Speech Technology Magazine

The 2008 Market Leaders

Recognizing the leading speech engine, speech self-service suite, speech security, speech analytics, professional services, and mobile application vendors.
By Leonard Klie , Ryan Joe - Posted Aug 22, 2008
Page1 of 1
Bookmark and Share

The speech technology market has matured to where vendors have changed their focus from developing new technologies to perfecting existing ones and expanding the range of where and how they can be used. That shift has prompted us to add a new category as we rate vendors in six areas: speech engine, speech self-service suite, speech security, speech analytics, professional services, and mobile search. Each category features one winner, two leaders, and a vendor contender, as decided by the editors of Speech Technology magazine and a panel of experts from the academic, consulting, and analyst communities. The results on the following pages are based on a blend of weighted criteria, several of which are presented in the accompanying charts, for which the experts provided ratings from one to five (with one being the lowest and five being the highest).




SPEECH ENGINE

The Market
: Throughout 2007 and the first half of 2008, the number and size of speech recognition deployments continued to spike as companies have become more familiar with the technology and its capabilities. The number of devices with embedded speech recognition technology has also increased steadily. Merger and acquisition activity continued, but several new vendors have emerged and others have expanded their roster of speech offerings, adding a handful of new firms to those considered for this award. Many of these firms have a strong presence in emerging Asian and Eastern European markets, highlighting the technology’s vital role among developing nations. One could say the market has matured, but new advances in natural language processing, expanded vocabularies, new language capabilities, new uses for the technology, and improved audio output show that this is a market still fraught with potential.
The Leaders:
LumenVox
is a newcomer to our leaderboard, but it has slowly gained market share and customers. Its speech engine sells for between $100 and $400 per port, prompting our analysts and consultants to give LumenVox high marks for affordability. What’s more, its pricing model allows users to prototype the application before buying it, so they "know what they’re getting before they make an investment," says Robin Springer, president of ComputerTalk.
LumenVox garnered high marks for its ability to tailor solutions to specific user needs and overall customer satisfaction. With a number of key strategic partnerships in place, LumenVox "shows a huge openness to work with others and go outside itself to get the greatest deal for its customers," observes Juan Gilbert, a computer science professor at Auburn University.

Nuance Communications scored an industry-leading 4.71 in accuracy, almost eight-tenths of a point higher than its closest competitors. "Nuance’s accuracy is fantastic," says Susan Hura, vice president of user experience at Product Support Solutions. "I don’t think anyone can touch them here."
The company also scored high in customization and integration, company reputation, and overall customer satisfaction. As in last year’s evaluations, Nuance again fell far short in the area of cost. But despite its high cost, it has more than 5,000 customers for its speech engines, and revenue has grown by approximately 20 percent for each of the past three years. It has seen especially strong growth in its healthcare dictation business, and expects to ship its technologies on more than 800 mobile phones this year.

The Winner:
The past 12 months have seen great changes at Loquendo, which welcomed a new CEO in Davide Franco, opened new offices in New York, and underwent an organizational change resulting in a new marketing and business development division. Despite all of these changes, the Italian company retains the top spot  in 2008. Loquendo captured the highest scores in all evaluation criteria except one. In accuracy, only Nuance and Voxeo scored higher, but neither could compare to Loquendo in the number and variety of languages and voices available. "For European languages in particular, they are great," Hura says. The company boasts nearly 60 voices in a total of 23 world languages.
Loquendo also scored very high in customer satisfaction, customization, integration, and innovation, and analysts painted a rosy picture for the company as it continues to open new markets and uses for the technology.

Vendor Contender:
Like Nuance and Loquendo, IBM also scored very high in the accuracy of its speech engines, pulling in a 4.1 rating there, and in its ability to customize and integrate solutions. "You can plug it into almost anything, especially if it’s compatible with a Unix or Linux operating system," Gilbert says. The company fell off the leaderboard this year, ranking just slightly behind Nuance. As in last year’s rankings, though, the company continues to excel in product innovation, research, and development. Ongoing development focuses on a speech interface for the Web on mobile devices, real-time language translations through mobile devices, social networking on the go, and a wide variety of other applications that will bring a new dimension to speech technologies.


SPEECH SELF-SERVICE SUITE

The Market:
As the interactive voice response (IVR) systems market matures, vendors are continually trying to cram more  capabilities into their packaged self-service suites, either through home-grown applications or partnerships with other vendors that can supplement their own offerings. The industry as a whole is also making a shift from legacy, proprietary systems to more open-source, VoiceXML-based applications. In fact, industry analyst firm Datamonitor predicts that 2008 will be the first year that VoiceXML-based IVR shipments will exceed more traditional applications.
This evolution is opening new possibilities for multimodal and video applications, which Datamonitor’s senior analyst, Daniel Hong, calls "the next phases of IVR." As application deploymnet costs drop, Hong also expects many more midmarket companies to make their first forays into speech self-service applications in a hosted environment.

The Leaders:

Avaya scored above a 4.0 in all areas but cost, where it received a mediocre 3.5 score. This, however, is likely to change as the company revamps its focus to make solutions more affordable, especially for new small-business customers and previous customers that invested in the Avaya Interactive Response Portal and now want to upgrade to Avaya Voice Portal. The solution costs $900 per port, and Dialog Designer is provided at no additional cost. The  company is also making strong moves toward the Web, multimedia, and unified communications, all of which are expected to position it well moving forward.

SpeechCycle
specializes in collections, technical support, and help-desk applications for the telecommunications and utilities markets. Despite its limited reach, "in their own little playground, they are very well-regarded and good at what they do," observes Susan Hura, vice president of user experience at consulting firm Product Support Solutions. This is evidenced in the fact that SpeechCycle was among the top scorers in depth of functionality, customization, integration, customer satisfaction, and cost, and did not score below a 4.0 (on a five-point scale) in any of these areas. The company is also strongly grounded in innovation, continually revamping and adding to its technologies to improve the customer experience.

The Winner:
Voxeo this year improved the business case for speech applications by removing the barriers of cost and complexity, scoring an industry-leading 4.5 and 4.4, respectively, in those two areas. Helping its case, its open-standards-based Prophecy 8 solution for IVR and Voice over Internet Protocol telephony applications is available as a hosted on-demand service or an on-premises solution, and its Web-based Voxeo Designer tool, text-to-speech, and speech recognition engines are all included at no additional charge. Beyond that, the company also excelled in depth of functionality, with more than 20 new features and tools added to the latest version of its platform alone. Its customer satisfaction also clocked in at an industry-leading 4.5, mainly because, as Bill Scholz, president of the Applied Voice Input/Output Society (AVIOS) points out, "they have excellent customer support."

Vendor Contender:
Intervoice repeats as this year’s vendor contender, having held that spot last year as well. While the company scored high among analysts in most categories, it still suffers from what many perceive to be a high price tag. The company more than makes up for it, though, in innovation, having made advances in the use of State Chart XML (SCXML), VoiceXML, multimodal, and video applications, and the release of its Intervoice Contact Portal, which brings its already powerful Voice Portal, IP Contact Center, and Advanced Notification Gateway products together into a multichannel suite of contact center solutions.  Where the vendor will finish when the dust settles from its recent acquisition by Convergys is anyone’s guess, though early indications are that Convergys does not plan to dramatically alter its product road map significantly.


SPEECH SECURITY

The Market: Speech security was a challenging category to rank. For starters, companies like BBN Technologies, SRI, and IBM have voice biometrics solutions, but they’re not really tailored to commercial enterprises. The other problem is that roughly half of the vendors offer hosted solutions, which gives them an edge in ease of use and, by extension, customer satisfaction. Consumer-facing deployments of voice biometrics are still proving themselves, particularly in the United States. While American financial institutions are interested, it’s difficult to find one willing to be an early adapter. And while there have been successful implementations overseas, such as European banking giant ABN AMRO, what happens in Europe stays in Europe.
Additionally, analysts had a particularly hard time pegging accuracy rates for speech security solutions. "There are no tests to compare these," says consultant Judith Markowitz. "There are tests to compare face and iris recognition and fingerprint recognition. But there are no tests out there run by a government agency for voice biometrics that are totally independent." Ultimately, scores among the top vendors were practically the same.

The Leaders:
Spanish company Agnitio has numerous international customers in both government and corporate sectors. These include the French Ministry of Defense, the London Metropolitan Police, and New York-based Medical Management Technology Group. Analysts were impressed not only with KIVOX, the company’s flagship product, but with the way it shifted its technology from the homeland security market into the corporate environment. "Their forensics stuff is pretty well-designed," adds one analyst.
VoiceVerified had some of the highest rankings across all of the judging criteria. Its biggest advantage is that it offers a hosted solution. Analysts further praised the company for its ability to remain flexible and accommodate enterprise customers’ needs. Additionally, VoiceVerified still boasts a great deal of company expertise and, in the past year, fused text-dependent and text-independent engines to enhance accuracy as part of its PSP Unforgettable Password solution.

The Winner:
Ultimately, it was VoiceVault’s expertise that gave that United Kingdom-based company its edge—barely—past VoiceVerified. Both companies had the distinct benefit of offering hosted services. "They’ve got an unfair advantage," Markowitz concedes. "If the customer doesn’t have to do anything, they’re really satisfied." 
Still, one of VoiceVault’s largest customers is ABN AMRO, which speaks to its strong reputation. VoiceVault replaced the bank’s PIN-based password system, allowing customers to more quickly and securely access their records. Analysts were impressed that this installation, which was initially used only by bank employees, was made available for customers. One analyst was swift to note that while voice biometrics can be a boon for banks, financial institutions are traditionally conservative and reluctant to adapt. This may soon change. Last September, Allied Irish Banks (AIB) announced it was implementing VoiceVault’s biometrics so its staff members could automatically reset their passwords. 
"Their customers like VoiceVault quite a bit," says Daniel Hong, senior analyst at Datamonitor. "Their approach is pretty good."

Vendor Contender:
Perhaps the biggest endorsement of PerSay’s voice biometrics solution is the 600,000 voluntary enrollments from Bell Canada customers this past year. One analyst even placed PerSay, which is based in Israel,  at the top because of this massive, customer-facing implementation. Though not as nimble as the hosted solutions, technologically speaking, analysts believed the company is one of the best, especially in the way it secures a large swath of components and systems within the enterprise.
In the past 12 months, the company also reportedly improved its solution accuracy, having slashed error rates by 50 percent, and added a multiengine, language-independent platform that can be text-dependent, text-prompted, or text-independent.



SPEECH ANALYTICS

The Market: The speech analytics market is on the upswing, having grown by 106 percent in 2007 alone, according to research by DMG Consulting, and "momentum in 2008 has remained strong, despite the economic slowdown," reports Donna Fluss, the consultancy’s president.
Though vendors have been expanding their products into other verticals, "primary buyers of speech analytics are still contact centers," Fluss says. "Vendors are not yet penetrating sales and marketing organizations, which represent the greatest opportunities and potential for business benefits."
Potential for speech analytics solutions can also be found in other modalities outside of the contact center. Research conducted by Frost & Sullivan earlier this year found that enterprises also need to look at applying the same analytics solutions to chat, email, and Web data because of the increasing multimodality of customer interactions. In this context, it will also be important to link analytics data into the rest of the business as well, notes senior analyst Keith Dawson.

The Leaders:
NICE Systems, a newcomer to the Speech Technology market leaders this year, scored an industry-leading 4.6 for the depth of functionality associated with its multidimensional Interaction Analytics solution. The Israeli company also earned high marks for accuracy, but did not fare nearly as well in cost. With solution costs starting and $100,000, NICE scored among the lowest of all vendors in the speech analytics space, though the high price tag did not seem to affect overall customer satisfaction; the vendor still earned a 3.7 rating there. It remains unclear, however, where its ongoing patent battles with rival Verint will leave the company. NICE Systems already lost two of its four court cases with Verint, and, depending on the outcome of the two cases remaining, could have to pay as much as $100 million in licenses and damages, if unsuccessful.
Verint, a company which this year increased its customer wins in the speech analytics space by 138 percent, retains its spot on the leaderboard for the second straight year. A favorite among analysts and consultants for the depth of functionality in its Witness Actionable Solutions product line (earning it a 4.4 score out of a possible five), Verint continued to impress with the December 2007 launch of its Impact 360 suite of analytics-driven solutions. Some of the functions contained in this solution suite include quality monitoring, call recording, data analytics, workforce management, customer feedback, eLearning, and performance management. Verint also received high marks for its solution accuracy, ease of use, and ease of installation.

The Winner:
Nexidia, which this year retains the winner’s spot it earned last year, was the top scorer in all but one of the judging criteria—it finished just slightly lower than NICE Systems and Verint in depth of functionality, something it hopes to rectify later this year when it releases an expanded version of its Enterprise Speech Intelligence solution. Other releases have included the QuickStart contact center solution built for rapid deployment, and AudioFinder, an application tailored specifically for  discovery in the legal profession.
The company also earned an industry-leading 4.0 in solution accuracy, and performed well in cost as well. In fact, its unique pricing structure allows companies to perform a 90-day proof-of-concept for a small fee before committing to a long-term software license or managed services contract.

Vendor Contender:
Though known more for its speech recognition and dictation solutions, Nuance Communications is slowly making a name for itself in the very competitive speech analytics landscape. The vendor scored extremely well for the accuracy of its Nuance Care Analytics solutions, which rely on the Nuance Analytics Hub to deliver insight. Also helping its cause is an extensive Nuance Deployment Databank and the wealth of experience in Nuance’s Business Consulting Practice. Nuance took the top spot (and a 4.0 score) for its ease of installation and integration with other systems, but, as with its speech engines, lost ground on pricing.



PROFESSIONAL SERVICES

The Market: The professional services landscape is swiftly changing. Only one top vendor from last year maintained its position this year, and it’s likely that the landscape will shift again next year, especially given the purchase of this year’s vendor contender, Intervoice, by Convergys was announced in July.
So what did analysts look for this year in ranking professional service suites? "It’s not just contact centers," says Datamonitor’s senior analyst, Daniel Hong. "It’s understanding the entire data network structure."
All of the top companies scored big points for services offered—which is precisely the category that limited SpeechCycle, last year’s winner, even though it scored consistently high in ability to execute and shared the highest customer satisfaction score with Vicorp. Reactions to mainstay Nuance was polarized. "Nuance is heavily involved in professional services, often helping some of the other listed companies in partnerships," says Bill Meisel, president of TMA Associates. But others saw it more in terms of Nuance needing help rather than helping. "They make it too complex, man," one analyst fires back. "They can’t do it alone."   

The Leaders:
Of the top vendors, Cisco Systems, has the most inconsistent scores, garnering high votes from some analysts in services offered and ability to execute, but middling votes in those same categories from others. "They know a lot in terms of understanding data-voice convergence and infrastructure," Hong says.
West Interactive is the only holdover from last year, in part because it’s adapted particularly well to the increased complexity of enterprise infrastructure. West has a huge speech services team and research group, supplemented with a Product Advisory Committee that includes more than 20 of its clients. It scored extremely high in customer satisfaction, one analyst notes. This is especially impressive given the company’s size and number of customers.

The Winner:
Vicorp was this year’s winner by a slim margin, squeaking by West with an ever-so-slightly-higher customer satisfaction score and more competitive pricing to its solutions. While one analyst protests that Vicorp’s services aren’t as diverse, this,  too, is only by a slight margin. Analysts roundly are impressed with the work Vicorp, which is based in England, has done in the financial sectors, notably with U.K. financial institutions Barclays and Lloyds TSB.
One of Vicorp’s  biggest strengths lies in its versatility; its open-standards-based xMP service creation suite allows customers to develop complicated applications with relative ease. The company received extremely high marks for customer satisfaction, mainly because of its reputation for providing a rapid response to concerns and its ability to turn around a solution for customers in a timely fashion. Additionally, Vicorp routinely consults with customers when updating its application suite, as it did when preparing for the August release of xMP Version 5. That solution is said to have a number of new features and enhancements.

Vendor Contender:
It’s rather tenuous for Intervoice to be this year’s vendor contender given its July acquisition by Convergys for $335 million. Still, credit must be given where credit is due, and the vendor’s scores in services offered and ability to execute equaled both West’s and Vicorp’s, due in large part because of the 300 voice solutions experts it employs around the world. In the past year, Intervoice released Contact Portal, a fully integrated multichannel suite of contact center software solutions that brings together elements of its Voice Portal, IP Contact Center, and Advance Notification Gateway. Essentially, the portal opened a variety of channels through which end users could communicate with the enterprise: SMS, Web chat, phone, or email. The company was stung slightly in its customer satisfaction score, not because anything bad was heard—surprising considering the vendor has 5,000 customers in multiple countries—but because nothing was heard at all. "Their customers aren’t as vocal," one analyst concedes.



MOBILE APPLICATIONS

The Market: This year, we sought to recognize voice-powered applications that are starting to gain a foothold in the mobile phone arena. It seemed like a straightforward enough goal, but given the number of different things one can do with voice on a phone, an attempt to compare them all would be futile. After declining to comment on specific applications, one analyst laughed and said, "Good luck with this!"
She had a point: Is it possible to compare Nuance’s application that gives users total voice control of their phones and the many applications on them with Vlingo’s open-search application? Of course not. The criteria that qualify each application is far too different for a side-by-side, apples-to-apples comparison. 
Since we already have a category for speech engine, we decided to look specifically at voice-powered search on mobile devices. Compared to our other award categories, this is a much smaller and decidedly more immature field, yet it’s an exciting one that warrants recognition. This playing field is likely to open up as more vendors emerge, voice-based search applications become more widespread, and interface designers optimize the convergence of voice and graphical user interfaces.

The Leaders:
Nuance Communications’ vSearch impressed many of our analysts, and most agreed that the company had the best speech recognition. This isn’t a surprise considering the company has had the server for years. Last year, when the company acquired VoiceSignal, it also gained embedded speech support.
When Nuance demonstrated vSearch on the iPhone earlier this year, the speech recognition was a hybrid client, using both embedded speech recognition and network speech recognition to optimize voice search.
Nuance got dinged on its multimodal interface—an important aspect of mobile search, the best practices for which are still being worked out. "[Nuance is] after all a speech company oriented toward speech," says one analyst. "They’re dragging other technologies with them, but they’d rather buy than develop."
Microsoft subsidiary Tellme Networks perhaps has the biggest upside because of the number of devices on which its applications are currently available. Most of its competitors in this category are limited to BlackBerrys, but Tellme works on a much wider array of  handsets, including ones from Samsung, Motorola, and SonyEricsson. Still, the vendor was criticized for its multimodal interface, which left many analysts dissatisfied. "They don’t have it yet," says Bill Meisel, president of TMA Associates, "but they will because they’re associated with Microsoft."

The Winner:
Vlingo had by far and away the most highly regarded application. The company made a huge splash at CTIA earlier this year when it introduced Yahoo! OneSearch with Voice—essentially layering its speech technologies over Yahoo!’s open-search application.
Recently, however, Nuance aimed a patent infringement lawsuit at Vlingo—one that nearly every analyst queried considers frivolous. "It’s typical Nuance," says one analyst. "If you have a technology they like, they’ll either buy you or sue you. It’s what happened to VoiceSignal."
But Vlingo’s victory in this category shouldn’t be seen as pity points. The company scored highest in customer satisfaction and multimodal integration. "Vlingo is heads and shoulders above everybody else," says Bill Scholz, an independent consultant and president of the Applied Voice Input/Output Society (AVIOS).

Vendor Contender:
United Kingdom-based Novauris turned heads earlier this year when it worked to bring its speech-enabled NovaSearch application to Verizon’s Get It Now feature, which allows the user to download ringtones, music, games, and applications from his phone.  Analysts are particularly impressed with the speech recognition’s accuracy, though Novauris’ search function is limited in that it runs a closed-search. In other words, it searches from a preset list. "They’re pretty good, but not the top of the heap," Scholz says.



CATEGORIES and CRITERIA  Speech Technology magazine’s Market Leader Awards names one winner, two leaders, and a vendor contender in each of six categories, using a proprietary selection formula that involves input from leading industry analysts and consultants. The selection of these leaders was based on a composite score of the analysts’ and consultants’ ratings for each one in several key areas, including customer satisfaction, depth of functionality, company direction, accuracy, cost, and ease of use, using a weighted scale for each of these criterion based on their importance to current or potential customers.

Page1 of 1
Learn more about the companies mentioned in this article in the Speech Technology Buyer's Guide:
Cisco Systems, Inc VoiceVault
Learn more about the companies mentioned in this article in the Vertical Markets Guide: