1. Ubiquitous Natural Language Understanding
On December 2, Yahoo! bought SkyPhrase, a start-up specializing in Natural Language Understanding, specifically tuned to let people ask about sports statistics. While many analyst saw the acquisition as a big step toward Yahoo adding a humanlike interface that might extend the service life of the company’s mail, news and financial information resources, it is most likely that Yahoo will use the new IP to help its fantasy football users to say something like, “How well does Adrian Peterson do in domes in November?” and get immediate, meaningful answers.
SkyPhrase will also enable fantasy players to receive text alerts when one of their players exceeds a designated threshold. A person only has to say, “Text me when Adrian Peterson rushes for more than 100 yards,” and it is done. Rather than being a “boil the ocean” incidence, whereby Yahoo adds a spoken interface as an alternative for interacting with its far-flung properties, the SkyPhrase acquisition is likely to lead to quite a few more “Speechable Moments;” those occasions in which people learn how powerful spoken words are for getting exactly what they want out of a complex system. E.g. AP’s stats, stat!
2. Virtual Assistants and Virtual Advisors
Like Yahoo!, search giant Google is destined to make huge investments into recognizing speakers’ intent as they enter search terms – regardless of whether they are talking or typing into PCs, phones or tablets. By the end of 2014, they will have conditioned millions of people to expect good things to happen when they say something into a browser-based device. Thus Google is putting the pressure on Apple, Microsoft, IBM and any other firm or individual that is trying to one up Siri or GoogleNow.
Intelligent agents on mobile devices, closely mated to enterprise knowledge management systems, will serve as the unifying force to support secure, personalized, multi-channel customer care.
3. Short-form and Long-form Dialogs Co-Exist
Google’s approach is far from “conversational.” It will continue to raise the bar for accurate rendering of brief utterances (I call it “barking”), but it does not appear to be investing in the sort of natural language understanding that enables people to speak in their own words and expect a spoken response. Apple’s Siri, by comparison, has continually added the ability to carry on a basic conversation with its users. the world of general search and device control, the battle is between long-form, human-like utterances and short form, “barking” with anticipatory search and thought completion
4. The Cloud Interacting with Embedded Resources
The result is “hybrid architectures.” Solutions providers need to sort out what’s “on the device” versus what’s in the cloud. The first order problems involve device “wake up” and application initiation. Also what can be done in a disconnected state. Still, solutions providers have a lot of work to do when it comes to defining how much information processing (such as speech recognition and speaker verification) should take place using chips or software inside a device, versus what naturally takes place “in the cloud.”
In 2014, expect there to be a number of advancements announced by the likes of Nuance and IBM, along with partners in automobile electronics, mobile device manufacturing and consumer electronics as they find the perfect balance between highly-responsive, personalized, always-on resources that are embedded in physical devices and highly accurate, multi-modal platforms that rely on the sort of computer power and memory that can only reside in large data centers connected to the Internet.
5. Personalization Drives Innovation
This feels like an old trend, but the next step is “personalization” based on rapid, confident authentication of the person in possession or in near proximity to a mobile device, automobile, point-of-sale terminal or even a “smart” TV. Our voices are unique identifiers and are totally underutilized as such. In 2014, thanks to the marketing efforts of consumer electronics manufacturers like Samsung and LG, as well as video programming distributors like Comcast and DirectTV, people are learning how they can use their voice and TV remote controls to cut-through the clutter of multichannel offerings.
In the next two years they, along with financial services companies, retailers and others will learn to add voice biometric-based speaker identification along with speech recognition to deliver a highly personalized, results oriented user interface.
6. Triumph of Passive Authentication
Using few words to navigate a complex menu of video programming is starting to be child’s play. Rapid recognition of who an individual is, as well as gleaning his or her intent will rely on a number of computing “engines,” acting as recognizers, interpreters, assistants and, ultimately, trusted advisors. Service and content providers are learning that personalized offerings rely on high level of confidence that the individual taking voice control of the phone, tablet, TV remote or other device matches his or her claimed identity.
Establishing a platform for trusted conversations and commerce has been the goal of service providers for decades (if not centuries). When Opus Research began investigating voice biometrics, it was under the belief that prompting a person through the process of creating a voiceprint wold be step one in establishing a trust framework. In 2013, we’re seeing an acceleration of interest in “passive enrollment” of voiceprints and “passive authentication” of individuals.
There are already a number of back-end resources charged with understanding context, assessing risk and applying business rules to help make multi-channel conversations productive for both parties. Steady improvement will carry on in 2014.
7. Enterprises Embrace Secure Assisted Self-Service
In 2014, we expect the trend toward enrolling and authenticating an individual to continue to move “to the background” so that retailers, banks and telecom carriers, not to mention government agencies can, at a minimum, detect fraudsters or, more to the point, quickly get the “go ahead” to carry out the instructions of a trusted individual. That translates into simple, speedy self-service and assisted self service. The giants of enterprise network security – like IBM, RSA (division of EMC), Cisco and their partners – have been building engines that take context and risk into account and apply the appropriate level of security to foster convenience as well as protection. They are joined by both giants and aggressive upstarts in contact center monitoring and performance optimization, led by Nice, Verint and their partners.
8. Here Come Conversational Cars
Finally! Speech-activated services in cars has become commonplace. While this has given rise to concerns about distracted drivers, it also reflects the fact that the manufacturers of automotive electronics, as well as the auto makers themselves are looking to differentiate their offerings with hands-free control and robust interfaces for smartphones to support popular apps.
There are still some glitches surrounding place names and points of interest, but great strides have been made in overcoming noise in the passenger compartment and isolating the voice of particular individuals to support personalized entertainment, navigation and communications services.
9. Authentication Becomes a First-Order Concern
In 2013, the need for strong authentication, data security was closely linked to privacy and fear of identity theft. But use of popular social media by individuals and a general migration of enterprise back-end systems (CRM, ERP, UC…) to “The Cloud” demonstrate that concerns about privacy and security are trumped by human nature and business economics.
In 2014, there will be much hand-wringing and outcry against privacy breaches, but it will not stop the overall trends described above: Hundreds of millions of individuals using mobile devices with personal data and apps interacting with computer resources and Web services that live in The Cloud. Authenticating the person in possession of those mobile devices will be the first order concern of enterprises and service providers.
10. Customers Take Control… Effortlessly
The whole point of “Conversational Commerce” is to put customers in command of the devices they use and in charge of the relationships they have with their selected vendors. It starts with strong assertion of identity and authenticity and moves from there.
Improvements in the accuracy of speech recognition, elegantly augmented by Natural Language Processing, Artificial Intelligence and Conversation Management resources in the cloud are making the objective of customer control a reality. Most of the heavy lifting (computationally) is carried out “in the background” without requiring customers to exert much effort to take control and achieve their goals.
Happy new year!