Following the post in #1111009377525186570 https://discord.com/channels/1108396290624213082/1421121173479096350 Currently PDF files can be added to the bot's knowledge pool as an indexed file during a user conversation instance. However in some contexts it would be preferable if we could read a PDF sent by a user to collect key data using something like an AI Task card. The important part is that this key data is considered personal information and as such should not be in the knowledge pool but a table row instead. This level of implementation is already possible with images using the Extract Content from Image card. The idea would be to do the same but for PDFs (OCR would be preferred as a company logo could be useful in certain contexts). This in effect would be quite useful for conversation bots that produce lead data, which a company would heavily depend on.