Recorded greetings on a company’s telephone systems that require callers to respond by pressing keys on their touch telephone key pad (Interactive Voice Response systems) have reduced the workload for service representatives at companies, but have often left customers feeling alienated when they have tried to reach a “live” person at these companies. IVRs have traditionally been inflexible and often frustrating, as a caller can get caught in “touch tone no man’s land” when none of the options provided by the response system seem to address the needs they have, and the caller keeps getting routed through the same meaningless choices.

The next generation of voice recorded greetings now use speech recognition software that asks a caller several basic questions, analyzes the responses and automatically routes the calls to the general area needed by the caller. This speech recognition greeting with open options is known as a “natural language system.” While it is more user friendly, it is also extremely expensive to program and keep updated.

The cost differential between a traditional IVR and a more open ended natural language system is huge. The more flexible the language system is the less likelihood there is that a customer will need to speak with a customer service agent, which helps keep service call costs down. Speaking with a live customer service agent can cost a company up to $6.00 per call, compared to a customer call handled 100% by an IVR that costs a company about $.30 per call.

There is a new trend in the field of automated voice recognition that customer callers will not hear, which is VoiceXML. Voice XML is often described as serving the same function for speech IVR that HTML serves for the World Wide Web. Proprietary systems that use legacy IVRs are becoming a thing of the past due to the tremendous amount of time required in human hours and the related costs to program them. New IVR systems are more likely to be sold with VoiceXML already embedded in the system’s programs.

Yet another IVR trend is the use of computer telephone integrations (CTI) where data already provided and input by the caller on the IVR appears on the customer agent's screen when the caller finally gets to that agent. Speaking from personal experience, this is a welcome change. How many times has the prompt asked for an input of personal identifying information and then when a live human service agent finally picks up the call, the caller has to provide the same identifying information all over again? It’s beyond irritating, and makes the caller want to opt out at their first opportunity to press “0” to reach a live customer service agent.

Enter Microsoft, which is working to provide customer call centers with open speech-enabled IVR application development via its new Office Communications Server (OCS) 2007, which will be officially launched on October 16th. The Speech Server component of Office Communications Server 2007 will reportedly offer much greater IVR functionality.

Microsoft has developed RT Audio Codec which it uses in OCS, the PC calling feature in Windows Live Messenger and the Xbox Live online gaming platform voice calling features. The software works by compressing digital speech samples into packets and then decompresses them. Even better, the list price for OCS is reportedly only US$649 for the standard edition. OCS requires that the user already have a server and router and other required software.

No matter how much IVR systems are improved, however, most customers despise them. When a customer has a question, a user issue or a complaint they want to talk to a live human being to discuss it, not a voice activated system. Customers generally grow shorter tempered as the number of requests for information by the IVR increases. In fact, there is now a Web site that informs customers calling large corporations how to bypass the IVR menu and reach a live agent: www.gethuman.com. We agree. Humans rule!!!

Read more here.