Hi Henning - Thanks for the question.  Infobase works for .doc, .txt. and .pdf files only at the moment with a 50MB file limit, but the setup would be a bit more complex than you’re probably thinking, as you’d have to use the new “Search Infobase” action that returns relevant snippets of the material, but you’d need to add additional processing steps to then summarize the findings in order to answer the question.  Depending on the nature of the material, you may need to load the full document into memory in order to gain the full context which could bump up against context window limits of the various LLMs.  While doable, it would be a bit tricky and have specific limitations.  Using our new Tables feature would probably be better suited to this task although it is not document based and is, as of now, enterprise only.  Regardless, the use case you’re describing is probably better suited for Glean.com, in this particular case.  Hope that helps!
Chris

Thanks for the reply, Chris - I have booked a meeting with Glean.com

I have talked to Glean.com, and they do not think that their solution is the right fit for my project. They will send me a list with other and better suited solutions

That’s odd because that’s essentially how we used them at my previous Company.  Either way, glad they have alternative suggestions.