From (Political) Show Business to 'Go' Business
The voice automation industry in the US can be generally situated in a pretty wide cross-section of political landscapes: you've got the Dallas metroplex (Bush country), California (Schwarzenegger's domain), and the Northeast Corridor (Virginia through DC up to liberal New York and Massachusetts). All of which is to say that the readers of this column, if they were to go see the latest Michael Moore film "Fahrenheit 9/11," might find varying numbers of places to see it and people in the theater.
If you do go see it, you might be struck, as I was, by how much politics really consists of show business - the makeup, the waiting for the camera to roll, the entourage of press watching the President tee up on the golf course, cut wood on the ranch, hob-nob with Bin-Ladens, etc. Don't even ask about Wolfowitz prepping for the camera.
These days I find a lot of clients asking us variations on "when do the cameras roll?" for speech applications. The upgrade to VoiceXML is done, the open standards makeup has been applied, the packaged apps are all packaged up and the Webinars are scheduled.
So, where are all the customers? The analogy to politics may serve us a little more (but not much), if we think in terms of the customer as a voter base. What is the equivalent IT reality to the political calendar, or 'terms of office' that may drive adoption of speech technologies? First towards the issue of cyclical 'terms' of IT infrastructure. As most of the world knows, the preponderance of IVR equipment installed since 1990, with an installed base of something like 300,000 systems and an annual revenue rate of $1.7 billion for everything from automated attendant to voicemail systems, is obsolete. Legacy IVR vendors have done the make-over, as we alluded to above, and are now shipping replacement designs that break apart the proprietary monoliths of yesteryear. Next-generation challengers like Envox, Holly, Paraxip, Vocalocity and of course VoiceGenie have climbed into the ring.
Customers with multiple systems have a growing problem in managing their 'hardware' problem. This end-of-life problem becomes more acute when you get more, different proprietary hardware as the result of a merger or acquisition. So, on top of the end-of-life issue there is Precipitating Event #1: M&A activity. We have talked to enterprise customers who have had critical speech automation decisions thrust upon them as a result of having to accommodate multiple platforms as the result of an acquisition.
A second, larger, transformation for enterprise IT decision-makers is the evolution from an application-oriented architecture to a service-oriented architecture (SOA). This is a hot trend these days, and lays the groundwork for 'utility' computing that will in turn enable a Web services environment and the kind of federation that makes grid computing and collaborative commerce the fun stuff that CIO's respond to. In this transition, the old vertical integration of application, server and database is delaminated into a horizontal integration where multiple applications ride a common 'service delivery' middleware layer, that delivers utilities such as data connectivity, authentication, clustering, etc.
A lot of integration work is required - some estimates put it at 40 percent of all IT spending. How does speech get to ride this cycle? Doing the homework on the prospect to understand where they are in the SOA continuum, or in plain English, what kinds of integration projects are underway, gets us to Precipitating Event #2. Keywords here are middleware, enterprise application integration and service delivery.
Finally, the verticals of financial services, utilities and insurance are getting febrile about outbound notifications. Whether it's fraud prevention, accounts payables and collections, medication shipments or claim status updates, the customer impact of event-driven communications—sometimes called intelligent notifications—is very high. Business processes are dramatically truncated by proactive communications; basically, why wait for the customer to call, why not call the customer? This is a speech opportunity when the execution is to a phone number and requires interactivity. It is driven by business processes and customer events.
Part of the notifications opportunity involves emergencies. As I mentioned at a SpeechTEK analyst panel a year ago, business continuity in a post-9/11 world is thriving, and speech-enabled phone trees for first-responder teams is part of this intelligent notifications opportunity. Thus, in a world that tends to think inbound, outbound notifications constitutes Precipitating Event #3. Here we find the essence of event-driven market activity: as enterprise IT gets its SOA act in gear, it is finding more and more applications that need to go out (or in) over the phone at some point. Other legs of the application may involve online, email, or retail, but multi-channel intelligent notifications must include phone when they get to a certain scale.
Think of these Precipitating Events as issues that are core to the enterprise IT decision-makers who are your constituency. Campaign on these issues, understand that speech must support the central themes that IT is focused on: integration and service-oriented architectures for the sense-and-respond organization. Stop waiting for showtime, it's go time.
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.