It's understandable that virtually all the headlines surrounding VoIP involve services aimed at consumers or business end users.
There is another dimension to VoIP, however, that deserves attention.
In most cases, enterprise VoIP traffic has to be translated to traditional time division multiplexed (TDM) status to be sent between enterprise locations, to other ISPs and to devices on traditional phone networks. This involves a confusing (to us, at least) array of gateways and other hardware and software elements. It also involves use of the public switched telephone network (PSTN) which, the last time we checked, was VoIP's competition.
It would be a lot simpler, cheaper and efficient if the traffic could stay in its original IP form. It also would enable better service, since not changing from IP to TDM makes it easier to maintain a standardized level of features throughout the system. Finally, it would put the fate of Internet telephone service more fully in VoIP providers' hands.
There always will be crossover from the PSTN to VoIP networks, since many calls that originate on one terminate on the other. The true economies of scale will kick in if a method is found to bypass the PSTN completely on VoIP-to-VoIP calls.
Expect to see more VoIP equipment, standards-setting and service provider confederations aimed at cutting the PSTN out of as much of the loop as possible. This trend will be energized by the emergence of the Session Initiation Protocol (SIP) as the predominant signal protocol.
Glitzy category-building, customer-facing services clearly were the first step. That will continue. The focus -- at least for network designers and telecommunications planners -- will shift to VoIP networking.