What AI technologies are being used

We use LLMs to primarily categorise and extract data from reports supplied to us or those that we find publicly, before, where appropriate, reviewing by our analyst team. No customer data ever gets passed to any LLM.

How we ensure security and quality

We use LLMs to primarily categorise and extract data from reports supplied to us or those that we find publicly, before, where appropriate, reviewing by our analyst team. We process snippets of text without any further context. We have trained a model which is now very adept at matching them up correctly in 297,979 different ways. No customer data ever gets passed to any LLM.

What we are NOT doing

We are not feeding in your private supplier or investments data into a 3rd party LLM. Your sensitive business data never leaves our secure environment.

Trust centre

Independently verified Scope 3 calculations

DitchCarbon provides verified emissions data for over 2 million organisations, so procurement, sustainability and finance teams can measure and act on supply chain and portfolio emissions from one source. Three separate pieces of evidence sit behind that word "verified", and they cover different things. UL Solutions verified our calculator against a standard. Globus Thenken independently assessed our emission factor methodology. Ten different third-party assurance providers have assured customer emissions reports built on our data. Every report we can publish is downloadable below.

Last reviewed 28 July 2026.

Assured downstream, ten times over

DitchCarbon data has been used in emissions reports that were subsequently assured by ten different third-party assurance providers, including Big Four firms. Ten separate reviewers applying their own scrutiny, and the data held up each time.

DitchCarbon is the only specialist Scope 3 tool with third-party assurance of its calculation methodology. See how every vendor we checked compares.
‍See how every vendor we checked compares.

How do we keep your data secure?

Four things, and you can check three of them yourself: an annual SOC 2 Type II audit, monthly penetration testing by independent security experts, a live status page, and a hard rule that no customer data ever reaches an external AI model.

SOC 2 Type II

We hold a SOC 2 Type II report covering the security controls around your data, audited annually by an independent firm. That covers access control, change management, monitoring and incident response across the systems that hold your organisation and spend data. Security reviewers can request the full report, and we answer security questionnaires directly rather than asking you to take our word for it.

Request the SOC 2 report

Monthly penetration testing

Independent security experts penetration test our environment every month, and we fix what they find. You can request the latest report, which names the tester, the scope and the findings, including anything still open.

Request the latest penetration test report

AI data privacy

Your proprietary data is never passed to any LLM or external AI service. We use a combination of in-house models and selected external tools to categorise and extract data from public reports and from documents organisations send us, and our analyst team reviews the output. The rule is absolute on the customer side: your spend files, supplier lists and portfolio holdings never leave our environment.

Read how we use AI

System status

Our status page shows live uptime and any open incident, so you can check availability before a reporting deadline rather than after one. Procurement decisions and board reporting both run to fixed dates, and the status page is public for exactly that reason.

Check live system status

How does DitchCarbon use AI?

We use language models for one job: turning inconsistently written public disclosures into structured, comparable data. Customer data is not part of that process.

How do we keep extraction accurate?

The hard problem is that organisations write the same thing hundreds of thousands of different ways. There are fewer than 30 standard GHG Protocol scopes and categories, and we have seen an enormous number of variations on how they are labelled, across languages and formats. We trained a model to resolve them, so "SCOPE 3 Emissions from Investments" and "Emissioni Scope III, investments" both become scope_3/investments. Rules-based validations then route anything anomalous to an analyst before it reaches your data.

Who relies on this data

Used by Fortune 100 companies in their sustainability reports

Amazon, SAP, Haleon, Coherent, Howden, EssilorLuxottica, Hikma and Grant Thornton use DitchCarbon data.

Which reporting frameworks does our data map to?

Our data is structured to the GHG Protocol from the point of extraction, so it maps to the frameworks you report against without you rebuilding it. Coverage gaps are labelled rather than filled silently, because your auditor will ask.

GHG Protocol

Structured to the Corporate Value Chain (Scope 3) Standard, all 15 categories

SBTi

Baseline and target data in the shape SBTi validation asks for

CSRD

Scope 3 data points and value chain coverage for ESRS E1 disclosure

ISSB

Scope 3 data with the provenance IFRS S2 disclosure requires

California SB-1368

Scope 3 data and provenance for California climate disclosure

Working on a portfolio? The same verified layer supports PCAF data quality scoring.

See the verification working on your own data

The reports above cover the calculator and the methodology. A walkthrough shows both running against the organisations in your own supply chain or portfolio, with the source behind every figure and the gaps labelled.

Request a walkthrough

Read the full calculation methodology

Explore company emissions data at scale

Explore company emissionsdata at scale