I have been working on several voice projects in the past with ACD features, mostly based on TDM technology. It's all commercial stuff, but I have the experience and I am willing to share that. If anyone wishes to start such a development I'm sure I can dig up a functional model and help with the design.