presented by ; Md Moshiur Rahman.
Talking to machines: Interactive voice response gets better
Improved technology makes it less likely that you'll get caught in 'touch-tone hell'
Such recorded greetings, inviting a response via the caller's touch-tone telephone keypad, are generated by interactive voice response (IVR) systems, which for two decades have been the principal communications interface between the public and corporate America, supporting self-service applications -- or at least reducing the workload on live call agents.
But these days, IVR systems are changing, leaving less and less likelihood of callers being trapped in "touch-tone hell." More corporations are switching to speech recognition so that callers are greeted by a voice that invites them to simply state their business. Reacting to the words they recognize, these systems route the calls accordingly.
Such an open-ended greeting is called a natural language system, explained Lynda Smith, division manager at Nuance Communications Inc. in Burlington, Mass., which makes the "speech engine" used in many IVRs. (Simpler, menu-structured speech interfaces are called "directed dialog" systems.)
Smith divides speech-based IVRs into four tiers. The lowest tier prompts the user to "press or say 1, and might have a "grammar" (the repertoire of words and phrases it can respond to) of 250 words. Tier 2 would be similar but with a grammar of up to 2,500 utterances. Tier 3 would add a natural language system, and Tier 4 would be capable of handling an open-ended grammar, such as would be needed for a directory look-up application. Prices range from $100,000 to $1 million, she added.
Speech recognition accuracy not an issue
Speech recognition accuracy is, oddly enough, not an issue, since the system can prompt for clarification if it's confused, explained Bob Meisel, telecommunications analyst and head of TMA Associates in Tarzana, Calif. The real issue with IVRs is containment -- how often the callers are able to complete their errands within the IVR application, without aborting the procedure by pressing 0 (or whatever it takes) to get to a live agent.
Smith said that containment rates are better for basic applications such as bank balance inquiries, but overall they can range from 40% to 90% -- assuming that a grammar has been assembled covering every way a caller might ask for the options in question. Meanwhile, the percentage of callers in touch-tone IVRs who immediately try to get to an operator by pressing zero varies from 10% to 40%, and the rate of misrouting ranges from 15% to 35%, she added. Smith claimed that the use of speech cuts the zero-out rate by up to 30% and reduces misdirected calls by up to 50%.
"Most banks were early adopters and had rates of better than 80% using touch-tone," added Mike Moors, director of sales at Genesys Telecommunications Laboratories Inc., an IVR vendor in Daly City, Calif. "A few have moved to speech and have seen a slight increase, but the rate is still in the 80s," he explained. "Health care firms usually see rates in the 15% to 20% range, since people are calling for more complicated reasons, but the system can still gather information about the caller, and 80% to 90% of the callers succeed at giving it within the IVR."
Others were less optimistic. "Normally, the success rate is 25%, rising to 45% or 50% if you put effort into it," said Bern Elliot, an analyst at Gartner Inc. in Stamford, Conn. "If you work at it and motivate people, you might get to 70%, but that is the exception rather than the rule," he added.
Lots of effort required
And considerable effort is required, indicated Skip True, an IVR and Web strategy manager at Chrysler Financial Co. in Farmington Hills, Mich.
"It isn't easy," said True of his migration within the plast year from a touch-tone system (which had an opening menu of seven options) to a speech system. "The biggest advantage of speech is that fewer callers are misdirected," he said. "But a constant effort is needed to edit and tune [the grammars] as opposed to the old touch-tone system, which functioned for years without anyone paying much attention to it.
"We captured 25,000 utterances and used them as a baseline to begin our grammars," said True. Basically, Chrysler Financial recorded calls to a simulated natural-language IVR, which actually used a live agent to route the calls, he recalled. The speech application required nine months to finish, and tuning continues as they encounter callers describing things in unanticipated ways, such as "payment coupon" for "billing payment."
In the end, the containment rate rose 2%. He declined to give the baseline, but customer satisfaction went up on all metrics. Time spent by callers with live agents has actually increased, because it's easier for a caller to get to an agent, he explained. Routine calls are typically contained within the IVR, True also noted
Meanwhile, True's emphasis on customer satisfaction is another trend in the IVR field, as corporations seek to do more than just save money on call agents. "Speech has moved the proposition from cost control to customer satisfaction," said Ken Goldberg, vice president at Dallas-based Intervoice Inc., which credits itself with inventing IVR 24 years ago.
But the cost equation remains compelling, since handling an automated transaction costs 30 cents to complete, as opposed to $6 with a live agent, estimated Elizabeth Herrell, analyst at Forrester Research Inc. in Cambridge, Mass.
"In large call centers, every 1% improvement in self-service can produce millions in savings," added Goldberg.
VoiceXML standard helps out
The other major trend in the field is one that callers won't hear: the advent of industry standards, especially VoiceXML, usually described as serving the same function for speech IVR that HTML serves for the Web.
"Previously, each IVR system had its own proprietary environment, and making changes was difficult and required a dedicated programmer," explained Herrell. "With VoiceXML 2.0 [finalized in 2004], you can make changes on the fly, and there is a wide corps of developers and libraries of reusable components. I've seen some big speech applications completed in 30 days that would have taken six to nine months in the old days, although it still usually takes a couple of months."
While some corporations may still be buying proprietary systems to expand legacy IVRs, new systems are typically sold with VoiceXML, noted Moors. "The game is over -- it's not a choice any more," he said.
Travelocity.com LP's call center in the Dallas suburb of Southlake, Texas, had been using a speech IVR for years, but it decided to migrate to a new one two years ago mostly to take advantage of VoiceXML, explained Ashok Narayanan, manager of software development at Travelocity.
Since then, Narayanan said, "writing an IVR application is like writing a Web application." Travelocity has since been able to mimic several of its Web applications on its IVR, including one that lets callers get copies of their itineraries sent to their mobile phones as text messages.
Narayanan said an application can take two months to create and can involve 10,000 recorded utterances, including city names, airline names, months and numbers.
"We have seen a lot of improvement in calls that terminate in the IVR for a good reason [instead of just being abandoned] or get routed to the correct agent," Narayanan said. "As for the monetary savings, we would not be doing this if the numbers were not to our liking."
Another widely cited trend in IVRs is the increased use of computer telephone integrations (CTI) so that data input by the caller will appear on the agent's screen when the caller finally gets to that agent.
"The leading users recognize that if you want people to do self-service applications, then you have to reward them, and that means getting them to the right agent and having the agent see what they are working on," Elliot said. "If the caller has to start all over after transferring to an agent, what's the point? Next time, they'll just try to get straight to the operator."
Call centers lacking CTI will typically tell the callers that after getting to a live agent, they must repeat their information for "security reasons" or because "the computer is down" -- a practice derided by multiple sources.
Here comes Microsoft
Those call centers may be getting some help from Microsoft Corp., which is seeking to open speech-enabled IVR application development to a much wider market with the introduction of its Office Communications Server 2007, whose Speech Server component offers IVR functionality, noted Albert Kooiman, a senior business development manager at Microsoft.
"Most of our customers are professionals, and the effort required to set up an IVR will be no different from setting up an Exchange server," he said. The list price for the standard edition is $649, he said, adding that the application will also require a server, a router and various software tools. One server should be able to simultaneously handle as many as 200 touch-tone calls, but the number will fall to about 30 with complex speech applications such as directory assistance, he added.
But however they're building them, IVR application developers must also come to terms with a simmering public backlash against automated customer service, Meisel noted. He pointed to the www.gethuman.com Web site, which details how to bypass the IVR menu and go directly to a live agent at the customer service numbers of hundreds of corporations. (It also grades the service that the callers receive -- and most of the IVRs get an F.)
No comments:
Post a Comment