Listen Up! Big Brother is Talking to You
Listen up! The government is talking to you and being helpful in the process. This might sound strange to anyone who has sat on hold waiting for a government agency representative. Ever called the Internal Revenue Service (IRS), Immigration and Naturalization Service (INS) or Department of Motor Vehicles (DMV)? You know what I mean. The cost saving measures forced onto cash-strapped agencies have driven many of these organizations to implement touch-tone IVR systems to handle the crush of calls, with more or less success. Whereas some of these applications are simplistic and easy to use - taking care of business - other applications are not so hot. With some I am reminded of the classic Married with Children TV show episode in which Al Bundy spends the entire show tunneling through the touch-tone application from hell, only to be put on hold and have his daughter hang up the phone when he fell asleep waiting for a CSR. The good news is that increasingly, great applications are being implemented using speech technologies, which are changing the way people interact with the government. "The government" is a broad body of hundreds of organizations and branches, which in turn represent hundreds of opportunities for speech. From the armed services to law enforcement and social services, new applications are emerging. In fact while researching this piece, I found applications in all government areas ranging from really customer facing applications such as the DMV to clandestine usages of speech technology. For the latter, it is nice to see the government fund research into the development of speech, and nicer that they are now a big consumer of commercial grade products. However, it is not so pleasurable to tell the reading audience that there are some really cool applications I just can't talk about. For the purposes of this article, let's just say that there were instances of many governments using speech, but we will focus on the US government, and that the term responsive applies to taking care of the needs of citizens, whether that be protecting them or personally helping them. As such, to see how speech recognition is helping these government organizations be more responsive to their customers, it is easier to look at applications by type, rather than by the type of government body using them. The Speech Recognition Beachhead - Auto-attendant Speech-driven auto-attendants providing call routing and directory services are a classic speech recognition beachhead for vendors selling into vertical markets - government included. As a proof of concept for speech, call routing is a great way to show how the technology works and how it can save money and resources. Reducing the number of operators to answer calls, providing after hours service, or eliminating the mundane job and associated costs of printing directories, positively affects a company's bottom line, which often leads to repeat business for the speech vendor. In fact it is perhaps one of the clearest ROI stories in speech as directory printing costs can be excessive and dial-by-name DTMF applications in large enterprises cumbersome. In the government sector for example, Phonetic Systems has installed directory services applications in all levels of government from the State of New Jersey and Orange County (Florida), to Westchester County (New York), as has iVoice with an auto-attendant for the Borough of Leonia in New Jersey. Similarly, Nuance, in conjunction with Amcom Software and General Dynamics Networks Systems announced a deal to implement auto-attendant and directory services for 29 U.S. Army bases. Information Access Directory services applications can be as simple as routing a caller to a person, or can be the entry point to an information portal for services as well. Following the lead of the Y2K voice portal craze, governments have been quick in providing portals of their own. These portals either provide access to personal information or generic information of a time-sensitive nature. Most notable are local and national 511 services that consumers can use to get real-time traffic and road condition reports. When the Federal Communications Commission (FCC) granted exclusive use of the 511 phone number to state and local transportation departments for use in traveler information systems more than a dozen states sign up, with more opting to implement 511 services each month. For example, Avaya and Nuance, in conjunction with Interactive Northwest developed a 511 service in the State of Washington, which provides drivers with continuously updated roadway incident and construction information, express lane status, weather and mountain pass road conditions. Nuance has also done others, including one for the San Francisco Bay Area Metropolitan Transportation Commission in partnership with PB Farradyne. These information services portals use both automatic speech recognition (ASR) and text-to-speech (TTS) technologies, as the amount of information tends to be both large and variable. A good example of TTS output is the U.S. Office of Personnel Management (OPM) application that uses ScanSoft's TTS to read job postings to applicants on OPM's USAJobs.com website. This site provides job seekers with access to thousands of government and non government jobs. Customer Service The government has plenty of situations that force a customer to stand in line or sit on hold, ostensibly to get or give information. Whether it is a call center agent or a desk clerk, often the person handling "customer service" has a heavy case load. As such, any process that involves permits, appointments or renewals have been targets for speech vendors. For example, SpeechWorks along with partner, Frank Solutions, have implemented a number of such applications, including a system for the city of Portsmouth, VA which allows customers to get information on building inspections, utilities and real estate transactions. They have also speech-enabled part of the call center for the State of Alabama's Medicaid office, allowing customers to check on their claim status and verify insurance coverage. In a similar vein, Nuance, along with the Georgia Technology Authority, created an automated license renewal system for the Georgia State Department of Motor Vehicles, relieving people of the burden of waiting in line at the DMV. Income tax is another particular area of broad concern, resulting in numerous speech-enabled applications, including the Washington State Department of Revenue and Fulton County, GA (SpeechWorks/Frank Solutions), with a business tax filing and tax information hotline, and the Michigan Department of Treasury (SpeechWorks/Intervoice) with a hotline for tax inquiries, payment schedules and tax Frequently Asked Questions (FAQs). The U.S. judicial system, in particular, has used ASR to shorten ACD queues, and allow customer access to information over the phone. Jury selection is being done by the Philadelphia County Jury Selection Committee and the State of California, which automates the process of calling in to see if a person is required to show up for jury duty. In Santa Barbara County, CA applications for both the phone and Web are being implemented to allow for 24-hour access to the county's Case Management database, including information on traffic citations, trial and verdict information (SpeechWorks/Frank Solutions/Artisoft). The system will allow not only access to this information, but will enable citizens to pay for fines with a credit card, rather than mail in the payment. Speeding up the process of government hiring is another cost saving application. The State of Wyoming, for example, employed Syntellect and Nuance to design call center help desk for employment inquiries. Security A tantalizing array of applications are appearing in the area of security, and are being developed along two fronts - over-the-phone ASR and voice verification and identification. First, following the previously mentioned applications law enforcement agencies are now using speech-enabled applications to be responsive to, and protect, the average citizen. For example, Appriss has developed an application for the Jefferson County Police. Their Call2Court time and attendance application uses Nuance Verifier voice verification engine and a badge number for automated court check in. The same application can also be used to provide notification calls to defendants, witnesses, and jurors about scheduled court times. Another such application by Appriss for the U.S. Department of Justice, as well as local and state govermments in over 1,200 counties in 36 states, is a victim notification system for notifying victims when a prisoner has been released from prison. iVoice has numerous installed applications along these lines including one for the General Sessions Court of Shelby County Tennessee, which provides information to lawyers and citizens on court cases, such as costs to file, when cases are set, case outcome and how much money is due or payable on a case. Each of these applications has shortened the time it takes to get or receive information that might be of a critical nature to protecting citizens. The correctional industry is a lucrative market for speech technologies, particularly voice verification. T-NETIX, a provider of telecommunications products and services to the corrections communications marketplace, has built a business around using voice for security and fraud control to over 1400 correctional facilities nationwide. Their SECUREvoice verification product (formerly called PIN-LOCK) uses voice prints in addition to PINs to positively identify inmates making outbound calls, preventing inmates from placing calls to numbers they have been barred from calling. For example, the Arizona State Department of Corrections has deployed SECUREvoice in 24 facilities housing in excess of 27 thousand inmates. Additionally, companies such as T-NETIX, Persay and Nuance technology-based ShadowTrak, for example, have developed applications which enable passive monitoring of parolees in community-based supervision programs to make sure they are where they should be. Second, are the applications that we don't hear about every day, particularly in light of recent terrorist activities. Homeland Defense agencies, in particular are looking at technologies such as those that use speech, to help in intelligence gathering to track and catch "the bad guys." The government is using ASR to "data mine" intercepted telephone calls for key words, used in unlawful activities, and is stepping up their use of voice identification in these types of applications to identify the speaker in recorded conversations as well. In addition, companies such as Fast-Talk use audio search technology, based on phonemes, to capture names, phrases or words, specific to different unlawful activities, in voice recordings. The government, who was one of the early supporters and adopters of infant speech technologies, has now become a big consumer of these technologies. Whether these applications show responsiveness to a problem that impacts consumers in general, such as long queue times to reach someone on the phone, or long lines in person, or whether the response is due to keeping the population safe, speech technologies are being used to provide real solutions to knotty problems. And just as important to the speech industry in shaky economic times, it is nice to know that this industry is being viewed as a positive solution worth investing in now, not when government prognosticators predict the economy will flourish again.
Nancy Jamison is the principal analyst at Jamison Consulting. She can be reched at firstname.lastname@example.org
Companies and Suppliers Mentioned