GreenProof by Brain.exe

Inspiration

Sustainability reports often mix facts with marketing language, making it hard to tell which claims are true. We wanted to fight greenwashing by creating a platform that verifies ESG claims against real-world data

What it does

Extracts and classify claims as general, quantitative, or future targets Summarizes which claims are consistent, inconsistent, or unsupported Tracks live data via simulated IoT sensors Calculates a trust score comparing claims with reality Simulates industry parameters to estimate emissions, energy use, and costs Maintains a verification log for historical tracking

How we built it

We combined React + Tailwind for an interactive frontend, Node.js + Python for backend and AI, and simulated IoT data + ML models to verify ESG claims in real time.

Challenges we ran into

Parsing complex PDF layouts Ambiguous claim classification (general vs future target) Simulating realistic IoT data patterns Designing a fair, interpretable trust score

Accomplishments that we're proud of

Converting unstructured sustainability reports into structured, actionable insights Implementing live verification and trust scoring Creating a simulation engine that helps assess future target feasibility

What we learned

Language can mislead — data verification is crucial AI extraction needs evidence-based guardrails Transparent scoring builds trust and accountability

What's next for GreenProof by Brain.exe

Integrate real IoT sensor data from partner companies Expand ML models for predictive ESG risk assessment Provide regulators and investors a dashboard to audit claims in real time

Built With

• express.js
• numpy
• pandas
• python
• react
• scikit
• tailwind
• vite