[Kon-Peki]

If it was me, I would be ingesting the raw filings from SEC EDGAR and using the robust xml documentation to create very accurately annotated data tables that would be fed to my LLM