Prizmo 2 can recognize business cards and differentiate between text, images and numbers. You can output your capture as JPEG, PNG, TIFF, PDF, RTF or plain text.
You can also export to Evernote, Dropbox, Google Drive and WebDAV services. In addition, the system recognizes 40 languages, can translate 23 languages on the fly and read documents aloud via text-to-speech.
I tested Prizmo 2 with PDF files, screen captures and my Epson XP-400 printer/scanner. The results were highly accurate and the OCR speed was very fast. The app took about 1.5 seconds to recognize the text in a single document. The text-to-speech was easy to understand. As it reads, the software highlights the words it is reading on-screen.
If you are a more heavy duty OCR user, there is a $24.99 in-app purchase that adds batch document processing, Automator actions and some custom export scripts. The basic version will be fine for most general users.
The only thing I would improve is the use of the non-intuitive File>Open command. Since I'm not much of a documentation reader, I scanned a file into a JPEG, and used File>Open to import it. That's not what you do. Instead, you choose File>New to import files. It was easy when I figured it out, but a bit non-intuitive.
Other than that, Prizmo 2 is powerful and reliable. If you need its features, I can recommend it without reservation. You can find a detailed feature list and video demos at the Prizmo website.
The company also has an iOS scanning app which I reviewed in 2010. It was also useful and reliable.