Novata

Their 'collect once, report many' value prop worked beautifully when manual data entry was the bottleneck — but Claude can now extract ESG metrics from annual reports, sustainability disclosures, and public filings without asking portfolio companies to fill out a single questionnaire. The platform becomes a very expensive spreadsheet when the data it organizes arrives pre-organized.

ESG analysts who spend their days mapping questionnaire responses to TCFD frameworks and calculating carbon footprints are staring down ChatGPT's newest function calling capabilities — it reads the same documents, applies the same methodologies, and never needs clarification on Scope 3 emissions boundaries.

The 400+ client relationships and three years of workflow integration create genuine switching costs, and the Ford Foundation/Omidyar backing signals patient capital that won't panic at the first revenue dip. But when the core workflow is 'help humans do what AI now does natively,' even sticky customers eventually ask why they're paying for the middleware.

Their 'Novata AI Assist' and 'AI Emissions Calculator' suggest they see the writing on the wall, but these feel like features bolted onto a data collection platform rather than a fundamental rethink of what happens when data collection becomes free. They're automating the forms instead of asking whether forms still make sense.

ESG reporting survives — it's increasingly mandatory. The question AI eliminates is 'how do we gather this data from portfolio companies?' When regulators require TCFD disclosures and LLMs can generate them from public filings, the expensive surveying apparatus becomes the expensive part nobody needs.