GreenPromise

Inspiration

Every business wants to be more sustainable, but most struggle with the first step: understanding their impact. GreenPromise solves that. Companies get fast, automated sustainability analysis, giving them the insights needed to reduce emissions and make smarter environmental decisions. GreenPromise empowers businesses to understand where they stand and how to do better, while doing so in an equitable way.

What it does

This website allows businesses to upload their financial and operational data — from CSV spreadsheets to invoices, PDFs, and even photos of receipts. Then the website parses their files, maps expenses to emission categories (energy, transportation, supply chain, waste, and other), and calculates their carbon impact. It presents this to businesses by showcasing their expenses in terms of areas that negatively impact their carbon footprint, comparing their emissions to those of other publicly available businesses, and offering suggestions and a timeline for when they could be completed. There are additional pages dedicated to explaining our purpose and introducing us to our users; as well as creating and using an account, allowing them to save their uploads and their outcomes, to reference for future use.

How we built it

We built it using the Anthropic API to analyze a variety of provided datasets (.CSV, .PDF, .JPG, .PNG). We also integrated a multi-format ingestion pipeline, where different formats get parsed differently but merge into one unified analysis. We used Claude Vision API to parse through images and extract structured expense data from them. Then we made it categorize emissions across key business activities and generated a breakdown of your footprint. Results are presented with clear insights and recommendations for improvement.

Challenges we ran into

Businesses don’t store their information in a single clean format, they upload CSV exports, scanned invoices, screenshots of receipts, partially filled PDFs, and more. Creating a multi-format pipeline that could reliably extract structured data across all these formats required careful parsing logic and validation layers.

Accomplishments that we're proud of

We’re especially proud of building a unified analysis engine that works seamlessly across CSVs, PDFs, and images, merging all inputs into one coherent emissions profile. But most importantly, we transformed sustainability reporting from something intimidating and expensive into something accessible and immediate.

What we learned

We learned that sustainability is as much a data problem as it is an environmental one. Many businesses want to reduce their footprint, but they just don’t know where to start. We also learned that clarity drives action. When businesses can see their carbon impact broken down into tangible categories, it becomes easier to prioritize changes. Technically, we learned the importance of building flexible ingestion systems. Real-world data is messy, and resilience in parsing and classification is essential for trustworthy outputs. Finally, we learned that automation can democratize sustainability. Tools that were once limited to consultants and specialized ESG teams can now be made accessible to companies of all sizes.