IndustrySustainability / CleanTechLocationUK / Isle of ManCompliance FrameworkISO-Compliant GHG Reporting, GDPRPlatform TypeCarbon Accounting & Energy Management SaaS + AITimeline2+ years (scope expanded with jurisdictions)Platform StackNext.js · TypeScript · IoT · ETLAI EnginePython · OpenAI SDK · DeepEval · RAGData & AgentsClickHouse via MCP · retrieval → anomaly → advisorInfrastructureSelf-Managed VPS · Docker · Database ReplicasRecognitionProduct of the Year, its category

Every carbon tool has dashboards. The problem was data architecture. Scope 1 and 2 data is clean; Scope 3 supply-chain data breaks everything: dozens of suppliers, dozens of formats, varying completeness.

I designed the data model for Scope 3 complexity from day one: NoSQL ingestion, normalised through ETL into a columnar store shaped for ISO and GDPR reporting.

On top sits the AI savings engine, a Data Retrieval Agent reading consumption from ClickHouse over MCP, an Anomaly Detection Agent surfacing spikes and waste, and a RAG-enriched Advisor turning raw data into personalised guidance. One product, SaaS + AI.

80% faster analytics queries

Zero manual monitoring — three agents, from data to advice

Product of the Year — won on data architecture, not features