2005: Yet Another Year of Intelligent Migration
OPUS Research is a long-standing and unrepentant booster of speech-enablement. We stand for judicious use of advanced speech recognition (ASR), text-to-speech conversion (TTS) and all the attendant hardware, software, professional services and tooling (soaking up something on the order of $28 billion in 2005) directed at a single goal. That is to take existing self-service infrastructure - on a Web site, in front of a corporate directory 'auto-attendant,' baked into a contact center or hosted environment - and "make it talk."
"Make it talk" should not be confused with efforts to "make speech mainstream." If the years between 1999 and 2005 have taught us anything it is that speech is seldom, if ever, the mainstream. Instead, it is better a tributary; one of many paths into the frequently used self-service channels. The year 2005 is emphatically not going to be The Year of Speech Enablement. Rather, it is Yet Another Year of Intelligent Migrations (YAYIM): from contact center to IT-driven customer service, from siloed to integrated solutions, and from switched networks to IP-telephony.
Always at the Threshold
Since the dawn of this millennium, some other IT project has been perched ahead of speech recognition in the backlog of mission critical IT projects. Y2K conformance was the must-do change in enterprise IT in 1999. Then came the World Wide Web and a series of e-commerce mandates to suck the air out of the prospects of adding speech recognition to self-service resources. The bursting of the dot-com bubble came next. Thus we entered 2004 with great hope for conversational access technologies to accelerate and improve upon ways for people to carry out their interaction with agents, IVRs, auto-attendants or hosted server farms that comprise the underpinnings of phone-based commerce.
The technologies work better than ever. What's missing is a clear set of guidelines - a migration path - from what-is-now to what-should-be. In the ensuing 12 months, the IT staff is going to have its hands full with all sorts of fire drills. "Compliance" and "security" issues reign supreme with IT these days. In the U.S., compliance to the security and disclosure strictures of Sarbanes-Oxley has IT managers rushing to install new utilities to beef up the ability to control and monitor access to enterprise data. IT analysts see the expense of SOx compliance exceeding that of Y2K hardening in terms of dollars spent. That alone could set the cause of speech enablement back another 12 months. The silver lining is a boomlet expected for deployment of speech-enabled automated systems for PIN discovery and reset.
Enter VoIP+Speech
The inexorable migration to Voice over IP (VoIP) conversion is also a major mandate, especially among multi-national, multi-site enterprises. The efforts of Cisco, Avaya, Alcatel and the horde of telecom infrastructure vendors have accelerated enterprise deployment of IP-based networks for traditional voice conversations, as well as impromptu conferencing and a few other IP-enabled enhanced services. VoIP conversion does not, in-and-of-itself; bring ASR or TTS into play. Nonetheless, the definition and deployment of 'VoIP+Speech' solutions will bring tens of millions of dollars to vendors or teams of vendors that demonstrate how an enterprise can use automated speech in conjunction with IP-telephony to leverage components of self-service infrastructure that are already bought-and-paid-for.
Common ground is defined by company-wide mandates to hit well-specified business objectives. Significant technology 'wins' used to come at the expense of 'seats' in corporate contact centers, in the name of saving money on customer service. Then 2003 marked the sea of change from cost savings to revenue enhancement, and 2004 added the idea of using technologies to promote customer retention and upselling. The underlying mandate has been to "use what we already have" and transcend traditional divisions between and among departments.
In 2005, solutions that involve speech will address the immediate need to break down silos on a call-by-call basis, in real time. Improvement does not start with "an IVR upgrade or replacement" to support speech recognition. It starts with mapping VoIP+Speech to business objectives. Solutions providers who speak to top management about return on existing assets and have the tools to map solutions to stated business objectives and enterprise workflows will be the winners; and, when speech is deployed as one of many channels into self-service resources that are designed to enhance the user experience, the caller is the real winner.
Mark Plakias is principal researcher at OPUS Research. You can reach him at mplakias@opusresearch.net , or better yet, connect to the resources available at www.opusresearch.net.