The software works a lot like Siri and Google Now, but it has some handy features like being able to create complex reminders and... singing. There are also some Xbox-like commands for being able to launch apps, and developers can use the API to let Cortana control in-app actions like queuing up Hulu videos. Cortana learns from how you use your phone, but you can use the Notebook feature to control what she knows.
Apparently, MS did bring in Jen Taylor to do the voice, but a synthesized voice is used for the majority of voice responses; otherwise, Taylor would have to record... a lot... of dialog, lol. Her voice is used to give Cortana [url=https://www.youtube.com/watch?feature=player_detailpage&v=6_YL5S2NwkE#t=69]some personality.[/url]
Also, you can have the assistant call you Master Chief.
[url=http://www.theverge.com/2014/4/2/5570866/cortana-windows-phone-8-1-digital-assistant]Article.[/url]
It's important to note that the software is still in Beta (as Siri was when first introduced) so we can expect to see some improvements down the road.
-
The synthesized voice still sounds strongly to me like Jen Taylor. I'd imagine they did the synthesis they same way they did Microsoft Sam and how Apple did Siri -- they just asked Jen to record all of the phonetic sounds that occur in English, and then the computer strings them together in the appropriate order to form a word we recognize. I'm sure that for some of the more common responses (such as joke questions/answers or common dialogue like "Okay I've added that to your calendar" or whatever), they asked Jen to record entire sentences. But it uses a lot less storage space to just have recordings of the phonemes -- even though the synthesizing process removes a lot of the emotion and inflection in normal human language and makes her voice sound like a computer.