Every tonne traceable.
Every euro explained.
How we build a carbon account your auditor can follow and an action plan your CFO can finance: the standards we apply, where every factor comes from, what AI does and what our experts sign off, and how sure we are of each figure.
- GHG Protocol · ISO 14064-1
- Bilan Carbone® Conform (BC-C)
- Official BENOR partner
- 1 · Source document
Electricity_2025.pdf · page 2
Ghent plant · 01/01–31/12/2025
Active energy, total ........ 526,000 kWh - 2 · Activity data
526,000 kWh · Ghent plant · 2025
- 3 · Emission factor
× 0.222 kgCO₂e/kWh · Grid electricity, Belgium
Database · version · year · geography stored with the line
- 4 · Result
116.8 tCO₂eGHG Protocol: Scope 2Bilan Carbone: Energy
Uncertainty: lowMatched by AI, checked by an expert
Standards
The frameworks behind every figure.
One set of activity data, read through the framework each audience expects: your auditor, your customers, your bank or a public tender.
- Company footprint
GHG Protocol
Corporate Standard, Scope 2 Guidance and Scope 3 Standard. It sets the scopes, the boundaries and how electricity is counted.
OutputScopes 1, 2 and 3
- Company footprint
ISO 14064-1
Our inventories follow its structure and documentation, so a third party can verify them.
OutputReady for verification
- Company footprint
Bilan Carbone®
Widely used in France, and in Belgium too. Same lines, other categories.
Bilan Carbone Conform (BC-C)OutputBilan Carbone report
- Targets
SBTi
Reduction targets and trajectories built on the Science Based Targets initiative criteria, checked year after year.
OutputTrajectory & targets
- Products
ISO 14067 · ISO 14040/44
Product carbon footprints and life cycle assessments, per unit you sell.
OutputPCF · LCA
- Products
EPD · EN 15804+A2
We prepare the LCA and the data. An independent EPD programme verifies and publishes it.
OutputConstruction products
- Reporting
CSRD (ESRS E1) · VSME
Your climate disclosures come from the same account. No second calculation for the report.
OutputSustainability report
- Tenders
CO2 Performance Ladder
Handbook 4.0 (Steps) and Handbook 3.1 (Levels).
Official BENOR partnerOutputPublic tenders
Process
How a carbon account is built.
Six steps, the same every year. Each one leaves a written trace, so next year's figures can be compared with this year's.
Set the boundaries.
Before a single invoice is read, we agree with you on what is in and what is out, and why.
What we decide together
- Consolidation approach: operational control by default, financial control or equity share if your group reports that way
- Entities, sites and activities included
- Reporting year and base year
- All seven Kyoto gases, refrigerant leaks included
- Screening of the 15 Scope 3 categories, with the reason for each exclusion
What you get
- A boundary note, signed off by you
- The list of data to collect, per site and per person
Collect activity data.
We start from what physically happened: kWh, litres, tonnes, kilometres. Spend in euros is only a starting point when nothing else exists.
Where the data comes from
- Energy invoices and meter readings
- Fuel cards, fleet and freight data
- ERP and purchase ledgers
- Supplier answers, collected in Carbon Cockpit
- Your own AI tools: Copilot, ChatGPT or Claude
What you get
- Every line linked to its source file
- Missing data flagged, not guessed
Match emission factors.
AI reads each line and proposes the factor. An expert approves it, or picks a better one. No factor match enters the account without a person's approval.
Stored with each factor
- Database and version
- Year and geography
- Unit and GWP set (IPCC AR6, 100 years)
- Who approved it, and when
What you get
- A factor library you can open and audit
- The best available factor for each line (see the hierarchy below)
Score quality and uncertainty.
Each line gets a data-quality score and an uncertainty. You see which figures are solid and which need better data.
How we score
- The 5 GHG Protocol data-quality criteria: technology, time, geography, completeness, reliability
- An uncertainty per line, combined into an uncertainty on the total
What you get
- Total emissions with their ± range
- The 3 lines to improve first
Review and sign off.
We go through the results with the people who run the sites. Outliers are explained, not smoothed away.
What we check
- Year-on-year changes above a threshold
- Ratios per site, per tonne produced, per employee
- Lines with low confidence or high weight
What you get
- Results validated by a D‑Carbonize expert
- A full audit trail: who changed what, and when
Report, then repeat.
One account, several reports. Next year, same scope and same sources, so the trend means something.
Rules we apply
- GHG Protocol and Bilan Carbone reports from the same lines
- Scope 2 shown location-based and market-based
- Base year recalculated after a change of 5% or more of total emissions
- Errors found later are corrected in past years, and the change is shown
What you get
- Reports for your auditor, customers and tenders
- Data ready for CSRD or VSME
- Exports in open formats
Traceability
Follow one line from the invoice to the report.
Pick a line. Switch the framework. For electricity, switch the Scope 2 method. The tonnes are always the same; only the way they are read changes.
1Source document
PDFElectricity_2025.pdf · p. 2
Ghent plant · 2025
2Activity data
526,000 kWh
Metered consumption
3Emission factor
× 0.222 kgCO₂e/kWh
Grid electricity, Belgium · average mix
4Result
116.8 tCO₂e
Read through the GHG Protocol
Scope 2 · Purchased electricity · location-basedData type: metered
Data quality · 5 GHG Protocol criteria
- Technology
- Time
- Geography
- Completeness
- Reliability
Uncertainty: low
What it tells you
Location-based shows the emissions of the grid you draw from. It is the figure to compare sites and years.
Illustrative example. Same Example company as on our Solutions pages.
Emission factors
The best available factor, for every line.
Not every line deserves the same effort. We use the most specific factor we can get, and we tell you where a generic one is still in use.
- Supplier-specificBest
Your supplier's product footprint or EPD, or your own energy contract. The most precise.
- Product and process averagesStandard
Recognised databases, per material, fuel, process and country.
- Base Empreinte® (ADEME)
- ecoinvent
- INIES
- DESNZ
- AIB
- Physical estimatesFlagged
A quantity estimated from a proxy (m², km, headcount) when a meter or invoice is missing. Flagged as an estimate.
- Spend-basedLast resort
Euros spent × a sector factor. Only as a starting point for Scope 3, flagged, and replaced year after year.
- Every factor is stored with its source, version, year and geography.
- A database update never silently changes a closed year.
- Your supplier data replaces the average as soon as it is checked.
- Biogenic CO₂ is reported separately from the scopes.
Illustrative example. Share of Scope 3 tonnes.
Uncertainty
We tell you how sure we are.
A footprint without uncertainty looks precise and decides badly. Every line gets a quality score on the five GHG Protocol criteria, and the uncertainties are combined into a range for the total.
- Technology
- Does the factor match the real process or product?
- Time
- Is the data from the reporting year?
- Geography
- Is the factor for the right country or grid?
- Completeness
- Does the data cover all sites and all months?
- Reliability
- Measured, calculated, or estimated?
601 tCO₂e± 9%
95% confidence range544 – 658 tCO₂e
95% confidence. Line uncertainties combined as independent errors. “Share of ±” is each line's part of the total uncertainty.
Improve first: Steel sheet and Buildings carry 89% of the uncertainty. With product footprints from your two steel suppliers (±10%), the total goes from ±9% to about ±7%.
AI and experts
AI reads. Experts decide.
AI takes the repetitive work: reading, extracting, matching. Every choice that changes a figure stays with a person who knows carbon accounting.
What AI does
Fast, consistent, every year.
- Reads invoices, meter files, fuel cards and ledgers as they are
- Extracts quantities, units, periods and sites
- Proposes an emission factor, with its confidence
- Flags gaps, duplicates and unusual values
What our experts do
Judgement, accountability, sign-off.
- Set the boundaries and the Scope 3 screening with you
- Approve or correct each proposed factor
- Review high-weight and low-confidence lines
- Sign off the results and explain them to your teams
Gas_invoice_2025.pdf · Ghent plant555,400 kWh
AI proposes
Natural gas, combustion0.202 kgCO₂e/kWh
IPCC default · Scope 1 · stationary combustion
Confidence94%
Line result112.2 tCO₂e
Approved by a D-Carbonize expert. Logged in the audit trail.
- A person approves every factor match.Nothing enters the account on AI alone.
- Missing data is flagged, not guessed.Gaps stay visible until they are filled.
- Every suggestion shows its source.Click a figure, see the document line.
- Your data stays in the EU.Never used to train AI models.
From tonnes to euros
The same rigour for the euros.
A reduction plan is an investment plan. Each measure is costed with the same traceability as the tonnes, and every assumption is written down and dated.
- CAPEX
- From quotes or reference costs, spread over the measure's lifetime
- Savings
- Energy and operating costs avoided, at stated prices
- Payback
- CAPEX divided by yearly savings
- € per tCO₂e
- Net yearly cost divided by tonnes avoided. Below zero, the measure pays for itself
- Subsidies
- Included when you are eligible, with the scheme named
Heat recovery · Ghent plantCuts gas use by 30%: 167 MWh a year
Example company- CAPEX
- €35k
- Yearly savings
- €7.5k
- Payback
- 4.7 years
- Avoided
- 33.7 t/yr
Cumulative cash, undiscounted+€40k after 10 years
( €3.5k CAPEX / yr − €7.5k savings ) ÷ 33.7 tCO₂e = −119 €/t
Below zero: the measure saves €4.0k a year, net of its cost.- Gas price: €45/MWh · 2026
- Lifetime: 10 years, undiscounted
- No carbon price
- Subsidy: none applied
Our limits
What we don't do.
A method is also defined by its refusals.
Offsets, if any, are reported apart. Your footprint only goes down when your emissions do.
Every estimate is labelled as one, with the proxy it comes from.
You see the range of the total, and which lines it comes from.
A person approves every factor match, and the approval is logged.
Your lines, factors and reports export in open formats.
FAQ
Questions auditors ask us.
Both. The GHG Protocol asks for both figures. Location-based reflects the grid you draw from; market-based reflects your contracts and guarantees of origin. Both are always available. You choose the one that drives your targets, and you keep it from year to year.
We screen all 15 categories of the GHG Protocol Scope 3 Standard. Relevant ones are calculated; the others are documented with the reason they are excluded. For most industrial companies, purchased goods and transport dominate. For makers of energy-using products, so does the use of sold products.
After an acquisition, a divestment or a method change that moves total emissions by 5% or more, the base year is recalculated so that the trend stays comparable. The reason and the impact are written into the report.
Yes. It is built to be verified by an independent body, with limited or reasonable assurance (ISO 14064-3, ISAE 3410). We do not verify our own work.
Yes. Every line links to its source file, its factor and the person who approved it. Your auditor can get read access to Carbon Cockpit or receive a full export.
Yes. We import it, check the boundaries and factors, and tell you what can be kept and what needs to be redone to stay comparable.
Usually both, for different readers. They use the same activity data and classify it differently. We produce both from the same lines, and reconcile them line by line.
Next step
See the method on your own data.
Thirty minutes, screen shared, with a D-Carbonize expert. Bring one invoice and one question.
- 300+companies
- 59countries
- 4.8/5 Capterra
Your 30-minute session
- 0:00Your boundaries and your reporting needs
- 0:10One of your invoices, traced line by line
- 0:20Factors, uncertainty and what to improve first
- 0:30Your questions