ESG Greenwashing Detector is an AI-augmented decision support platform designed for ESG analysts. It helps identify potential greenwashing signals in corporate sustainability reports through explainable, evidence-based analysis.
Every signal is traceable to specific evidence in the source documents. No black-box scoring.
Nuanced risk categories instead of simple pass/fail judgments. Context matters.
All claims backed by excerpts with page references. Full auditability.
Supports and augments professional judgment rather than replacing it.
Extracts environmental sustainability claims from ESG reports. Identifies climate-related statements including emissions, net zero targets, energy transition, and renewables. Classifies claims by type (quantified targets, intensity metrics, reduction commitments, qualitative pledges) and extracts attributes like baseline year, target year, and scope coverage.
Assesses the substantiation strength for each extracted claim. Detects numerical disclosures, Scope 1/2/3 reporting, historical performance data, third-party verification, and methodology transparency. Flags gaps such as missing baselines, non-quantified targets, absent verification, and selective scope disclosure.
Compares disclosures across reporting years when multiple reports are uploaded. Uses vector embeddings for claim similarity and clustering. Detects target changes, baseline resets, metric disappearance, and scope boundary shifts to identify temporal inconsistencies.
Combines all detected risk signals into a structured profile. Evaluates four categories: vagueness risk, omission risk, substantiation risk, and temporal inconsistency risk. Applies weighted scoring with rule-based escalation and confidence calibration.
Generates analyst-ready narrative outputs explaining the risk assessment. Includes risk rationale, evidence citations, detected signals, assumptions made, and uncertainty communication. Produces plain-language explanations with confidence levels.