AJAX isn't a technology, or a language, and there's no recipe to implement it; it's just a combination of various components to achieve something you otherwise couldn't: asynchronous http requests. However, since early 2005, when Google and Flickr popularized the concept, its use has grown rapidly.
AJAX uses XHTML for the data presentation of the view layer, DOM, short for Document Object Model, which dynamically manipulates the presentation, XML for data exchange, and XMLHttpRequest as the exchange engine that ties everything together.
Because of these requirements, AJAX works on I.E. 5.0+, Mozilla 1.0+, Firefox 1.0+, Netscape 7.0+, and Apple added it to Safari 1.2+.
Traditional HTML sends a request to the server, which processes it and either returns a static HTML page or dispatches the request to some scripting language such as ColdFusion, which creates and returns an HTML page for the browser to render. When this method has to retrieve new data from the server it has to repost and reload another HTML file. In many cases perhaps only a small portion of the returned HTML code varies and the shell itself remains the same resulting in huge overhead because the data has to be downloaded every time.
Some classic examples of applications that would benefit from AJAX are searching for information and displaying it back in a table, related select dropdowns, or checking if a user exists before submitting an entire form.
The below url contains a detail implementation of how it works.