns the purchase-order export you already have into a classified supplier list and a sourced, spend-based estimate, with the arithmetic behind every number.

Request a demoSee how it works
po_export_FY2025.xlsx5 suppliers · 4 sourced · 1 flagged
SupplierCountryIndustrySpend · factor
Northport LogisticsMalaysiaTransportation and travel servicesUS$10,710 sourced
Kestrel PackagingSingaporeManufacturing and industrial productsUS$9,940 sourced
Halden SteelworksGermanyMetals and metal productsUS$8,010 sourced
Brightdesk Office SupplyUnknownRetail and trade servicesUS$1,180 default
Tasik Catering ServicesMalaysiaHospitality and entertainmentUS$730 sourced
Illustrative rows. Invented suppliers, real Bursa CSI categories.

Scope 1 and 2 come from meters and bills. Scope 3 is the hard one: it lives in thousands of supplier invoices nobody wrote with emissions in mind. We start there, with the data you already have.

(01) How it works

From export to estimate, in six steps.

41,285purchase-order lines
1,130suppliers

One real year of spend, collapsed.

  1. 01

    Start with the export you already have

    Purchase-order lines exactly as your ERP produced them, as .xlsx or .csv. No template, no retyping.

  2. 02

    Columns read as they come

    Carbon2Data works out which column is the vendor, the amount, the currency and the country, and tells you exactly what it found before anything is committed.

  3. 03

    Lines collapse into suppliers

    A year of spend becomes a supplier list you can actually work with: one row per supplier, every transaction still underneath it.

  4. 04

    Classified against a real disclosure taxonomy

    Every supplier lands in the Bursa CSI supplier-industry taxonomy (30 categories), the one your exchange already uses. Bring your own list instead if you have one.

  5. 05

    Country and industry mapping is where you take over

    What the data does not settle, you decide. Once. The decision is remembered for every future import.

  6. 06

    Emissions by reporting period

    A spend-based estimate for every supplier and period, with a methodology page that shows the arithmetic behind any single transaction.

(02) The AI

Here is exactly where the AI sits.

Most of the work is rules and registries. The model is the fourth thing we ask, not the first, and it is allowed to say it does not know.

  1. 01

    Your own data

    A supplier you have already mapped stays mapped. Your decisions always win first.

  2. 02

    Legal suffix rule

    Sdn. Bhd., GmbH, S.A., LLC: a legal form says a lot about where a company is registered, in any jurisdiction.

  3. 03

    Global LEI registry

    The Legal Entity Identifier record, when the supplier has one.

  4. 04

    The model

    Only now does a language model get a say, on what the rules above could not settle.

  5. 05

    Unknown

    When it still is not sure, it returns unknown and asks you, rather than inventing an answer.

(03) Method

An estimate never becomes a fact. Every number shows its working.

For a single transaction

estimated_emissions (kg CO₂e)
  = spend_amount
  × currency_rate_to_usd
  × (country_factor + industry_factor) / 2

A spend-based estimate, the method the GHG Protocol’s Scope 3 guidance accepts when supplier-specific data is not yet available. Coarser than a supplier’s own reported figure, and labelled that way.

Country factor
World Bank, carbon intensity of GDP (EN.GHG.CO2.RT.GDP.PP.KD). 191 of 249 countries have a real value; the rest use the median, never a guess.
Industry factor
US EPA Supply Chain GHG Emission Factors v1.3: 1,016 NAICS sectors, aggregated into 13 broad ones.
Provenance
sourced factors are marked sourced. default factors are flagged everywhere they appear, down to the transaction.

(04) Suppliers

Missing something only the supplier knows? Ask them.

The outreach email is drafted for you. Suppliers answer through a private link: no account, no login, no portal to learn. Their reported figures replace the estimate, and the record says so.

(05) Workflows

Then let it run itself.

When a supplier is

  • High spend
  • High risk
  • Unmapped
  • Missing data

Send it to

  • Slack
  • Webhook
  • ESG database
  • Email template

Scope 1. Scope 2. And the hard one.

Bring a real export. We will walk through your own suppliers, live, and show you where every number comes from.