Jonathan Taylor, President and CEO, Voxeo

Voxeo focuses on providing standards-based IVR platforms and hosting. Voxeo helps enterprises automate and connect their most common phone calls with its Interactive Voice Response (IVR) or Voice over IP (VOIP) solutions. Its IVR and VOIP hosting, turnkey platforms, and developer services are powered by VoiceXML and CCXML.


Speech Technology Magazine sat down with Jonathan Taylor, president and CEO of Voxeo, to talk about Voxeo's plans to revolutionize speech technologies and how these plans became possible.


Q. Welcome to SpeechTEK West, Jonathan.  I understand that you are unveiling several announcements at the show.  What are these new initiatives?

A. Thank you.  Voxeo is announcing three things at the show:  First, our new Prophecy voice platform. Second, two new key partnerships with Vicorp and Apptera. And third, the fact that we've just closed our 24th consecutive month of profitability.


Q. As your "biggest" announcement, please explain a little more about what "Prophecy" is and why it is different from any other offerings out there today.

A. Prophecy is our newest VoiceXML and CCXML speech development and delivery platform.  Simply put, it differs from existing speech and IVR platforms in four key ways: affordability, simplicity, scalability, and accessibility.


You see, every year we survey our customers and partners and we ask them - What could we do to help your business the most?  How could we make our products and services more valuable to you?  And every year for the last three years, the vast majority of our customers and partners have said the same thing:  make ASR and TTS less complex and less expensive.


This comes as no surprise, of course.  Nearly every analyst, advocate, and enterprise in the space has been reporting or asking for these changes; they all know that cost and complexity have held our industry back immensely.   What has surprised us, however, is that the leading speech engine vendors have barely reduced their prices.  In fact, in many areas prices have gone up.  So four years ago, we decided to do something about it.  We decided to build our own ASR and TTS engines.  Now that they are complete, we're bundling them for free with our VoiceXML and CCXML platforms.   Prophecy is the bundled result of our efforts.


We started Prophecy four years ago by doing the same thing Nuance and SpeechWorks did at their start - we licensed best-of-breed, core speech technology from some of the leading speech research organizations.  We then spent 40 man-years building a complete product around that core.  We developed our own advanced speech signal processing technology.   We extensively optimized and tuned our own speech models and phonetic dictionaries.  We built a highly scalable and reliable MRCP server.  We built grammar engines to seamlessly support any speech grammar format - including GRXML, ABNF, BNF, JSGF, Nuance GSL, and ScanSoft OSR.  We then tested these engines in our hosted platform for over a year.  They're now ready to ship.


Don't get me wrong - we're not claiming that these new engines can do everything the Nuance or ScanSoft engines can do.  Prophecy delivers extremely high-quality speech engines, with a focus on the features that 80 percent-plus of speech deployments require.  For example, Prophecy does not include support for grammars with more than 50,000 words or for biometric speaker verification.


We will continue to promote, sell, and support Nuance and other engines with our platform, to the customers who require and can afford them.  Our new engines - The Prophecy Speech Recognizer and Prophecy Speech Synthesizer - are designed to address a market Nuance and others do not: companies that want to use speech, but who cannot budget or justify the traditionally high cost.


Q. What impact will this new product have on your Fusion product line and your customers who have deployed this solution?

A. I mentioned that simplicity was one of our four goals with Prophecy.  Many of our customers who purchased turnkey VoiceCenter Fusion servers did so because it is an extremely simple, ready-to-go, pre-configured VoiceXML and CCXML platform. Prophecy brings that level of simplicity to a pure-software, downloadable product.  Prophecy is available as a single 77 MB download, with VoiceXML, CCXML, call conferencing, call recording, SIP VoIP telephony including a SIP softphone, and the speech recognizer and synthesizer engines built in.  There is absolutely no configuration or other third party software required.


Customers who previously would have bought our simple turnkey Fusion boxes can now just download the Prophecy software and deploy it on any Windows machine, with Linux and Mac support coming later this year.  They can run up to 24 ports on a low-end PC or notebook computer and up to 96 ports on a single high-end server.  If you put two or more servers running Prophecy on the same LAN, they auto-discover each other and work together to share call load.  That capability addresses another Prophecy goal - scalability.  For example, we've already deployed a single-customer system of over 2,000 ports using Prophecy's auto-discovery and clustering technology.


Our Fusion servers will still be desired by many customers, as they also bundle analog or digital PSTN and PBX interface cards, a PSTN/PBX to SIP VoIP gateway, and a reliable dual-CPU server pre-installed with the Prophecy software. 


All of our customers with active support agreements - including those who have previously bought our premise VoiceCenter software, our turnkey VoiceCenter Fusion servers, or the Vocomo platform, which we acquired last year -- will receive a free upgrade to VoiceCenter 6.5 that includes the Prophecy engines.  



Q. What impact do you think this announcement will have on the speech technologies' market?

A. We're hoping that the impact will be significant.   Cost and complexity are major barriers for enterprises looking to deploy speech, and by removing those barriers we hope to enable exponential growth in speech deployments.   Remember, our industry still sells to a small minority of the enterprise IT market.  Most of the market has been waiting for a breakthrough in speech quality and pricing.  We think Prophecy could be that breakthrough.


I think we've seen a relevant example in the Web.  Ten years ago, the Web was not widespread.  Today, it's pervasive.   In comparison, speech technology has had relatively stagnant growth.  The Web took off in large part because it was extremely inexpensive to build and deploy applications for, and because the software you needed to do so was extremely accessible.  Then and now, a developer could go to the Web site of almost any Web technology vendor and download their products for free - without being screened by a salesperson, without the burden of up-front payment, and without signing an NDA.  Again in comparison, speech has remained an exclusive club.  Access is screened by sales, cash-in-advance is standard, and NDAs are required.  Those requirements restrict growth.


Open accessibility is another tenet of Prophecy.  With this new release our complete platform -  again including VoiceXML, CCXML, ASR, TTS, and more -  will be openly available as a free two-port download.  Four ports will cost just $249, and additional ports will cost $549 or less.  The free solution makes it easy for developers and projects to get started, and the $549 per-port all-inclusive price makes it affordable to deploy.   At $549 we're charging less than 20 percent of what we and our competitors previously charged for an equivalent solution.  That is a dramatic decrease, one that took us four years of engineering to deliver.


This technology was built as a direct result of market demand from our customers.  If the traditional speech vendors had addressed ours and other requests for a low-cost solution previously, we never would have started this project, and if they offer one in the future, we will gladly sell it to our customers and prospects.


Q. You had mentioned partnerships with Vicorp and Apptera, what do these agreements entail?

A. I think they demonstrate the enormous pent-up demand for a product like Prophecy in our market.  On day one of Prophecy, we have one of the leading VoiceXML Service Creation Environment (SCE) vendors and one of the leading speech application providers announcing a partnership with Voxeo to use, market, sell, and bundle the Prophecy platform with their tools and applications.   These bundles will again decrease the cost and complexity of deploying speech solutions.  For example, Vicorp will be bundling a version of their high-end SCE tools at no-charge with the four-port, $249 Prophecy download.


Vicorp and Apptera are visionary leaders of our space, and we're thrilled to be working with them on Prophecy.  In the last few months we've revealed Prophecy to about 10 such companies, and every single one of them has signed on to build and bundle Prophecy-based speech solutions. 


Q. Now that you are forming an Application Channel Program, what effect will this program have on current and future partnerships?

A. Voxeo has always been a channel-focused vendor; in fact over 80 percent of our business is done with a channel partner.  We are unique in the speech industry - we do not build any speech applications ourselves.  We feel that an in-house professional services team would conflict with our channel partners and customers, many of whom provide value by building applications themselves in vertical markets they know well.   In fact, most of our long-term partners are not speech companies per se; they are vertical industry partners, and we enable them to deliver speech and phone-based solutions.


We also felt that in-house app-dev professional services conflicted with our goal of making the platform extremely simple to use and deploy.  When a platform vendor is getting paid to build applications, that vendor has a hidden incentive to make application development hard for the customer to do on their own.  We don't like that.


As a result, we bring a partner in for all application development work; in turn, they bring us in as their preferred platform provider.  In the past, these shared opportunities were only done with companies we've worked with for many years.  With the launch of Prophecy, however, we've experienced burning interest from new speech-focused partners such as Vicorp and Apptera.  So we've formalized our partnership program more, to fairly distribute opportunities.  Our goal is to work closely with our new speech industry partners while still remaining very loyal to our long term vertical market customers and partners. 


Q. What are the benefits of the Application Channel Program to your customers as well as the vendors who are participating in the program?

A. Really, the benefits are the same as the goals for the Prophecy project itself - reduced cost, reduced complexity, improved scalability, and open access.  By bundling best-of-breed development tools and packaged voice-applications with the Prophecy platform, speech solutions can now be deployed as quickly and easily as Web applications.  These bundles grease the rails, if you will - they make it easier for customers to buy speech solutions and easier for our partners to sell them.


Q. Is there anything that you would like to add?

A. I mentioned that we'd just closed our 24th consecutive month of profitable operations, and I'd like to thank every single one of our customers and partners publicly for helping us achieve that goal. 


I'd also like to invite all the attendees at SpeechTEK West to try the Prophecy platform.  As I mentioned, it's absolutely free, and anyone can download Prophecy now at http://www.voxeo.com/prophecy.

SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues