How would you go about converting a book to an audio version

I am very fond of audio books: I listen to them while I exercise and when I take long drives.

I have a fairly large, prioritized list of books in text form that I'd like - at some point - to "read", or rather listen to.

More often than not, the audio version does not exist anywhere.

Modern deep learning based text to speech engines are really good for paragraph-sized text, but fail miserably on long form text.

So, question: has anyone figured out how to automate the conversion of a traditional book in text or PDF form to a listenable audiobook of somewhat acceptable quality.

More specifically: are there ML/AI based tools out there to segment a text or PDF files into chunk of a size that can be digested by modern TTS tools?