Closed document set analysis

I’m a psychologist but I work with lawyers. We all deal with folders of documents usually pdfs and often want to ask questions of closed document sets (i.e. those relating to a specific case). We really need a natural language processor to be able to extract relevant information according to a long list of relatively set prompts. Ideally then we would construct sections for reports and other documents we are producing based on our source documents. There are a few critical issues. One is the whole thing of having to upload documents when we already have them on cloud systems. Another related one is data security. The other one, but this is more of a hope for the future, is being able to train models easily on existing reports so that things like interview transcripts can be turned into structured sections.
There seem to be a few products that can do some of this but they often have limitations or security concerns. I’m not a developer obvs but have some understanding of what’s involved. Is this something that is ‘just around the corner’ or is there some more fundamental barrier?

With respect to “data security” you’ll need to specify que a bit more about what your exact security requirements are to give you any kind of answer on that front.

As far as training models on existing reports goes, you can do that but not with ChatGPT.

It’s not all bad news though, with respect to needing to upload documents which are already on the cloud you’re in luck—as long as you use Google Drive or OneDrive. They are currently testing Connected Apps and may be rolling the feature out to more people soon.

Thanks, that’s interesting. Re data security, it’s perhaps these issues: MFA for a login account; where the files are being uploaded to and saved; whether another organisation has actual or possible access to the files, chat prompts or responses. MS Copilot for Office 365 seems to have these issues addressed but doesn’t do what we want it to do i.e. you can’t do generative chat in relation to a set of documents (e.g. PDFs in a single folder).