The 2008 Star Performers
As with any technology, vendors of speech systems and solutions constantly need to innovate to deepen their competitive footprint. In the process, several vendors have also moved the wider speech technology industry forward in grand fashion. The vendors that we honor with the 2008 Star Performer Awards are already well-known and established companies that have brought the industry to new heights through promising product or service launches, renewed attention to existing products or services, and an overall focus on the future. Their endeavors during the past 12 months have involved a wide range of technologies, targeted at very different vertical markets, but they have all led to a greater acceptance of speech as a viable solution with the power to alter how everyday tasks get done.
Sync: A Macro Splash by Microsoft
Software giant Microsoft doesn’t do anything on a small scale, and its launch of the speech-enabled Sync in-vehicle communication and entertainment system with Ford was no exception. Backed with almost ubiquitous advertising, the Sync system—which uses speech recognition technologies from Nuance Communications and Bluetooth technologies from CSR—came as standard equipment in most 2008 Ford, Lincoln, and Mercury models. The software, which includes voice-activated controls for Bluetooth-enabled phones and MP3 players, targets an increasingly tech-savvy, younger consumer base and capitalizes on a growing collection of legislation that restricts the use of mobile or other electronic devices while driving.
Sync is "the first significant product beyond the basic IVR that has speech recognition and has captured widespread attention and the imagination of the consumer," says Jim Larson, an independent consultant. "It’s certainly very well-engineered and very innovative."
Sync’s success in the field reflects the increasing demand for speech-enabled applications in hands-busy, eyes-busy environments, and has prompted analysts to predict that speech-enabled automotive infotainment solutions will become standard features on all cars within the next five years.
Microsoft is making other moves to speed up that time line. In May it struck a deal with Korean automaker Hyundai-Kia Automotive Group to make the same technology that powers Sync available in select Hyundai and Kia cars starting in November, when Ford’s one-year exclusivity agreement with Microsoft ends. At that time, it will also be able to license the technology to any other automobile manufacturers.
Though Microsoft is keeping a tight lip on what, if any, additional features the newest versions of its automotive technology might include, many people have speculated that future iterations will come equipped with capabilities for control of navigation devices and DVD players, and enhanced safety features, all activated by voice.
Microsoft is also expanding its automotive reach with other technology manufacturers besides Nuance. In June it announced a collaborative arrangement with Swiss embedded speech solutions provider SVOX to further speechify in-vehicle communication, infotainment, and navigation products. It has had a similar deal in place since September 2007 with Germany’s Siemens VDO Automotive to develop in-vehicle video entertainment and Internet connectivity through an integrated USB port in the car. These are likely to be the first of many such deals that Microsoft inks during the next few months to enable a much faster turnaround of projects.
Datria's Pick 'n' Pack: A Warehouse Breakthrough
A year ago, Datria changed the paradigm for voice-enabled warehousing operations by creating the first order fulfillment application that allows warehouse workers to communicate directly with central warehouse management systems using wireless phones and headsets instead of expensive, proprietary computer terminals and other hardware. By enabling inexpensive IP phones to be used in its Voice Pick ‘n’ Pack solution, Datria allows warehouse operators to spend less than $300 per worker; other devices can cost $1,500 or more apiece.
In developing the solution, Datria integrated Cisco’s telephony and Wi-Fi technologies, though the highly scalable solution can work with any off-the-shelf IP telephony product to create a two-way audio interaction between a warehouse worker and the central warehouse management systems through a company’s VoIP network. The solution also incorporates speech recognition technologies from Nuance Communications.
Pick ‘n’ Pack is also the first to make use of network-based speech recognition, allowing warehouse operators to take advantage of service-oriented architectures (SOAs) and Web services, and to better integrate their speech system with other systems throughout the warehouse. The platform-agnostic solution complies with a range of voice and data industry Web standards, including XML and VoiceXML 2.0.
Furthermore, as one of the first speaker-independent systems on the market, Pick ‘n’ Pack eliminates the time needed to train the system to each specific worker’s voice.
"Companies can now use network speech, VoIP, and off-the-shelf products like IP phones and headsets in their warehouses to reduce operational costs," observes Daniel Hong, lead analyst at Datamonitor. "This approach is new, and Datria has capitalized on this opportunity by leveraging its assets and expertise in speech-enabled enterprise mobility to bring forth an innovative solution for the warehouse market."
The Voice Pick ‘n’ Pack solution is already being used by Coca-Cola Enterprises, the largest Coke bottler and distributor in the country, to automate the workflow of more than 2,000 employees across a number of sites.
Syntellect: Voiyager of Testing Solutions
In application design, errors can appear at any time and without warning. With no opportunity to test every call path in a large-scale interactive voice response system, designers have to do their best to design and build a solid system and then hope and pray that no major errors pop up.
It’s in that environment that officials at Syntellect first envisioned Voiyager Dynamic Application Discovery. When they finally introduced the application at last year’s SpeechTEK event in New York, they knew they had something special.
Since its official release late last year, Voiyager has been touted across the industry as a breakthrough product that will fundamentally change how VoiceXML applications are designed, tested, deployed, and ultimately maintained. It is the first product of its kind that can test all aspects of an application from end to end across every call path, every state, every time, penetrating the full call-flow design path to expose potential design flaws, errors, and suspicious behaviors that will impact customers and the level of service.
From simple tasks, like renaming states to match the documentation, to higher-value tasks like selecting a state and getting the browser phone to place a call that gets the user to the state, the developer can rapidly rerun the problem as often as needed to isolate and verify the fix. Using a Voiyager bookmark, a developer can rapidly rerun any saved call to accelerate the call testing process.
The platform-independent Voiyager application, which works with all versions of VoiceXML, is capable of not only testing a massive amount of code, but also can do so in record time. In one early demonstration of its capability, Voiyager completed a month’s worth of application testing in about an hour. In another instance, it uncovered production-level bugs that would have generated more than $500,000 in development resolution and opportunity costs.
Another nice feature: Voiyager does not require extra setup; users simply point it at the URL and let it run. By eliminating the professional service requirements, telephony and speech recognition infrastructure investments, and setup, testing can happen early and often.
Blinkx: Using Speech to Find Videos
Blinkx has been gaining a lot of attention the world over for its video search engine, which uses a combination of conceptual search, speech recognition, and visual analysis software to search for and present online video content to Web browsers. The company’s founders set out to offer an alternative to basic keyword-based search technologies that only scratch the surface as TV and Internet user-generated video content explodes. They’ve done so with technologies that listen to and see the Web in all of its forms.
In addition to several key customer signings in late June and early July—including Russian Internet portal Rambler, MSN UK, and Elo, a Brazilian distributor of films, documentaries, music videos, and fashion shows—the company recently launched the Red Label program, which allows customers to deploy blinkx’s patented video search technology on their own Web sites for free. The service is being launched in two tiers: one for larger customers using blinkx’s own services, and the other a self-service portal for smaller sites.
To date, blinkx has more than 26 million hours of premium audio and video content, including favorite TV moments, news clips, documentaries, music videos, and video blogs, in its index. The company already has partnerships with 350 video and online advertising firms around the world, and it has no intentions of stopping there