CASE STUDY: Planning, details key to audio ATM rollout

atmmarketplace.com


15 October 2003

    Making an ATM "speak" can be a complex process -- but one that will
become essential to all financial institutions and operators of ATMs in
the United States, according to Timothy Hoyle, a senior consultant with
IRB Consulting Group.

    Hoyle's company is assisting a large Mid-Atlantic and New England
area bank with more than $40 billion in assets with its rollout of
audio-enabled ATMs. Hoyle expects the majority of that Bank's 1,000 ATMs
will "speak" by the spring of 2005 -- at roughly the same time
MasterCard has mandated that all ATMs must be capable of running Triple
DES encryption.

    After discussions with advocacy groups for the visually impaired,
Hoyle said, the bank agreed to pilot 32 audio-enabled ATMs in the fall
of 2002. Shortly afterward, he said, IRB Consulting Group was asked to
develop a plan to mitigate the costs of this large audio rollout.

    The resulting plan outlined the changes required to implement audio,
place ATMs on the bank's wide area network and implement Triple DES at
the same time. In addition, these same capabilities would be implemented
as a part of the bank's regular ATM attrition program.

    Some upgrades - a faster processor and more memory -- are required
for both audio capability and Triple DES. "You're generally going to
need a Pentium II or better processor and at least 256 megabytes of
RAM," Hoyle said.

    Other key components include a sound card (for audio) and an
Encrypting PIN Pad (EPP) for Triple DES.

    Hoyle said that the bank also decided to make some infrastructure
changes from which it expects to reap financial benefits, chiefly an
upgrade to an Internet Protocol (TCP/IP) network at its branches, again
timed to coincide with Triple DES deadlines. The switch will save his
client "millions of dollars" a year, he said.

    With a project as large as this, Hoyle said it was essential that
his client be able to leverage the impact of the changes. One such
benefit that resulted from the implementation of audio was a
streamlining of the transaction sets used by the bank.

    "There were hundreds of ATM images in use at the bank," Hoyle said.
With the implementation of voice guidance, the bank will offer more
consistency to its customers across the network. Working with Concord
EFS, which drives the bank's ATMs, Hoyle said the bank reduced its
possible screen configurations from more than 100 to roughly a dozen.

    Evaluating the technological "big picture" while trying to keep
costs to a minimum was crucial, said Hoyle, who estimated that the
upgrade's final cost will be more than $10 million.

    IRB Consulting Group has helped its client roll out more than 180
audio-enabled ATMs since last December and plans to add another 150 or
more in 2003. By the end of 2005, he expects the bank will have at least
one audio enabled ATM in each of its approximately 600 branches and at
170 non-branch locations.

    Catching the WAV

    The ATMs will "speak" both English and Spanish. Offering two
languages will become less complicated, Hoyle said, when the bank
introduces text-to-speech software, currently expected in early 2004.
Today, voice guidance is provided through the use of WAV files, a
process that presents management and implementation challenges for the
bank, he said.

    WAV files do not support dynamic content changes; a visit to the ATM
is required to make any change, including one as simple as a surcharge
amount. "That's a big deal when you're talking about 1,000 machines,"
Hoyle said.

    Text-to-speech software uses a computer-generated synthetic "voice"
rather than a human one. While the sound isn't as natural, it offers a
less costly and simpler alternative to WAV files, Hoyle said. Changes
can be made dynamically and implemented without a visit to each ATM.

    Hoyle recommended the bank use text-to-speech when it introduced
audio, but Concord does not yet support it for its primary ATM vendor,
Diebold. Kevin Carroll, Concord's director of ATM services, said
additional development is required because Diebold ATMs will continue to
use states-and-screens in a Windows environment. Concord expects to
offer support for text-to-speech on Diebold machines by the end of 2003.

    In the interim, Hoyle noted that the bank decided to have Concord
use text-to-speech software to generate WAV files. This resulted is a
more natural-sounding voice that can run in an OS/2 environment -- but
without the costs of hiring human "talent" and recording them in a
studio.

    Concord develops a "script" of ATM transactions for its clients,
Carroll said, then runs it through a text-to-speech engine and produces
a CD of the WAV files. The client reviews the CD in either their own
test lab or Concord's and makes any necessary changes before a final
version is produced.

    Picturing it with words

    Both Hoyle and Carroll said that writing scripts is a challenge,
largely because they must be modified for each make and model of ATM in
a bank's network. "The cash presenter is in one place on a Diebold 1062
and in another on a 1072," Hoyle said.

    Script writing presented more fundamental difficulties as well.
"We'd start out saying, 'Insert your card into the card reader' and then
realize that the user didn't know where the card reader was," Hoyle
said.

    At the suggestion of various advocacy groups for the visually
impaired, Hoyle noted that the bank decided to use the concept of a
clock face. Vendors created drawings for each of their ATM models,
showing where the peripherals would appear on a clock face: card reader
at one o'clock, 6 inches from the keypad, for example.

    The logistics

    Upgrades for both audio and Triple DES present a major logistical
challenge, Hoyle said. Updating site information, including which make
and model of ATM was deployed at each location was critical. Diebold and
NCR, the bank's ATM vendors, helped with site visits. "What we had on
file wasn't always accurate," Hoyle said.

    IRB Consulting also required vendors to ensure that service
personnel were adequately trained on adding audio capability. Early
upgrades took up to 10 hours, but the time was reduced to three to four
hours following an increased emphasis on training, Hoyle said.

    Service technicians must contact a member of the bank's ATM group
before work begins and when it is completed so that the bank can monitor
each upgrade. Working with the bank and its ATM vendors, IRB Consulting
Group developed a checklist to ensure that an ATM's audio capability is
in good working order. When the technician completes his work, a branch
employee puts on an audio headset and runs through voice-guided
transactions with the checklist - before the technician leaves.

    Early upgrades allowed the project team to work out the bugs -- not
scheduling work on Fridays, for example, to lessen the chance of any
problems occurring over a weekend. The bank now has it down to a
near-science, Hoyle said. "We've got five or six upgrades scheduled for
each day this fall."

    For a PDF version of this article, click here:
http://www.atmmarketplace.com/pdf/IRB_CaseStudy_03.pdf


VICUG-L is the Visually Impaired Computer User Group List.
To join or leave the list, send a message to
[log in to unmask]  In the body of the message, simply type
"subscribe vicug-l" or "unsubscribe vicug-l" without the quotations.
 VICUG-L is archived on the World Wide Web at
http://maelstrom.stjohns.edu/archives/vicug-l.html