Recently we received a request to create an AI Agentic solution to automate Methane emissions reporting for ESG & CSR compliance. Based on our Umbrella AI agentic platform we forked a dedicated functionality, which was customised into independent AI Agentic solution for methane emissions reporting.
OGMP 2.0 and EPA Subpart W are both frameworks for reporting Methane emissions, but with different focuses:
OGMP 2.0 (Oil and Gas Methane Partnership 2.0) – Voluntary international framework led by UNEP.
- Goal: To drive deep methane reductions across the oil and gas sector.
- Methodology: Requires shifting from generic emission factors to direct, site-specific measurements.
- Reporting Levels: Has five levels, with “Gold Standard” being the highest, which requires comprehensive, measurement-based, source-level data.
EPA Subpart W – Mandatory U.S. regulation under the Greenhouse Gas Reporting Program (GHGRP).
- Goal: To require reporting of greenhouse gas emissions from large oil and gas facilities.
- Methodology: Facilities must collect data, calculate emissions, and follow specific reporting procedures.
- Reporting Threshold: Applies to facilities emitting 25 000+ metric tons of 𝐶𝑂2 equivalent per year.
Our AI Agentic solution Methane emissions reporting for ESG & CSR compliance
- OGMP 2.0 and US EPA Subpart W reports are only quantitative (clients data integration required or/and documents).
- US EPA Subpart W has over 600 critical validations checks and 10 segments, its reporting format is XML, OGMP 2.0 format is MS Excel.
- Gen AI can find data points from available corporate sources given as Excel spreadsheet or database or anything else client has.
- Gen AI can conduct a gap analysis between available data points and what is required by reporting standard.
- Gen AI is by default non-deterministic, that’s why is needed to develop a separate app in a combination with existing AI Compliance Reporting tool. This app will produce XML, Excel reports, that will allow to “paste” specific data precisely into specific cells.
- The project required up 3 months of full time work.