Green0meter

Problem: Small and medium companies, comprising 99% of all businesses and employing two-thirds of the workforce, possess significant potential to mitigate CO2 emissions. Unfortunately, the majority of attention and funding is directed towards large corporations, leaving smaller entities without the necessary resources and expertise to curb their emissions effectively.

Solution: Green0meter is a B2B SaaS platform that calculates companies' carbon footprint and then with the power of AI suggests actionable and personalised recommendations on how to reduce the carbon footprint. It uses Azure ML models, lists of recommendations, industry averages per NACE code. We reduce the number of calls/compute power to only when we have all the necessary inputs. We also include the compute power-related emissions to the company's overall emissions in order to have a full-blown impact.

Impact: Green0meter offers a comprehensive solution that not only informs companies about their primary carbon footprint impact but, more significantly, provides personalised and actionable guidance on how to reduce it. Through this approach, Green0meter effectively minimises the overall carbon footprint, while also offsetting the substantial environmental impact of AI technology.

Feasibility: Green0meter strives to be highly accessible and affordable for small and medium-sized companies, ensuring its widespread impact. The primary challenge lies in efficiently gathering data from businesses and continuously enhancing the AI model to provide highly personalised recommendations. By overcoming these hurdles, we can collectively drive change towards a more sustainable future. Join us in making a positive difference together!

What you built: Set up MS Azure cloud instance, created MS SQL Tables with recommendations, we created Power Automate flows using APIs to OpenAI, we trained AI PowerBuilder models to recognize and read energy invoices. https://colab.research.google.com/drive/1adGKZiwkkFVPd6MxUpwafzXr3N_C_YHP?usp=sharing

What you had before: We had the FE platform written in React.

What comes next: We need to keep improving the list of recommendations, train the model with additional data set (past recommendation), extend the CO2 calculation with third party data (satellite imagery, Statistical Office data).

Video: https://youtu.be/O4CmVb5kJUY