Last Edited: 7-17-2004

The 1997 CTI/IN Conference was held on October 21-22 in San Francisco. Sponsored by the IN Forum.

I attended to get a better feel for how the CTI industry relates to the IN industry. A good source of market research. As usual, I have some general comments and some specific items.

GENERAL OBSERVATIONS:

• Genesys hosted the meeting. About 20 people attended. Nortel sent the largest group, about 5 people.  The group seemed to be intimidated by Mircosoft who was not present. In general, Microsoft does not attend standards meetings. Their view is that their products automatically become the standard, so why go to a standards meeting. The CTI vendors view Microsoft as a 4 mph steam roller. You have to be going at least 5 mph or you will get rolled over. Therefore, one of the group's open questions is "What is the role of Microsoft?". Companies invited to this meeting include Bellcore, Lucent, SBC, Nortel, GTE, Genesys, Ameritech, DEC, Dialogic, Atlas Telecom, Sonetech, Excel, and Vergent. A number of the attendees are former BellSystem/AT&T/Lucent employees. The Excel rep. used to work at Lucent. The Vergent rep, Dave Bernstein, left AT&T a few months ago. Tony Roug from Dialogic was a long-term member of the IN project at Lucent.
• As the name implies, computer telephony is the integration of computers and telephony.
  From this meeting, the main services appeared to be:
  • Call Centers: Basically a Network ACD with Account Match service. Calls come into a company
    and get routed to an operator. Some form of Caller ID is used to do a database dip into
    a corporate database to retrieve info about the caller and show it as a "screen pop" on the
    operator's monitor. Network ACDs may also include outbound calling.
  • Personal Number Service (Single Number Reach)
  • 800 Service
  • Voice over the Internet (VoIP)
  • Integrated Messaging: Your browser is no longer just for e-mail. You can send/receive
    voice messages and FAXes as well.
  • Integrated Voice Response: Also called a "telephone tree".

A FEW TIDBITS I PICKED UP:

• Customers hate telephone trees (IVRs). Visual is better that audio for complex services.
• Voice Dialing is viewed as "low hanging fruit" by some. However, Ameritech says that it was a dud. There were 3 main problems:
  • Quality of voice recognition is bad.
  • Enrollment procedure is bad. It is hard to enter your "calling list", therefore, people don't use it.
  • It cost too much
• Single Number Reach has one of the same problems as Voice Dialing. People can't update the calling list.
• Given the problems with IVR, VAD, SNR,... human factors is BIG!!!
  • Passive services are great! Caller ID has 44% penetration in SBC.
  • A browser-type visual interface may be easier than audio.
• For internet voice, most vendors are looking at making circuit-to-packet gateways (a.k.a. Sync/Async Converters SACs) (a.k.a. Access Interfaces from the old fast packet days). About a $2B/yr market in a few years.
• The group felt that the issues dealing with a stable "talking path" from PSTN to IP will be worked out. However, signaling will be a big issue. For example, for a phone to call a PC, it will be hard for a phone to enter an IP address. Seamless call setup is the goal. Dave Bernstein of Vergent has a potential solution for this. He suggests treating the internet phone as a "cell phone" and using the HLR to get to it.
• The big 800-service customers (e.g., Citibank) have a mix of ACD/PBX/IVR/POTS lines.
• Some companies are looking to outsource everything: call center equipment & people.
• About 10-12 companies have call centers with 5000+ agents
• The real $$$ goal may be in increased network traffic (minutes) as opposed to charging directly for advanced services.
• As a policy, Ameritech does not want to open up triggers
• Additional applications may include SS7 on Windows NT, Mobility of data access, Wholesale to Gov't Agencies.
• Ericsson has the "phone doubler" to do internet call waiting. Others have products also.
• Trivia: California has 16 million lines.

Anthony Clark