Is this possible using visual intelligence? - Point your camera at a receipt and Books automatically reads the vendor, amount, date and all the line items

Not directly from what’s shown here.

What is shown is that Visual Intelligence can recognize things like:

• bills / receipts-like scenarios for actions such as splitting a tab
• calendar events
• objects, places, and screen content

And separately, the Vision framework has a newer API, Recognize Documents Request, that can extract structured information from documents like:

• tables
• lists
• text
• barcodes
• emails, phone numbers, URLs

So if your goal is:

point the camera at a receipt and have Books automatically read the vendor, amount, date, and all line items

that sounds more like Vision’s document recognition capabilities than Visual Intelligence as described here.

I don’t see anything in the provided WWDC context that says Books specifically does this automatically with Visual Intelligence.