"Ajax" transfers the data over HTTP, just like regular web pages. So, just use whatever you use to scrape data from regular web pages. You will need to treat the results a bit differently - if JavaScript is returned, you will need to interpret it from your script.

Alternatively, you can try to automate the website from the outside, by using, for example, Win32::IE::Mechanize, and then capturing the traffic using Sniffer::HTTP. Where exactly do you have problems?