Technical Architecture
Carbon Intelligence™
Raw data in, strategic intelligence out. Carbon Intelligence™ combines a transparent, auditable emissions methodology with impression-level granularity and an AI optimization engine that proves carbon reduction drives better performance. Every calculation is traceable, every recommendation is backed by data, and every result is exportable — from real-time dashboards to CSRD-ready reports.
1
Data Ingestion & Smart Analysis
Flexible multi-source connection
Connect your campaign data through the method that suits you best: CSV upload for quick one-off analyses, API script for automated recurring syncs, or Data Warehouse for enterprise-scale integration. Whatever the source, the Carbon Intelligence™ engine instantly triggers a comprehensive automated analysis of your data.
CSV · API · Warehouse→
Column detection→
Granularity identification→
Total rows cleanup→
Normalization
Column Detection
Automatic header mapping: impressions, clicks, spend, conversions, video views, sites, apps, formats, devices, countries...
Granularity
Detail level identification: by site, by app, by campaign line, by creative, by device...
Cleanup
Automatic removal of "Total", "Grand Total" rows, subtotals, empty rows, and duplicates
Validation
Data integrity checks, metric consistency verification, anomaly detection
⬇
2
Data Calculator
GMSF Aligned · GHG Protocol · ISO 14064
The calculation engine applies certified methodologies to convert each impression into its carbon equivalent, with the maximum available granularity.
// gCO2 per impression calculation — GMSF aligned
gCO2/impression = (
EnergyPerImpression(device, format, duration)
× GridCarbonIntensity(country, region)
× DataTransferFactor(creative_weight, ad_calls)
+ ServerProcessingEmissions(supply_chain_hops)
)
// Applied per row with max granularity:
// → per Site / App
// → per Device (mobile, desktop, CTV, tablet)
// → per Format (display, video, native, audio)
// → per Country / Region (local grid intensity)
Total Campaign CO2 = Σ (gCO2/imp × impressions) // per row
gCO2/impression = (
EnergyPerImpression(device, format, duration)
× GridCarbonIntensity(country, region)
× DataTransferFactor(creative_weight, ad_calls)
+ ServerProcessingEmissions(supply_chain_hops)
)
// Applied per row with max granularity:
// → per Site / App
// → per Device (mobile, desktop, CTV, tablet)
// → per Format (display, video, native, audio)
// → per Country / Region (local grid intensity)
Total Campaign CO2 = Σ (gCO2/imp × impressions) // per row
⬇
3
Intelligent Analytics
Advanced Calculations Engine
Calculated data feeds the advanced analytics engine that generates granular recommendations, simulates scenarios, and enriches business metrics.
Advanced Recommendations
Multi-dimensional calculations crossing carbon performance × media performance → actionable recommendations per site/app/format/device
22 What-If Scenarios
Reduction simulations: exclude top polluters, device shift, frequency cap, green geo-targeting, format switch...
Carbon ROI & ROAS Enricher
ROAS enrichment with carbon dimension: cost per gCO2 saved, green ROI per dollar, carbon-adjusted CPM
Blocklists Generator
Automatic exclusion lists compatible with all DSPs (DV360, TTD, Xandr, Amazon) — highest-polluting sites/apps
⬇
4
AI Analysis Engine AI-Powered
Multi-LLM · Agnostic · Natural Language
The entire intelligent analysis is translated by the Carbon Intelligence™ engine and sent to a multi-LLM agnostic system that interprets it and returns its response in natural language.
Internal AI Pipeline
Full analysis
from engine→ Structured
translation→ Multi-LLM
agnostic→ AI response
natural language
from engine→ Structured
translation→ Multi-LLM
agnostic→ AI response
natural language
ROI Forecasts
Return on investment projections integrating the carbon variable — short and medium-term predictions
Carbon Predictions
Predictive modeling of future emissions based on detected patterns and market trends
Deep Insights
Counter-intuitive discoveries, hidden correlations, emerging patterns — clearly explained in human language
Implementation Roadmap
Prioritized action plan with concrete steps, suggested timeline, and estimated impact per action
⬇
5
Exports & Deliverables
Multi-format · CSRD Ready
All outputs — AI recommendations, full report, enriched raw data — are exportable in every required format, ready for CSRD reporting.
Excel
Complete dataset
CSV
DSP / BI import
JSON
API & integrations
HTML
Interactive report
PDF
Official report
PowerPoint
Client presentation
Export Contents
AI Recommendations — natural language analysis, insights, roadmap
Full Report — all metrics, charts, benchmarks
Enriched Data — every row with gCO2, scores, flags
Blocklists — ready-to-import files for all DSPs
Full Report — all metrics, charts, benchmarks
Enriched Data — every row with gCO2, scores, flags
Blocklists — ready-to-import files for all DSPs
🇪🇺 CSRD Ready — Compliant with European extra-financial reporting requirements