When it comes to setting up a script for a customer, when and how do the recordings get made ?
Personally I have no problem using my voice, I was a DJ for 20 years and been in telecoms for nearly 40 years the first 15 years being a service engineer for telephone answering machines. So I have got used to listening to myself.
But .... judging by the response I get from engineers on my training courses, when I get them to write a script which includes a recording. 9 times out of 10 they opt out in favour of the reliable standby," beep.wav"
So the solution is fairly easy, use an online text2speech web site (a lot of these have charges to download the result, but as engineers I doubt its beyond anyones ability to record from the screen.... not that I would condone such deceptive acts). For the financially challenged, which I guess means anyone not in sales, there are some free options like google translate that do a reasonable job.
The advantage of being an Apple User, thing are a little easier, for my autovoicemail project that has multi-language voice mail messages that change dependant on status text, presence and outlook, I used Apples built in text to speech engine to create over 100 wav files in less time than it took me to write this blog. (You can find a video demo of tha project on my youtube channel http://bit.ly/2ouIytz )
For fellow mac users I attach a small app, written using automator, to let you save a WAV file to the desktop from any text typed into the dialog. A choice of female or male is also available. Modifing the script to handle other languages should be a simple as changing the voice names in the first bloct of the script.