The

Live Webcast: Voice Biometrics: Overcoming Barriers to Adoption

Thursday, February 2, 2012 — 10 a.m. EST / 7 a.m. PST
Sign Up Below!

Or click here to learn more and to register to participate.

This Report made available courtesy of Validsoft. Contact Pete Headrick (pheadrick@opusresearch.net) to receive a copy

Click Here to View the Report Summary

According to coverage in the Australian press Salmat, which handles over 100 million inbound and outbound calls each year on behalf of its clients, had “scoured the world” for a new solution provider. Its vendor selection started with 10 prospective winners but, in the end, was winnowed down to Avaya and one other (unnamed) candidate.

We’ll presume that Genesys, as incumbent platform provider, was considered as a candidate from the start. Indeed, it had shown an impressive array of enhancements to its core Customer Interaction Manager (CIM), including the introduction of a Conversation Manager to support “intelligent” handling of multi-channel and multi-modal interactions. Genesys also has a longer history of hosted implementations of CIM as well as the use of the Genesys Voice Portal (and its precursor VoiceGenie platform) for applications involving voice-based authentication.

The transition to a totally new platform is, to the best of our understanding, something of a rip-and-replace operation during which business continuity could be something of a challenge. This is especially true because Avaya, has had a love/hate relationship with “hosted” or “cloud-based” implementations. The move to IP-based implementations provided some initiatives toward hosted renditions of IP-based ACD and Contact Center resources. According to literature, both Telecom Italia and Sprint have offered hosted renditions of Avaya’s IP-based call processing resources.

However, to the best of our knowledge, Avaya has yet to bless or integrate voice biometric authentication into its IP-based Contact Center or its Experience Portal. By contrast, Salmat has made a point of differentiating itself by offering caller authentication as a service to financial institutions, healthcare providers, telecom service providers and government agencies in Australia and New Zealand. We’re anxious to see how smoothly the re-hosting takes place in the coming year. And we want to welcome Avaya into the community of voice biometrics-based solution providers.

Register Now – Discounted Offer Ends November 30th:
The first 30 people to register for Voice Biometric Conference New York can register for $299.00 That’s over half off the full conference rate of 699.00. Use registration code, ‘earlyvbc’ to receive the special rate.

Here are the important findings as reported by Nuance:

62 percent of respondents have more than 11 usernames and passwords, and that they forget their passwords between one and five times a month. Each time a password is forgotten it’s lost time that could be spent being more productive at work or at home. In addition, 74 percent of those surveyed have been blocked access to important information while using a mobile device because they could not remember their log-in information. The survey also found that 80 percent of respondents use the same password for multiple accounts if it fits the password criteria.

More detailed findings can be found here.

When viewed as a whole, the responses reveal that present solutions to mobile security are both inconvenient and insecure. There are no surprises in that. There should be no surprise in learning that 77 percent of respondents would be comfortable using voice biometrics as an alternative if it meant tighter security, and that a combination of voice and a password or PIN was preferred by 61 percent of those surveyed.

These findings are a significant upgrade from a survey conducted in 2009 by contactcentres.net in Australia and New Zealand. When asked about their preferred methods for identity verification, voice biometrics rated first, but only by 45% of participants, followed by PIN (21%), password (18%) and personal details or history questions (16%).

Nuance’s Twitter poll shows how dis-integrated our mobile identities have become. I fall into the category of individuals with well over 11 user names and password combinations across mobile communications and commerce sites. I was also one of the respondents who was locked out of a site after mis-remembering my password while trying to rebook a connecting flight through a mobile Web site. How I look forward to the day when I can use my voiceprint for more convenient authentication.

The companies are looking for “consumer-facing organisations that typically handle five million or more customer contacts per year, have customer authentication requirements, and currently operate a cross-channel customer service strategy. The incentive to participate will be to gain first-hand experience with voice biometric-based authentication systems and to take part in the development of a multi-channel authentication strategy. Companies are expected to get their applications in to Pierce Buckley at VoxGen before October 28th.

The companies are on track to finalize the deal by the end of September and have already started contacting Loquendo’s customers and go-to-market partners regarding Nuance’s roadmap for combining the companies core product lines. Dan Faulkner, Nuance’s Vice President of Product Management and Marketing, sees a very good fit between Nuance and Loquendo. For one thing, the Italian company has real strengths in markets where Romance languages are spoken, meaning Western Europe and Latin America.

Loquendo takes great pride in its investment in R&D, with an emphasis on product development. It relies on go-to-market partners to bring complete solutions to enterprise customers. This approach will help avoid channel conflict as the combined company defines both its product portfolio and go-to-market strategies.

As Faulkner puts it, the purpose of the acquisition is “to amp up the adoption of speech on a global basis” by adding high-quality ASR, TTS and voice biometrics assets into Nuance’s product portfolio. For skeptics, who see the elimination of competitors as lost opportunity to promote price competition, Faulkner points out that there are at least three formidable competitors on a global basis making significant investment in “broad solutions” that include speech processing, business process automation and self-service.

As I see it, the acquisition marks a watershed for speech applications, self-service, and conversational commerce in general. It has been a while since enterprise customers bought speech-enabled IVRs or added speech self-service to their offerings based on the accuracy of the recognizer. Decisionmaking has moved to how quickly a system integrator or other solution provider can bring up a new capabilities, features or functions that promote a better user experience. Nuance and SpeechWorks were strong rivals (until the merger). Then IBM or Loquendo or Lumenvox or Prophecy would appear as the “strong seconds” in the context of competitive bids and RFPs. Going forward, that role will be played by other speech technology providers (like Lumenvox, Voxeo/Prophecy or regional specialists).

At SpeechTEK 2011 in New York last week, there was a lot of talk of divorcing speech processing (ASR and TTS) from interactive voice response (IVR) and marrying it more closely to the computer resources that promote natural language understanding (NLU), speech analytics, customer relationship management, call routing and other business processes that are key to quick resolution of a caller’s issues. Speed to market and rapid response to customer queries are the first-order concern of today’s customer care professionals

If you “zoom out” from the enterprise server close up (and even of you keep that focus, but add speech-enabling Web services, rather than IVR apps) you see strong competitors to Nuance/IBM/Loquendo’s product portfolio. Real competition (and source of price/performance pressure) will come from Google, Microsoft and a few others. And if the battle is for market share on a regional level, there are strong, niche-y competitors with a variety of languges and dialects for ASR, TTS and embedded products.

In recent days the leaders of Telecom Italia have confirmed the imminent sale of Loquendo to a large American company (competitor) with a product portfolio and technology that is similar to Loquendo’s but a hugely disproportionate market cap.

The article’s writer seems to point to Nuance as the acquirer by asserting that the prospective acquirer has a history of major acquisitions in the speech technology space (with the further comment that it has led to disposal of assets along with loss of know-how, expertise, patents, etc.

Loquendo was established as a research-oriented subsidiary of Telecom Italia in 2001. Its line of speech processing technologies spans server-based and embedded automated speech recognition, text-to-speech rendering and voice biometrics. As the posting points out, it has over 100 employees and has been profitable since 2004 with revenues of roughly $13 million (US) in 2010.

The timing (of the post) means that this is bound to be the talk of SpeechTek, which is one of two major conferences dedicated solely to speech processing. On the one hand the specter of antitrust and anti-competitiveness rears its head. Loquendo is one of the few core technology providers with as broad a portfolio of software and technology as Nuance. On the other hand, as I will be discussing on a panel with Dan Hong and Bill Meisel on Tuesday, speech processing customarily comprises only part of a product or solution.

There is an “opportunity continuum” that starts with speech detection/acoustic processing, speaker detection/ID, speech recognition/understanding, transcription, translation … and so on — each presenting opportunities on servers, in-the-cloud, on mobile devices, in cars, household appliances and toys.

In each of these domains there are multiple players and specialists. In voice-based search/translation/transcription, Google has a leadership role – along with Microsoft. There are dozens of “players” in speaker verification and identification (which itself is customarily part of a “layered, multifactor” solution. Mobile or embedded resources is an area with a high level of competition and high levels of demand for personalized services that are “conversational” in nature.

I don’t know what the terms and conditions are surrounding the sale of Loquendo by Telecom Italia to Nuance (or another competitor in the speech-enabled world), but I will observe that incumbent telcos (former PTT’s, if you will) have not demonstrated the ability to deliver innovative products or services at a global scale.

As I write these stories, I sense an acceleration of interest in fraud-prevention solutions that include voiceprints. It is driven by growing recognition on the part of “card issuers” (be they banks, government agencies, healthcare providers or retailers with loyalty cards) that the use of mobile devices to carry out commerce brings greater exposure to fraud. It is the sort of fraud that can be prevented by “stronger” user authentication as required. Voiceprints fit the bill because they provide “proof of life” (as VoiceTrust puts in its product literature) and a true second factor (“something you are” as opposed to two instances of “something you know”).

I’m also pleased to announce that I’ll be giving a short presentation on the power of voice biometrics at the upcoming 5th Annual Security Confab, which is a venue for thought-leaders to meet and discuss the developments shaping the security world.

Sensationalism makes headlines, but what is lost in the coverage is that Russia’s largest bank has been testing and is ready to deploy an ATM-like kiosk that will be able to scan passports, record voices, capture fingerprints, perform a 3D facial recognition scan as part of a registration process. The ability to use this data to prevent future fraudulent access went almost unnoticed.

As for the use of speech processing technology from SpeechPro (aka Speech Technology Center or STC), Anna Belash, who heads Marketing Communications at the company explained in a post, that the existing prototype is using speech recognition and text-to-speech synthesis along with speaker recognition to provide a way for individuals to use their voice to navigate through the process of using the ATM. As she said, “it’s a kind of speech-enabled IVR.”

The ATM will employ speaker recognition and authentication if needed “to ensure higher security level while accessing to confidential data.” In other words, the unique deployment for Sberbank – combining speech recognition, speaker recognition and emotion detection is not designed to be spooky or invasive. “It is just one more step forward the creation of humanlike systems that speak to the clients, understand and recognize a speaker.”

What’s different is the incorporation of emotion detection in the enrollment process, which is probably a very good idea if enrollments are going to be conducted without human assistance or supervision. The machine will be able to talk with the prospective enrollee (and later on the client) in plain Russian, and will be able to autenticate his or her unique voiceprint while, at the same time, test voice levels for signs of nervousness, anger, or deceit.

The potential scale of this deployment is impressive. According to the newspaper reports, Sberbank has the largest branch network in Russia with more than 19,100 retail outlets in 3,800 cities. It operates a network of about 28,000 ATMs and 17,000 payment terminals.