This is just a guess on my part, but I believe the reason the Echo cannot act as a speaker phone is that Amazon has avoided giving skills any direct access to the microphones. Instead, skills only have access to the textual component from the mics, after Amazon performs the speech-to-text translation. I suspect such access has been withheld out of security concerns: They simply don't want skills to be able to surreptitiously listen in on what the Echo is hearing. But, until such an interface is forthcoming, it would be very difficult to produce a speaker phone app.