Ted Cwiok, Convergys

NewsBlast: In a recent interview, AT&T CIO Hossein Eslambolchi said, "Speech technology has been on the back burner for many years, but the rate of innovation is probably five times faster than Moore's Law." Do you agree and what is the basis of your opinion?
Ted Cwiok: I would agree, based on the fact that advances in core speech technology are fostering additional innovation in application design and deployment. In my opinion there are three technology components that have greased the skids for innovation: recognition accuracy with larger grammars and, more recently, natural language processing, quantum leaps in TTS naturalness, and the proliferation of applications based on VoiceXML.

What these technology drivers enable is innovation across many fronts, driven by the realization that customer satisfaction and cost savings can be uniquely achieved by deploying speech technology. For Convergys, innovation means delivering applications with new network architectures and complementary Web-based products that more closely link live agents with automated speech solutions. Another example is packaged speech application components that improve quality and speed time to market.

NB: What are the communications needs that speech can assist businesses in solving?
TC: Every organization has to respond to a customer base that chooses to communicate across many channels such as the Web, e-mail, phone, etc. Speech is the only automation technology for the phone that truly addresses customer satisfaction, safety and cost savings. Speech is a more natural human interface, and you simply cannot realistically staff a call center with enough trained agents to handle spikes and not have idle ears during slow periods.

Speech uniquely allows call centers to adopt a layered communication strategy, targeting specific activities such as checking on the status of a order or transferring funds between accounts with speech due to speed, reliability and accuracy while using agents for cross-selling, up-selling or delivering premium service.

NB: Why should a company outsource their speech application instead of keeping control of the process on their premise?
TC: Time to market, ability to dynamically scale for peak calling periods, constant performance monitoring by professionally trained staff, reliability and redundancy, automated software and hardware updates, elimination of staff and training, and little if any capital expenditure are arguments in strong favor for outsourcing speech applications. Additionally, Convergys addresses investment protection and the ability to maintain control of the speech application by deploying standards-based VoiceXML solutions and offering a unique Open Hosting model to let customers choose where they would like to locate their data, application, media and even some hardware components.

It's also simply more cost-effective to outsource speech applications, especially over the long term. Remember, speech technology innovation is more than doubling every 18 months. Don't forget improvements to processing power and network and telephony infrastructure. Outsourced solutions automatically keep pace. Is there anything worse than maintaining a dated speech system that is so "visible" to your customers and management while your competition leaps ahead?

NB: Does this not make it more complex to deploy?
TC: Exactly the opposite. Speech outsourcing by companies such as Convergys is based on systems and facilities created for the most robust of requirements and extremely high call volumes. These already-deployed systems make speech deployment an everyday occurrence and standard operating procedure with respect to best practices requirements.

NB: Provide us with a few instances of a successful speech app being deployed.
TC: We are finding that speech applications are being deployed across many different industries. Three examples of successful Convergys applications include:

Pharmaceutical - A large pharmaceutical organization has deployed a speech application that allows callers to obtain a list of prescription drugs that qualify for a financial assistance program, and to obtain an application for the program. The application can also capture a caller fax number to fax back detailed information.

Almost 20 percent of all calls are now resolved without agent assistance, and without hold times: Average call duration has significantly decreased with a bonus of positive caller feedback. Not only are agents more productive, but the company can identify significant annual savings due to speech.

Retail - One of the largest specialty retailers implemented a speech solution that automates call routing, offers speech applications for automated interactions, and provides for quality control and training needs with full session recording capabilities. After a brief introductory message, callers are presented with options that include automated order status checks with speech or transfer to a knowledgeable agent.

Speech automation at this retailer delivers more rapid access to information while allowing agents to focus on areas of expertise. This speech system has delivered cost savings in excess of $1 million annually while improving customer service.

Travel - A major airline needed to reduce costs, improve agent productivity and achieve higher customer satisfaction rates. The airline also sought improved call capacity to prevent delays in answering calls due to temporary, but severe, increases in call volume due to bad weather and other external factors.

Speech proved invaluable at this organization when a major blackout spiked call volumes to levels that would have quickly overwhelmed agents. The application hosted at Convergys not only dynamically scaled to handle the unexpected massive call volume, but it also supplied automated flight status and lost luggage information. The result has been better customer support while achieving annual savings estimated in the millions of dollars.

NB: Discuss some of your partners in delivering speech technologies. What are you looking for in selecting partners?
TC: Convergys works closely with some key technology providers such as Intel and ScanSoft to deploy leading-edge technology. Convergys also works closely with application providers having a specific area of expertise. An excellent example of a partner in this regard is TuVox. They have the ability to quickly deliver FAQ details for technical support and other information generated from the same knowledge base used by the Convergys Answer Station product. Answer Station supports agents for e-mail and chat, and it provides a uniquely integrated solution for Convergys customers.

TuVox is a good example of the type of application partner Convergys is looking for. An Open Hosting model allows application partners to extend their market reach with a hosted solution requiring little, if any, up-front capital investment and faster time to market. It also allows Convergys to offer a greater range of speech solutions, taking advantage of focused expertise to more quickly and costs effectively address customer requirements.

NB: How are you and your partners providing lower TCO and enhancing the caller experience for your clients?
TC: Enhancement to the caller experience and lowering cost of ownership are the key drivers to innovation within Convergys. Convergys is applying resources to newer technologies such as Natural Language Understanding and speaker verification to provide an even more natural caller experience and add options for security. What's different with Convergys is that it's not simply an effort to integrate existing technology. Convergys has the resources to develop and extend new technology to even further move the bar toward excellence and cost savings.

Convergys differentiates from virtually all other vendors by being able to offer a solution set that encompasses live agents, Web-based automation, and speech. Convergys offers cross-channel integration designed to reduce overall costs and make agents more productive. For example, CTI integration for screen pops from a speech application is a standard component for Convergys. Benefits include not making callers repeat information while allowing agents to focus on more product activities.

Other areas of benefits such as continual software and hardware updates make the Convergys Open Hosting model more effective and efficient due to technology shifts and performance characteristics. Ongoing telephony upgrades and built-in redundancy also positively impact cost of ownership while delivering quality for callers.

NB: What are ten best practices an enterprise client should focus on when improving customer service with speech technologies?

  1. Consider deploying speech within the context of Web and agent interaction. Don't deploy a stovepipe speech application.
  2. Identify customer satisfaction "hot spots" and establish metrics for success.(e.g. hold time, consistency of response, agent focus on higher value, etc.)
  3. Listen and learn from recorded and live calls with agents
  4. Identify timeframe for deployment with critical dependencies
  5. Determine ability to maintain the system (e.g. staff, ongoing training, software and hardware upgrades, etc.)
  6. Grade importance of uptime and ability to handle spike conditions
  7. Take the time to test the user interface design. (Take advantage of speech to promote the corporate brand.)
  8. Integrate with standards wherever possible
  9. Deploy a speech system with robust reporting. Monitor reports for business as well as technical issues on a regular basis.
  10. Educate customers about the use and benefits of the speech system

NB: What will be the biggest event this year in offering speech technologies solutions?
TC: Deployment of speech applications based on NLU will be the most notable technology event in the upcoming year. Convergys Labs has completed a rigorous, multi-vendor technology assessment of NLU and will begin deployment this year.

NLU is particularly interesting as it offers a more flexible user interface and even more natural dialog interaction. It's not a magic bullet, appropriate for every application. However, NLU will bring speech to a new level for organizations supporting callers that may not necessarily know how to ask for what they need, or for where a large audio menu would be painful to listen to.

SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues