Amit Desai, Chief Technology Officer and Founder, Voxify

Voxify develops and hosts automated call center applications, referred to as Automated Agents^. The company is headquartered in Alameda, Calif. with sales, support, and development offices and additional representation through its global channel partner network in locations around the world.

Speech Technology Magazine sat down with Amit Desai to discuss the benefits of hosted services and how that fits into Voxify's offerings.

Q. Amit, please tell us a little about Voxify and your role with the company.

Voxify delivers speech self-service solutions to call centers at 20 percent of the cost of a live agent while improving customer service.  Voxify speech applications are pre-trained to handle call types for specific industries so they can be deployed in as little as 8 weeks.  Voxify speech self-service applications are deployed on a hosted managed service basis so deployment is fast and customers only pay for the calls that Voxify takes.

As the CTO of Voxify, I evangelize and research advanced speech technology and voice user interfaces. I provide expertise in voice user interfaces, speech recognition, and platform software to help guide technology direction.

Q. We've kicked off another year of SpeechTEK in New York and I understand you are speaking at the show. When and what you are speaking about at this year's conference?

I am speaking in the VUI Track on "VUI: Controlling your Personas" where I'll be talking about designing your optimal persona by identifying persona design factors, variable factors and offering best practices.  The session is Tuesday, Aug. 8, 2:00-3:30 p.m.

In the Tools and Techniques Track on "Reusable Software," on Wednesday, Aug. 9, 3:45-5:00 p.m., I'll be talking about designing highly usable systems and how to reuse the code by componentizing behaviors which are interaction skills common to human-live agent interactions.

Q. What trends have you seen emerging in the speech technology industry so far for 2006?

Automated speech technology is rapidly becoming a required part of the customer experience with its high level of accuracy.  Today, the real value in the speech technology stack has moved to the speech application level.  All the component technology below the speech application layer is of similar, high quality and has become commoditized.  Companies that want to automate calls with speech only have to look at the speech application vendor to see if the application can handle specific types of calls.

Hosted speech is an important trend especially with increasing market pressures for rapid time to deploy, companies have less time to adjust and improve how they deliver customer service.  In addition, with tight budgets, speech applications offered on a hosted basis frees companies from the cost of acquiring speech expertise, the cost of capital to run the speech system, and the cost of maintaining a 24x7 operation.  Another important aspect of a hosted system is that it gives companies almost unlimited ability to handle call spikes while only paying for the calls taken.  Speech vendors that host speech applications have economies of scale enabling their customers to have access to as many speech ports as needed without having to pay for them in advance. 

Another important trend is that speech applications now come as pre-built 'templates' with conversational behaviors, industry business rules, and call type functionality.  Companies implementing speech today should not be looking at custom-built, hard-coded implementations that take a year or more.  Working from speech application templates is far less expensive, less risk, and has reduced the time to market to weeks instead of months.

Q. What role does hosted speech solutions play in these trends?

A hosted speech model offers the perfect solution for companies to rapidly use speech technology without compromising high quality service.  A recent Forrester report states that "self-service must not equal less service." Companies can now leverage a trusted partner like Voxify whose core competency is building, implementing and maintaining speech applications. 

Speech applications offered on a hosted basis is less expensive to deploy, has lower business risk, and gets a company to ROI faster.

A hosted system can give companies one thing that a premise solution can not offer - almost unlimited ability to handle calls spikes.  With a hosted solution, speech ports are shared amongst different customers of the hosted vendor giving economies of scale and making extra ports available when certain companies need them.  Hosted vendors know how to optimize usage of the ports based on different customers' seasonality in different industries so there is always excess capacity.  A premise solution cannot offer this because it would require the company to buy more speech ports than it normally needs.  The economics are not there so companies usually by just what they need and not more. 

Q. Why should an enterprise deploy a hosted solution versus a premise-based solution?

If a company uses a hosted solution all it pays for is the upfront set up cost and the per-minute rate for each call taken.  It is very simple and low risk.

If a company deploys a premise-based solution it has to buy the entire speech platform infrastructure including VoiceXML browsers, speech recognition engines, text-to-speech software, operating systems, and hardware.  It has to have speech experts to train, tune, and maintain the system.  Most companies do not have those types of resources at their disposal and they will not reap the economies of scale in hardware, software, or IT resources by creating a premise-based system themselves.

A hosted solution developed from templates can offer rapid time to market.  Working from templates hosted applications can be done in as little as eight weeks.  A premise-based solution requires time to purchase hardware and software, and hire speech experts.  You will not see a premise-based system being deployed in eight weeks.  It will be more like eight months or over a year.

Q. Please tell us about one of your latest hosted speech deployments to an enterprise user.

We have an exciting new customer deployment in the Financial Services market.  Nelnet is the National Education Loan Network and is one of the leading education finance companies with student loan assets of $21.3 billion. Every May, the Federal Government resets the Federal Student Consolidation Loan Interest Rate, which takes effect July 1.  If the rate increases sharply, the Nelnet call center is swarmed with callers wanting to consolidate their loans before the July 1 rate increase.   Additionally, Nelnet's own direct response promotions to thousands of student loan holder's further increases the call spike. 

Nelnet turned to Voxify to handle the call spikes.  For Nelnet, every call is potentially a new consolidation loan so they cannot afford to miss a single call. 

The results are that Voxify Automated Agents were able to take all of the calls and no caller were left on hold or lost.  The Automated Agents qualified callers for loans and if qualified, captured the loan application information.  These are the same tasks as a live agent.  Abandonment rates during peak periods were greatly reduced and loan assets increased.  Nelnet is very pleased with its decision to implement speech self-service.

Q. Was there anything that stood out as unique in this implementation versus previous deployments?

Nelnet had an urgency to be up and running by June before the major call spikes came.  Voxify typically has short deployment times but because of Nelnet's call crunch coming, we deployed this sophisticated application in only four weeks.  Nelnet is a great customer and they worked with us to get this done.

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

Voxify is hearing from our customers that their customers are demanding speech self-service.  Callers are demanding that they be served when they want and in the way they want.  They are increasingly impatient and do not want to be put in a hold queue if they can get the information quickly and efficiently through a self-service system.  Our customers are saying that the way to improve customer service is to offer speech self-service as another channel in which to be served.