Why does a big company like CR:USA not have any data beyond the last 12 months or so? Arghh, Factset.
For this reason, and the inability to download all raw data I’d suggest the following.
I think the most valuable thing P123 could do for its users at this point is not fancy AI tools, but rather the construction of its own proprietary database and data feed directly from filings at EDGAR and SEDAR and the equivalents in Europe and Japan. While this is an enormous endeavor (so is AI), I think the capitalized value of such a feed must be similarly enormous if Compustat and Factset and Reuters can charge 10-30k for a single license. Is this not worth hiring 100 people to accomplish?
P123’s largest COGS is probably data, and data providers will always raise prices to gobble up P123s margins, so this just makes sense for the business. P123 has the know how to make their own database and feed.
Additionally, I think AI people just want a lot of data to wrangle with their programming skills, not a dummed down semi programming AI interface to use on un-downloadable data.