Speech Automation's False Choices

As a growing number of enterprises make their conversion to VoIP (Voice over IP) or IP telephony, the notion of automated speech is getting short shrift. Even though most of the major moves from traditional switched networks (also called 'TDM' by telecom aficionados) to IP often coincides with deployment of new solutions that marry voice portals or speech-enabled contact centers to corporate WANs, you never hear anyone talking about SoIP (Speech over IP).

Whether you talk to Alcatel, Avaya, Cisco, Microsoft, IBM, Oracle, Siemens or any other major enterprise infrastructure provider, it's clear that IP deployments are accelerating. By some measures, the number of IP-PBX 'ports' shipped in 2005 will exceed new shipments of traditional telephone switches. While the architecture has changed, the objectives remain remarkably constant. Procurement decisions and vendor selection are still ROI driven. That's unavoidable. The bigger the ROI and the shorter the payback period, the better!

The 'bigger is better' frame of mind has put contact centers at the forefront of both speech enablement and IP telephony conversion. That's because decisionmakers have used contact centers as showcases for cost savings from the early days of IVR. Expecting the past to be prologue, they continue to regard contact centers as the place where they will see the biggest bang for the IP buck.

At an analysts' meeting convened by Avaya, Mark O'Gara, senior vice president of internal computing and operations at AOL was asked the following question: "If you could turn back the clock, would you pursue your original strategy (IP-enabling the contact center) or would you address enterprise networking first?"

O'Gara has a high profile among both the IP telephony and speech processing crowds. Over a two-year period he converted his global empire of 14 contact centers to IP telephony. At the same time, he speech-enabled a customer care application on top of his IP contact center platform to derive $20 million in annual savings and calculated an ROI of 150 percent in the first year.

Given his high-profile success, his answer was surprising. With the benefit of hindsight, he said he would have preferred to IP-enable the enterprise first. That way, he could have gained experience with the new technology and get his IT and telecom teams trained and aligned with the new architecture in advance of moving to a more mission critical environment.

Deciding to IP-enable the contact center was a life-changing moment. As he recalls, when management decided to convert his customer care contact centers to IP telephony in 2001 his verbatim was, "You have to be (choose your adjective) nuts!"

From today's perspective, O'Gara understands why management would place its initial focus on the contact center: it is a showcase and is the area where you have the chance to generate the largest amount of savings and ROI potential. However, that sort of visibility comes with high-profile risk which must also be taken into account when calculating recognizable and quantifiable return.

Contact centers are highly complex, high-visibility venues and upgrading them on the fly is often a high-risk endeavor. We are learning that there are many features and functions that are embedded in the contact center that are of great relevance to support 'self-service' throughout the enterprise.

In the relatively near term, it will become more and more apparent that "platform versus app" decisions are predicated on answering the wrong question.

An Avaya executive characterized O'Gara's decision as making the choice between upgrading 'the platform' (meaning the entire enterprise communications and IT infrastructure) or 'an application' (referring to self-service in a contact center, sales support, or dispatch of the field service force). The point is: it's the wrong question to answer. In the context of broadband IP telephony conforming to a service-oriented architecture, the platform is the application and vice versa.

Not to sound too McLuhan-esque, but "the solution is the platform." The people who buy or influence the purchase of speech processing, call processing and application development environments want to see solutions that leverage their existing infrastructure. These solutions, to real problems in the enterprise and contact centers alike, create tantalizing ROI. Without the solution, the platform is just a set of inert technology components - Web servers, voice portals, media servers and application development tools.

Applications and platforms bring cost. Solutions bring value. Solutions that bring value with reduced cost because of common and shared architecture to house their applications are the power of SOA and IP telephony.

Dan Miller is a senior analyst for Opus Research. He founded Opus Research  and published Telemedia News & Views, a monthly newsletter regarding developments in processing and intelligent network services. He can be reached at dmiller@opusresearch.net .

SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues