Speech Technology Magazine

 

What Happened to Voice Portals?

What is – or, more accurately – what was a voice portal? One definition offered by the “Computing Dictionary” describes voice portals as “an interactive voice response (IVR) front end to a data retrieval system … being positioned mainly by carriers as wide-scale applications, offering things like specialized audio ‘infotainment’ services, real-time news, and stock quotes.” These sites offered engaging speech-enabled dialog available 24/7 from any standard telephone. …
By Steve Chirokas - Posted Aug 30, 2005
Page1 of 1
Bookmark and Share

What is – or, more accurately – what was a voice portal?

One definition offered by the “Computing Dictionary” describes voice portals as “an interactive voice response (IVR) front end to a data retrieval system … being positioned mainly by carriers as wide-scale applications, offering things like specialized audio ‘infotainment’ services, real-time news, and stock quotes.” These sites offered engaging speech-enabled dialog available 24/7 from any standard telephone.

I remember working at a speech technology vendor in early 2000, when the topic of voice portals couldn’t have been hotter. Fledgling companies such as Quack, iDine, FoodLine, TalkingWeb, TelSurf, and PhoneRun offered exciting new ways of accessing information simply by using the telephone, a prospect that seemed ready to burst onto the scene. Voice portals were popping up with great frequency, characterized by a Web-like business model where the goal was to capture “ears” and hope that traffic volume would sustain an advertising-based revenue stream. These sites provided restaurant reviews, traffic updates, and other information potentially attractive to a large audience.

The trade media and industry analysts were on board as well. In going back through my notes, I found analyst comments such as this one: “By 2005, more people will surf the Web from phones than PCs,” and voice-based services will “humanize the Internet.” There was a concept of a speech-based Internet, seamlessly linking data and organizations with speech recognition and telephony. Hey, we were on a mission!

What happened to the voice portal bubble? Most of the companies established to deliver Web-like services in a speech format either faded away or morphed into something else. It also looks like the most popular current form of Web access is via computer and not the phone. So was the “era of the voice portal” a blip on the screen of failed technologies/business models?  

The answer is a resounding “no.” These trailblazing efforts helped shape the speech services that are exceptionally successful today. If there was a mistake, it was assuming that callers would use the same method to access information on their phones as they do with their computers. Many industry pioneers thought you could simply use the same pool of data (i.e., the Internet) and present it either on screen or in audio format without much modification.

I recall working, for example, with a product that delivered advertising over the phone just like a pop-up ad would on the Internet. It didn’t fly.

Voice Portals Move to the Enterprise

So what happened to voice portals?

The original sponsorship-based voice portal business model never achieved profitability. Telephony infotainment simply did not deliver enough value to drive sustainable revenue.

However, lessons learned, especially in areas such as call flow and voice user interface design, have carried over to commercial Web sites and “customer portals” that are far more focused on specific tasks. The challenge now is to ensure that callers are engaged in a dialog that allows them to complete a task and also encourages them to call back. In many respects, the measure for success is close to what the original voice portals sought: a growing list of callers, including a healthy base of repeat users. The difference is that the ultimate metric is a combination of customer satisfaction and savings from automation, not just call volume.

Speech deployments in the travel and hospitality industry are good places to start to look for what voice portals have become today. The operative word is “focus.”

Travelers can now access automated speech applications over the telephone to book a flight, train, rental car, or hotel room. Primary customer benefits include 24-7 availability and the ability to make or to confirm reservations without waiting to speak with an agent. While these customer service solutions might not necessarily be considered traditional “portals” because of their fairly narrow focus, they deliver value for callers and organizations that deploy them.

For example, a major airline today allows callers to check flight schedules and book flights over the telephone as easily as they could with either an agent or through a Web site. Not only has the carrier achieved its customer satisfaction goals, but it also has reaped the rewards of automation by requiring fewer trained agents.

How do organizations deploying speech systems know they are succeeding? Following in the steps of webmasters who used graphical applications, such as WebTrends to track traffic patterns, page popularity, and other metrics, speech vendors are implementing the same type of reporting capabilities for speech deployments.

In many cases, organizations want to monitor overall performance while also profiling traffic patterns to ensure they are delivering exceptional service to their callers.

As an example, a Convergys Corporation client in the insurance industry utilizes a Web-based Business Information Portal (BIP) tool to track a host of caller and performance metrics. This company not only verifies that callers are completing their tasks from statistical data, but it also uses information derived from the BIP to influence decisions in other parts of the business. Speech meets the Web again.

Early voice portal developers also explored the use of “cookies” to track usage. The hope was that cookies, which identify repeat visits and customize pages on Internet sites, could perform similar tasks in a speech setting. But given their reliance on text files -- something that is difficult to put into effect with a telephone -- cookies were never successfully deployed as part of a voice portal.

The ANI Picks up Where the Cookie Left off

The cookie concept returns through the innovative combination of automatic number identification (ANI) and dynamic grammars. Such technological advancements now make it possible to personalize calls based on the caller’s location and other criteria. An ATM locator service, for example, would only provide callers with addresses for nearby machines.

Additional refinements (e.g., the ability to recognize family names and access health care-related information) enable speech systems to more quickly and accurately determine what the customer wants to accomplish and respond appropriately. Not exactly what voice portals originally set out to accomplish, but the ends are targeted toward voice portal goals that always included customer satisfaction and greater use of speech systems.

Completing sophisticated transactions by utilizing speech was not a priority for the first voice portals. Speech solutions have since evolved to incorporate complex back-end system integration with security and performance considerations that are analogous to the Internet. Early voice portals started to broach the transaction barrier with automated reservations. Speech solutions deployed for financial institutions now regularly allow callers to check balances, transfer funds, and execute a wide range of transactions.

As such, security has become an important consideration as account numbers and other personal information could be illegally obtained by accessing stored audio files or via speech logging data. A number of these issues are quite similar when comparing speech solutions to the challenges faced by those deploying Web-based applications.

Emerging speech technologies such as natural language understanding (NLU) give new meaning to the portal concept. Designed to support open-ended voice user interfaces, NLU is an excellent solution for handling the initial contact with a caller, especially when callers are uncertain of how to phrase their requests and unclear on the specific service they need.

In this context, an NLU solution provides the “portal” into the organization for everything from basic call routing to technical support. Benefits range from reducing the number of toll-free numbers required to support customers to making it easier for callers to complete their tasks without navigating several menus and adhering to strict grammar requirements.

While the original voice portals may not have been dead-on in their effort to make telephony Web surfing and infotainment a part of every day life, they did contribute greatly toward the evolution of speech within the enterprise. Early voice portals helped to bridge the gap between the Internet world and speech-enabled telephony along with the emergence of the VoiceXML standard.

In the final analysis, voice portals never really disappeared. A sharpened focus, improved technology, and a Web-like infrastructure for speech simply combined to make the voice portal the enterprise speech solution as we know it today.


Steve Chirokas is senior director of products and channels for Convergys Corp. His responsibilities include product management for the Speech Solutions Group and managing channel relationships with organizations that are partnering with Convergys to leverage the value of their speech applications within a hosted infrastructure. Prior to Convergys, Chirokas was director of product and partner marketing at ScanSoft and has held previous marketing positions with Process Software, Compaq, Brooktrout and Scitex.

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