The year 2007 has arrived, and without lack of predictions for what the New Year will bring. In the case of speech technologies, a forecast for the upcoming year points to increased satisfaction and improved automation, among other areas of growth.

 

Today, speech is being integrated into more and more products and services and as advancements continue to occur, speech technology has found itself at the forefront of a new wave of possibilities.

 

For a closer look at this and other aspects of the speech technologies market over the next year, I asked Joe McFadden, Vice President of Enterprise Product Marketing for Intervoice (News - Alert), a few questions on the subject.

 

 

 

One of the most significant developments is Natural Language Recognition technologies. Natural Language Recognition is based on an SLM (statistical language model). Unlike standard automated voice recognition systems that use a directed dialogue function that locks the customer into a menu. Sophisticated Natural Language Recognition systems with SLM allow the customers to bypass prescribed steps in a menu and more naturally direct a call or access information.

 

A tremendous advantage, for example, to larger call centers where hundreds of thousands of contacts are processed daily, is that customers who interact with a Natural Language Recognition system are more apt to work with the system as opposed to zeroing out to a live agent. Thus, these companies have the opportunity to handle more contacts during the day, get more customer inquiries answered by the system, as opposed to a live agent, and in the end experience a significant operating cost savings and more satisfied customers.

 

 

Absolutely - I believe areas we will continue to see growth in over the next year will be in the ability for businesses to now integrate legacy applications and systems into their voice solutions not only for external customer reach, but also for internal company applications such as helpdesks, general communication applications, and CRM activities.

 

 

With the advent of IP-based contact center systems, we’re seeing a much more seamless integration of self-service, deferred assistance and live assistance. These improvements along with speech technologies for self-service are defining a new customer-centric standard for the customer experience.

 

The immediate assistance service model is typically delivered by live agents by phone, but can also be delivered via web chat and immediately escalated emails. Deferred assistance is typically provided by live agents responding to emails and voice messages. And as mentioned earlier, self-service is typically delivered over the web in the form of knowledge base information or over the phone via an IVR.   Increasingly, the immediate channel is able to benefit from speech applications as well by using natural language routing to get the call to the correct agent and to automate the routine identification and security questions before handing the call off to the live agent.

 

The degree to which these service models are integrated determines how intelligently and cost-effectively a company responds to a customer. These service models traditionally were designed around proprietary hardware and software solutions in a silo fashion with little to zero integration. The inefficiency of the silo architectural design and the cost of integrating these service models with proprietary hardware are often prohibitive, and therefore seldom considered. Today’s customer routing management software applications leverage IP technology to integrate all three service models into a single consolidated software application on the corporate data network.   

 

 

I believe we will see any number of new mobile speech applications that fall into a few major categories:

 

Hands-free Mobile Applications – further improvements and wider deployments of speech technologies for mobile devices that advance email access will extend worker productivity.

 

LBS (Location Based Services) - We have been saying for years that LBS are useful and necessary. Today, for example, we are seeing highly interactive, sophisticated LBS applications locate your nearest ATM or McDonalds simply by using your voice.

 

Multi-modal Applications – such as the ability to speech enable request for credit card balances and the ability to deliver the last 10 credit card transactions to the cell phone and physically display those last 10 transactions on the cell phone screen, thereby delivering the information in audio and visual formats.

 

 

The IVR is comprised of software and business rules which all companies typically are looking to continually improve in both the technology and business process area. I believe we can look forward to significant enhancements in customer service over the near term such as personalization of the IVR and speech self-service solutions. Such solutions would never forget which user you are and what you were doing and in those instances where you are transferred to a contact center agent, the agent would receive a quick screen pop and be able to tell very quickly who you are and what you were calling for. Personalization sounds easy, but in this instance, the requirement may exist for the integration of multiple databases to be accessed in order to provide the end user with all of the required information requested.

 

No, I do not see touch-tone IVRs disappearing in the near future. How simple or how complex the application is will drive someone to a particular solution. The two deciding factors regarding touchtone versus speech are 1 – user-choice and demand, and 2 - the simplicity versus complexity of the application deployment. As speech technology improves and deployment becomes easier it will continue to find more applications, but touchtone systems are still efficient solutions to some problems and are likely to be with us in some areas for many years.

 

 

 

Yes, I do think this has been slower than the industry expectations. I do see adoption accelerating as a result of several factors:

 

-User acceptance of the new speech technologies

-Improved technologies such as Natural Language Recognition

-Improved design and deployments of applications

 

Businesses are being very cautious with their spending. Contact centers and IT organizations are taking inventory of their environments. They are spending the time to study where they want to grow to and then take their time investigating their options in the market to see who has the best solution to grow and help them integrate and change their business over the next several years. These are fairly major shifts in customer service and end user touch points that affect a company’s brand, so these decisions are not being made lightly, quickly or without lots of deliberation.

 

 

 

As with other standards, adoption of VoIP will lower the cost of system deployment, leverage existing infrastructure and lower the cost of telephony services. It can also drastically simplify some applications and enable others by allowing more flexible call control across systems and geographies, without the need for complex and costly CTI projects.

 

 

Absolutely, as stated before:

 

-Consumers will continue to adapt to these new technologies and ultimately prefer the speech-enabled applications

-Technologies for speech recognition such as Natural Language Recognition will continue to improve and evolve

 

Perhaps the most important reason is that vendors will continue to improve upon the design and tuning of speech applications.

 

 

Challenges that customers will face will be working with and deploying multi-vendor solutions. In effect, if I purchase my platform from one vendor, it may be separate from the application provider, which in turn may be separate from the application consulting team. Finding a single speech application vendor solution for the end-to-end speech application should be the customer’s goal.

 

 

Absolutely – Again, a good example we will see is multi-modal applications such as the ability to speech enable requests for credit card balances and the ability to deliver the last 10 credit card transactions to the cell phone and display the last 10 transactions on the cell phone screen for both audio and visual access. Natural Language Recognition will also allow consumers to dialogue more naturally with automated systems.

 

 

We’ve always been an end-to-end customer service provider. Of course this also means that we are an end-to-end speech application provider. Intervoice is a single source provider for speech platforms and packaged custom applications and application consulting services. We’re also addressing market demand for delivery choices. Today we offer enterprise, premise-based solutions, managed co-location services, and hosted solutions.

 

For more information about speech technology applications or to contact Joe McFadden, directly please email him at [email protected].

 

 

 

-----