Since the introduction of Interactive Voice Response (IVR) technology over 10 years
ago, many revolutionary telephony-based products have infiltrated the marketplace. Today,
some of the most popular uses of IVR systems are with automated attendant applications,
call routing, and information retrieval. This new generation of IVR systems allows a
caller to have voice-activated access to large databases of company information. For
example, if you want to call an employee in a large organization, you need only to pick up
the phone and speak a name. Gone are the days of having to dial by name and fumbling at
your numeric keypad.
These new IVR systems have provided a tremendous improvement to the industry. However,
there are still some drawbacks. Most IVR systems are set up to only listen for certain
bits of information. Once heard, they process the information, search a database, attempt
to retrieve the correct data, and produce the desired result (i.e., connecting the caller
to Steve Dailey). The IVR systems only need to ascertain a small amount of information
before making an "educated" guess and connecting the call. But, what if the
database held a record for a Steve Daley or three people named Steve Dailey? Would the
computer be able to recognize which Steve Dailey the caller requested? Or, would it find
the first, closest fit and then connect the caller? Unfortunately, the latter is most
often the case, leading to an unacceptable percentage of incorrect connections. The only
way to combat this problem is by utilizing a new technology that incorporates a
conversational-style search.
The Conversational Search
A conversational-style search is the most natural method for conducting search and
retrieval functions in voice-driven auto attendant, call routing, and information
retrieval applications. This technique provides a caller with the ability to obtain the
desired result through a natural dialog similar to that of a live operator. The
conversational process takes the caller through a series of questions to ensure easy
access to the caller's desired information.
The ideal conversational-style search product should contain certain features to ensure
accuracy. Sophisticated speech recognition techniques should feature the ability to search
large databases of information on the basis of speech input, accomplished while conducting
an on-going, intelligent conversation with the caller. Let's again use the example of
having two people named Steve Dailey in a directory. The typical IVR system would most
likely look for the first Steve Dailey it found and connect the call. A conversational IVR
system is intelligent enough to know that there are two matches, and it will ask a
question to further qualify and ascertain the correct match. The following example
demonstrates this system in action:
- Conversational System Operator: Thank you for calling Main Street
Technology. How may I direct your call?
- Caller: May I speak with Steve Dailey, please?
- CSO: Did you say Steve Dailey?
- Caller: Yes.
- CSO: Did you want Steve Dailey in accounting?
- Caller: No.
- CSO: Did you want Steve Dailey in receiving?
- Caller: Yes.
- CSO: Please hold while I connect you.
The system ascertained that there was more than one Steve Dailey in the database. By
asking a few more questions, the system was able to identify the correct Steve Dailey and
connect the call.
A conversational IVR system also has a unique built-in spelling feature to further
enhance the probability of finding a successful match. What is the first question live
operators asks you when they are having trouble understanding you? Don't they say (and
don't you expect) "How do you spell that?" Another good example of a technique
that a conversational IVR system uses to solve a real problem is when a company has two
employees whose names sound exactly the same, but are spelled differently. For example:
Steve Dailey and Steve Daley. The spell-by-name feature and the operator-like conversation
allows you to be transferred to the correct individual. A typical voice recognition
product would be stumped and would either transfer you to a live operator or give you a
choice of two names that sound identical.
Another conversational feature that a traditional IVR system could not provide is the
ability to tell when a name is not in the database (also known as "garbage
rejection" or "false positives"). Conversational voice-driven IVR systems
have fine-tuned their searching technique to a point where they can conduct a conversation
with the caller in order to "not" find someone that does not exist in the
database. Again, speech-based IVR systems relying on traditional technology will keep
guessing at the name until the caller gets frustrated and either hangs up or transfers to
the live operator.
Conversation In Action
A conversational-style directory search product is ideal for speech-enabling a number of
applications: in corporations for auto attendant, corporate directory assistance, and
personal directory services; in telephone companies for public directory assistance and
frequently-called numbers services; and in call centers for speech-enabled transactions
and for intelligent front-ends.
Corporations utilizing a conversational-style, speech-enabled auto attendant and
directory assistance system realize numerous benefits. Most incoming calls can be handled
by the automated system, thereby providing 7 x 24 coverage to ensure that their employees
and customers around the world can quickly and easily contact anyone in the directory at
all hours of the day simply by speaking their name. It can also help decrease the
operational cost issues related to adding additional staff to augment over-burdened
operators.
Moreover, conversational IVRs eliminate the need for a printed internal directory.
Employees can pick up their handset, say the name of the person to whom they wish to
speak, and be connected -- quickly and easily. Considering that company rosters change on
a consistent basis, paper directories are out-of-date even before they are printed. With
this new system, the database of information can be updated immediately when a change
occurs. A tremendous cost saving is achieved with the elimination of printing the
directory, and a better class of service is offered to the employees.
Telephone companies are ideal locations for automated directory assistance systems,
because the number of live operators can be significantly reduced (i.e., a cost savings is
achieved). Additionally, consumers can now be guaranteed that they will receive the
correct phone number. How many of us have received a number through directory service,
called the location and determined it was the wrong number? With the conversational,
highly accurate, search solution, incorrect answers are a thing of the past.
In the U.S. market, customer service is a critical factor for success. Consumers are
constantly looking for the easiest, fastest way to obtain order, product, and price
information. With an automated system, call centers can ensure customer service is never
compromised. Speech-enabled solutions may be deployed in the call center environment in
two basic ways: providing assistance with simple transactions (like providing the status
of an order, quoting price, and availability of products, etc.); and serving as a
front-end to accumulate information from the caller and then route the caller to the
agent, saving precious agent time. Corresponding benefits to these features include
providing better customer service, allowing agents to be more productive and handle more
complicated customer service issues, and eliminating wait time imposed on the caller --
all of which save a company money and reduce caller aggravation.
With greater emphasis on providing better customer service, companies worry about the
extra expense needed to provide the service customers are demanding. Conversational IVR
systems help bridge this gap. Whether utilized at a large organization, telephone company
or call center, a conversational-style, speech-enabled search solution is a large factor
in achieving increased productivity, cost savings, and exemplary customer service. This
new type of technology is paving the new speech-enabled IVR road of the future.
Mark Bannon is the vice president, Sales, Marketing and Technical Services of
Phonetic Systems Inc. Phonetic Systems is a leader in the innovation, development, and
deployment of state-of-the-art, conversational style, speech-enabled, telephony-based
directory search solutions. www.phoneticsystems.com. |