Network Computingwww.networkcomputing.com

Virtual Assistant Technology Makes Its Way Into the Enterprise

Posted by Ericka Chickowski
August 13, 2012

Apple's Siri technology has proven that real-world users can and will embrace speech recognition and virtual assistant technology if applied even reasonably well on mobile devices. The question now is whether the enterprise can leverage the interest stoked by Siri and advances in speech technology to create mobile enterprise products that actually help the business run better.

According to experts like Dan Miller, senior analyst at Opus Research, the answer is yes--as long as enterprises don't try to build tools with the same kind of "boil the ocean" mentality that Siri brought to the table. Instead, successful tools will be targeted at specific tasks and will help mobile enterprise users more easily do their jobs and consumers more quickly interface with targeted sets of information to get better customer service.

"There will be a short list of routine things that people will ask mobile natural speech technology to do once they discover that it's pretty good at it," says Miller. "It's not going to take us to Mars or discover new elements in the universe, but it will get really good at helping us do routine things and make those annoying corporate tasks easier."

Just as FAQs can offer people help by providing information from a pre-defined knowledge base, enterprise mobile virtual assistants have the potential to be very good at helping people with "frequently invoked task lists," he says.

And the engineers behind the technology responsible for Apple's Siri over at Nuance Communications believe enterprise developers can produce meaningful virtual assistant mobile applications if given flexible access to the underlying speech recognition technology to customize it for very specific user bases. That's why Nuance recently took the covers off of an SDK for Nina, an enterprise-focused mobile virtual assistant that can be customized and baked into enterprise apps for a range of customer-facing functions.

"The SDK is about easing the integration and being able to really abstract the complexity of the technologies that provide the power and intelligence of Nina to render her quickly into a mobile application for customer service," saysChristy Murfitt, senior solutions product manager at Nuance.

The SDK will be used in a forthcoming pilot planned for this month by the financial institution USAA to integrate Nina into its mobile banking application.

"We believe that the virtual assistant has tremendous potential to make it simpler, faster and more satisfying for our members to manage their financial affairs on their mobile devices," says Neff Hudson, assistant VP of emerging channels at USAA.

According to Nancy Jamison, principal analyst at Frost & Sullivan, Nina is compelling because it offers an easy way to combine a mobile business application with natural learning to more quickly put a user in touch with the right information, person or process.

"Plus, because it isn't all voice, navigation errors are reduced and information choices can be quicker because the user can see query results on the screen of their device," says Jamison. "It is also flexible for the customer because they don't have to use voice--they can use typing and tapping, as well.

"The intent of Nina is different than that of Siri," adds Jamison. "Nina is designed to help facilitate customer service in a quick and intelligent way, allowing a customer of a business to get something done. Siri is a more broad-based, speech-driven personal assistant that allows a user to find or do a broad range of things."

While the initial use of Nina is specifically focused for customer service applications, Murfitt says that as Nuance has piloted the technology in enterprises, it tends to get "creative juices" flowing within IT groups interested in broadening the use cases of Nina. So the potential for internal mobile applications utilizing natural language understanding definitely remains on the table.

After all, says Miller, "enterprise users are a sort of consumer, too."

Related Reading