10171422Gaia - an AI Tool for Nature-related financial disclosureClosedFast Start ResponseInnovate UKFinancial institutions must assess nature-related risks in financed companies (e.g. high water use, CO2 emissions). These risks can lead to material losses such as production halts, regulatory penalties, or pollution-related taxation. FiData is developing GAIA, a novel AI tool that enables financial institutions to assess nature-related financial risks---a growing requirement under UK SDR and EU Taxonomy regulations. At present, assessing these risks is time-consuming and expensive. It typically involves teams of analysts manually extracting and interpreting data from corporate sustainability reports. This leads to delays of up to two years and makes it difficult for asset managers and private equity firms to assess thousands of companies each year. GAIA transforms this process, delivering a fully automated solution that extracts and interprets sustainability data from unstructured corporate reports within minutes. The market is crowded with marginally fine-tuned LLM tools with minor improvements from foundation models. Most lack transparency on how outputs are generated and can't guarantee completeness---making them unreliable for high-stakes tasks. Unlike existing tools, GAIA does not merely extract metrics---it models nature-dependency and risk severity by combining retrieved data (by the LLM) with a Nature Dependency Matrix. GAIA will make risk assessments cheaper, faster, and more scalable, expanding access to nature-risk data across portfolios. The grant will support Fidata's Proof of Concept, demonstrating its performance and scalability.