Tapping Synergy Between
Chatterbots And Live Agents
Meet Andrette. She's your ideal customer service representative: smart, helpful, and
demonstrative. She likes "knock knock" jokes. She's also a software robot, or
Chatterbots, as defined by Tom Rearick, the president of Big Science Company, are software robots with a
conversational, text-based interface. They're programs that can "converse" with
humans through text chat.
Regardless of her droid status, Andrette can take you on a Web site tour, answer
questions about company history, and point you to product specs. But what if your next
question stumps her? You want immediate help, and you don't want to talk to a software
program. What you want is to speak to a real, live call center agent. Enter, FaceTime Communications.
This week, Big Science announced a partnership with FaceTime that combines automated
customer support through Big Science's Klones (chatterbots like Andrette) with live
operators using FaceTime's flagship product, Message Exchange.
Big Science Lets Klones Do The Chatting
The brain behind Andrette is Big Science's Klone Server. Klone Server is middleware that
works with any operating system supporting Java 1.1 and a Web server with support for Java
servlets. This includes Windows NT 4.0, Unix (Solaris, HP-UX, SCO, and Linux), Netware,
Mac, OS/400, OS/390, OS/2, and AIX on RS/6000. The list of Web servers is equally
expansive, including Apache, Microsoft IIS, Lotus Domino, and Netscape. Just one Pentium
II processor in a server can power more than 500 simultaneous conversations.
Klones, which can be thought of as knowledge worker avatars, can:
- answer simple questions across subject domains as broad as your company's knowledge
- apply process knowledge to infer customer goals and accomplish them;
- access back office processes on the customer's behalf;
- solve complex problems within narrow subject domains; and
- escalate problems (including the question and answer session that's already transpired
between Klone and customer) to a call tracking system, telephone call-back, instant
messaging client chat, or e-mail address.
All end users need is a frames-enabled Web browser. By default, applications don't
necessarily set cookies or download embedded script, although they can. Each interaction
with the Klone typically involves a total end user refresh of less than 12 K worth of HTML
code and .gif or .jpg graphics.
The operation and performance of chatterbot applications can be monitored in real-time
via the Web by call center managers or Webmasters. You'll have transcripts of
Klone/customer interactions, a list of the top 20 questions asked, and other traditional
call center performance metrics at your fingertips.
For more information on the technology behind the Klone Server, read Big Science's white paper. To
talk to Andrette, the Klone on Big Science's Web site, go to www.bigscience.com/.
FaceTime's Live Interface To The Web
FaceTime is an ASP, or application service provider. There are a number of benefits to
using an ASP as opposed to buying the technology outright and hosting the solution
yourself: lower capital outlay, lower risk, and much lower maintenance. Scott Sherman, a
FaceTime sales rep, sums it up: using an ASP "takes all the fear out of something
The FaceTime Message Exchange has two services, which can be purchased separately or
FaceTime Instant Messaging allows customers to click a button on your Web site to initiate
a text chat with a live call center agent. During the interaction, both the agent and
customer can exchange text messages, and can embed active hyperlinks within those
messages. Agents can also pop open the customer's browser and direct them to a particular
Web page that's relevant to their discussion. FaceTime's messaging is compatible with all
Yahoo!, AOL, and MSN's chat systems, too.
FaceTime Instant Messaging comprises three components: the FaceTime Messenger that the
customer uses, the FaceTime Agent Workstation which manages conversations between the
agent and customers, and FaceTime tools to administer, monitor, and report usage of the
FaceTime E-mail Management services are similar to the offerings of other e-mail ACD
programs on the market. With FaceTime Message Exchange, all incoming customer e-mail is
tracked, routed, and catalogued for later retrieval. You can eliminate lost or languishing
e-mail by setting specific routing criteria that automatically re-directs unanswered
messages after a set time period based on your settings.
This module covers four basic areas of e-mail management:
- Sends customers an automatic receipt acknowledging their e-mail
- Routes incoming messages and maintains conversation "threads"
- Manages service levels so you can prioritize replies and set alerts when service level
objectives are not being met
- Catalogs all messages in a database for easy reporting.
For a demo of FaceTime's Messaging Exchange, go to: www.facetime.net/Products/demo/demo.html
There are two scenarios in which Big Science and FaceTime are positioning their
partnership. In the first case, live agents are the default and Big Science's Klones pick
up the slack during off-hours (after live agents have called it quits for the night) and
during peak traffic times. In this way, customers have an immediate, interactive
interface that goes beyond self-help menus to get answers even when live agents
In the second scenario, Big Science's virtual agents are on the front line of customer
service. Customers go to the chatterbot first to find answers to frequently asked
questions and general information. If the Klone is unable to help the customer, it
escalates that interaction to a live agent through Messaging Exchange. This eliminates
much of the strain on call center agents, as they are relieved from answering repetitive
basic queries and information requests. The costs go down, too, as fewer agents are
utilized for service now being handled by the chatterbot.
Both solutions meet common company goals: reducing costs (with fewer call center
agents) and keeping customers happy (by giving them multiple contact options). The ink on
the Big Science/FaceTime deal is barely dry, so it will be interesting to see actual
deployments of this joint solution in the coming months. Expect to see similar
partnerships in the industry announced throughout this year as customer support via the
Web grows and matures.
Dara Bloom welcomes your comments at firstname.lastname@example.org.