Escape From Voicemail Hell

Updated: August 20, 2012

Hoping to rescue valuable customers from voicemail hell, an increasing number of companies are putting their money where their mouth is - literally. Speech recognition technology is fast gaining ground as a friendly alternative to curmudgeonly call center operators and touch-tone madness. What's more, when coupled with VoIP , a speech solution can deliver countless business benefits, like bolstering customer satisfaction, cutting call center costs, and simplifying the end-user experience.

But talk isn't always cheap. In fact, previous incarnations of proprietary speech recognition solutions were often complex and costly, leaving a trail of abandoned implementations and disillusioned IT departments. Says William Stofega, a research manager at IDC: "There are a lot of costs and bad memories associated with speech recognition technology in terms of prior iterations."

Open Source Solutions

These days speech applications are making a comeback. According to Daniel Hong, a Datamonitor senior technology analyst, global spending on total speech solutions will reach more than $2.7 billion by 2009.

Fuelling the rebirth is a shift from proprietary software to open standards such as VoiceXML. Developed through a collaboration of AT&T , IBM , Lucent Technologies and Motorola, VoiceXML is an application of the Extensible Markup Language, which allows users to access a website by phone and interact with it through speech recognition technology.

In the past, businesses had to rely on vendors to custom-develop proprietary speech recognition applications. As a result, if a company decided to switch communication providers, it would be forced to abandon its existing speech solution altogether and rebuild from scratch - an extremely costly and labor-intensive proposition.

Speech applications programmed with VoiceXML, on the other hand, are portable and can run on multiple vendor platforms. In addition to slashing development costs, VoiceXML also offers natural sounding text and extensive libraries of words and phrases.

Big Cost Savings

Further driving the adoption of speech technology is the lure of enormous cost savings. The average U.S.-based call center call costs $5 if handled by an employee, but only 50 cents with a speech-enabled system, according to Datamonitor. And while speech recognition solutions can cost upwards of $300 per user, Carmi Levy, a senior research analyst at Info-Tech Research Group, says "the upfront investment clearly justifies the savings in back costs," particularly in reduced call center manpower.

That's not to suggest that the pairing of speech recognition technology and VoIP is a panacea. Warns Levy, "This is not something you're going to implement simply because someone in the organization thinks it's cool. Ultimately, there has to be a business driver within the rest of the organization that dictates that this will add value to the way the company communicates."

Elizabeth Herrell, vice president at Forrester Research, points to call completion as one reliable measure of success. For example, she says, "If companies have 30 percent completion of calls in their Interactive Voice Response system, they need to demand that speech technology is going to give them 40 or 50 percent completion to have a really solid return on investment." Herrell adds that although speech solution costs can run anywhere from $250,000 into "the millions," it can take less than a year for a company to realize a return on its investment.

Quality Control

Another step to ensuring the most bang for your buck is selecting the right vendor . If the goal is to overhaul an entire communications system, Levy recommends turning to industry behemoths such as Nortel Networks, Cisco Systems or Avaya for a full-blown IP-enabled telephony implementation accompanied by a speech recognition engine. For those companies that have already replaced their antiquated analog-based PBX systems with VoIP, Levy suggests point solutions from vendors such as Voxeo, Lumenvox and Tellme Networks.

In the end though, a speech application is only as good as the network it runs on. VoIP-centric pitfalls such as latency and jitter can greatly compromise the quality of a call regardless of a speech engine's bells and whistles. Subsequently, it's up to companies to optimize their VoIP networks and to verify that their IT infrastructure can support the additional voice traffic that can accompany a speech solution. After all, says Levy: "Quality of service will continue to be the make-or-break issue related to the success or failure of speech recognition technology."

Top Hosted Speech Providers

  1. Tellme Networks The Tellme Platform
  2. West Advanced Speech Recognition solutions
  3. Convergys ExpressWare, Convergys ExpressAs, Intelligent Call Routing, Natural Language Understanding, Text-to-speech
  4. Voxeo Prophesy
  5. AT&T Natural Voices
  6. Voxify Automated Agents

Top Premise-based Speech Products

  1. Nortel Networks Advanced Speech Recognition
  2. Genesys Voice Platform
  3. Intervoice Solutions Framework
  4. Avaya Voice Portal
  5. Aspect Customer Self Service
  6. Cisco Personal Assistant
Featured Research