Speak Up, 'Yer Talking Too SoftlySpeak Up, 'Yer Talking Too Softly
Nuance's software can understand dialects and turn up the volume.
Gender detection and the ability to recognize a Texas twang are some of the new features customers can expect to find in Nuance's latest speech-recognition software.
Nuance 8.0 features auto-tuning, which over time improves the system's accuracy in recognizing dialects or accents within a certain group of similar callers. The system also can recognize callers who access the system via cell phones or from noisy environments, eventually making adjustments to regularly filter background noises.
The enhanced system includes a personalization feature that detects a caller's gender. Gender detection lets companies that use the system in their call center determine which products or services to market to the caller. For example, a retailer could program its automated speech system to market upcoming sales on women or men's fashions, says Marcello Typrin, Nuance senior product marketing manager.
It also can detect if a caller is using a cell phone or a landline unit. If a caller is speaking too softly, "it prompts the system to say 'Can you speak louder,'" Typrin says.
A new feature called AccuBurst can detect the availability of memory and processing power in the speech-recognition system. When necessary, AccuBurst can apply underused memory to the recognition process to cut error rates by up to 30%, Typrin says. The memory redistribution can help save money. "You don't have to invest any more in computing hardware to get a boost in accuracy," Typrin says. Nuance 8.0 is available now, starting at $200 per port.
Improving accuracy can also lead to improved sales. "The more you can minimize the times the system doesn't understand you, the better the experience for customers," says Dan Hawkins, voice business program manager with consulting company Datamonitor.
Nuance is working with several interactive voice-response and telephony-hardware partners such as Edify Corp. and Cisco Systems, who will integrate the speech-recognition system into their respective customer-relationship management or IP telephony products to be sold to their customer base.
About the Author
You May Also Like