For the last half decade or so I have generally only worked with one client. Early on when we were making the first decisions around what eLearning was going to look like at their organization, it was decided that they would use text-to-speech. The rationale was that if six month or a year went by, they could have anyone update the course material rather than hire back an expensive voice actor; makes sense.
Fast forward five years and I'm still the main person creating eLearning for this organization and I could have easily supplied them with my voice for no additional cost to them. In every update that has been done to my courses in that time, only one was redesigned and updated by another designer.
Take a look at this video to see me compare the two methods of creating narration for Adobe Captivate.
One complaint that I have had in the past is that it was difficult to get the studio quality that clients would expect. One way around that is to use Adobe Audition to clean up your audio recordings and remove any undesirable noise or pops. Check out this video to see me apply the noise reduction to another human recording