Smart carbon data preparation.
Claia reads accounting exports and client documents and hands back classified, traceable, analysis-ready carbon data. Days saved on every assessment, line-level precision, your rules and your templates: you keep the analysis, Claia does the rest.





to 1 hourof preparation per engagement
For carbon experts.
Claia reads your clients’ accounting and documents, computes emissions, classifies every expense and hands you a deliverable with what remains to review and why.
Emissions are computed from expenses, by category and by scope.

The tools that speed up your analysis.
AI-enriched accounting analysis, document reading, supplier lookup.
- French accounting exports recognised, private and public sector
- Business rules and reference data, completed by AI when needed
- Every classification is traced and can be revised
- Quantities (kWh, litres, tonnes), periods, amounts, supplier
- Measured data replaces the spend-based estimate
- Actual activity and industry code retrieved automatically
- Category and emission factor corrected accordingly
- Environmental product declarations for materials and construction
Less manual work.
More time for analysis.
- Data hosted in France
- No training on your files
- Every result traceable
Consulting firms first.
Do you produce carbon footprints for clients?

Gerardo Gonzalez, computer engineer, is building Claia after several years at EY Paris as a data science manager, where he designed document-analysis products supporting audit engagements. Claia is built with the firms that carry out these analyses, on their real files. Claia was born at Station F in Paris.
Calibrated on French data, ready to travel.
Claia is calibrated on French accounting data: statutory exports, public-sector ledgers, the ADEME Base Carbone® emission-factor database, supplier identification through the official French registries (SIRENE, INPI). The chart of accounts and the factor database are pluggable. Working with another country’s data? Get in touch and we will look at your case together.


