IVR

This may sound like a strange task for the world’s largest provider of hosted VoiceXML (News - Alert) services: create something for developers who want to write voice applications in different languages.

 

But that’s exactly what Orlando-based Voxeo Corporation is targeting with Tropo, a high-performance, cloud-based telephony platform that allows developers – in fact, already has seen developers – create and deploy applications in Groovy, JavaScript, PHP, Python and Ruby.

 

Here on the company’s blog, Voxeo’s (News - Alert) Dan York says that Tropo, currently in beta form, “is and will always be absolutely free for development use.”

 

TMCnet had a chance to put some questions to Voxeo President Chief Executive Officer Jonathan Taylor, and during our interview – printed in full below – we learned that the company views Tropo as a kind of exciting experiment to find out what developers want to do.

 

Taylor told us that the platform likely will be out of beta at summer’s end, and that at that time, the company will charge 3 cents per minute for production use. Even then, Tropo’s pricing structure requires no contracts and no monthly minimums, Taylor said.

 

Spurred by Voxeo’s own look into the architecture for the next generation of its Prophecy VoiceXML platform – as well as a trend that’s seen Web developers move from direct coding of markup languages such as HTML, into Web application development frameworks, such as Ruby on Rails and the Google (News - Alert) Web toolkit – Tropo could produce telephony applications and mashups that the industry has never seen.

 

And that’s just what Taylor says he likes most about it.

 

“Any company that claims there is only one ‘right’ way to build telephony applications is, simply put, wrong,” he said. “VoiceXML will continue to grow, telephony tool use will continue to grow, and API-based telephony will grow as well. Our business is about supporting that growth, not promoting one religion over another.”

 

Our exchange follows.

 

 

Jonathan Taylor (peek-a-booed and pictured left with his son): There were two things that really drove us to launch Tropo. First, we have been exploring the architecture for the next generation of our Prophecy VoiceXML platform, and as part of that we asked ourselves, “If we were going to achieve a density of 5,000 concurrent VoiceXML calls per server, what would that platform look like?”

 

In answering that question, we concluded that performance of JavaScript was key, as VoiceXML applications are increasingly using an AJAX-like model, where JavaScript within each VoiceXML document drives much of the dynamic caller experience and back-end data integration. So we began prototyping an extremely high-performance JavaScript environment for telephony applications.

Second, Web developers have increasingly moved away from direct coding of markup languages like HTML and have moved to Web application development frameworks, such as Ruby on Rails and the Google Web toolkit. A core part of our business has always been enabling non-telephony developers – primarily Web application developers – to easily deliver telephony applications.

 

In combination, we saw an opportunity. We could extend our work to enable a high-performance, direct JavaScript telephony platform with support for other leading languages – like Java/Groovy, PHP, Python, and Ruby – to deliver a new platform that mirrors the move toward high-level scripting frameworks in Web applications and complements our other voice products. So we grabbed that opportunity and built Tropo.

 

 

JT: Fundamentally, hosted, developer-focused telephony is what we’ve been doing for ten years now. When we started the company in 1999 our mission was to make it easy for anyone to create and deploy telephony applications. We launched a completely free, hosted developer program then and as a result we now have over 36,000 members in that developer program.

 

While other vendors have focused exclusively on the obvious and large call center applications – and on building most of the VoiceXML applications themselves via professional services – we’ve focused on enabling tens of thousands of companies and developers to build any telephony application they can imagine. The model has been enormously successful, leading to 20 consecutive quarters of profitability with yearly revenue growth over 60 percent. There are very few, if any, companies in our industry who have achieved similar growth while remaining profitable. 

Specifically, the financial advantages of hosting are extremely compelling in times of economic trouble. Hosting removes the often significantly large up-front costs to purchase, install and mange a new telephony platform. Many telephony applications enable customer self service of various forms, and in doing so significantly reduce enterprise support costs. Telephony applications save companies money. In parallel, you see the trend towards high level scripting frameworks we discussed previously. At the intersection of all of these we found an ideal environment for the launch of Tropo.

 

 

JT: We arrived at the 3 cent per minute price by looking at our known costs to deliver hosted telephony over the last ten years and combining those with the cost savings we achieve via Tropo’s extremely high performance and calls-per-server density. However, I think the other attributes of Tropo pricing – no contracts or monthly minimums – are even more significant than the per minute price. The combination makes Tropo very easy to do business with.

Tropo will be out of beta in the next three to six months. The beta isn’t about improving reliability or adding functionality, it’s about learning how developers want to use the platform, what challenges they face, and what capabilities they use the most or the least.

 

We’ll be learning right along with innovative developers and improving the platform every week with that knowledge. We’re not looking for Tropo to make a significant impact on our revenue this year. This year is all about working and learning with the Tropo developer community. I’m certain we’ll see some significant early successes. In fact, just two days after Tropo’s release, one developer had completed an application and wanted to launch it commercially. We said, “Sure, no problem – but we still won’t charge you.” Voxeo is extremely product-focused, and is in many ways a perfectionist company.

 

Our high growth and profitability lets us focus on creating the best products we possibly can. In contrast, our competitors are losing money and need to drive as fast as they can toward new revenue. That said, I do expect that Voxeo will see a significant revenue impact from Tropo in 2010.

 

 

JT: It was ten years ago this week that AT&T, Lucent, and Motorola launched the VoiceXML Forum. Since then, VoiceXML has been enormously successful – certainly the most successful telephony application standard ever. Over 30 different vendors have built VoiceXML platforms, and over 80 percent of all new enterprise IVR and speech platform deployments are done with VoiceXML.

Voxeo is the world’s largest provider of hosted VoiceXML services, and we’re the fastest-growing provider of on-premise VoiceXML software. We’re seeing significant growth with VoiceXML this year in both North America and Europe. However, we know from ten years of experience that there is not “one right way” to create and deploy telephony applications.

 

We know that there are at least three complementary ways to build telephony apps. First, XML-based telephony, using technologies such as VoiceXML and CCXML. Second, tool-based telephony, using products such as our own VoiceObjects (News - Alert) platform. Third, API-based telephony, using our SIPmethod SIP-servlet platform and now Tropo.

 

Any company that claims there is only one “right” way to build telephony applications is, simply put, wrong. VoiceXML will continue to grow, telephony tool use will continue to grow, and API-based telephony will grow as well. Our business is about supporting that growth, not promoting one religion over another.

 

 

JT: That’s my personal favorite thing about Tropo. We have no idea what applications will come out of it, and we love that. Tropo is enabling an entirely new generation of developers to quickly and easily create telephony applications, and I think we’re going to see enormous innovation as a result. Over the last few days we’ve seen people use Tropo to build extremely clever voice apps in just a few hours. I’d love to come back in a few months and talk more about what developers are using Tropo for. Until then, the sky’s the limit.