|
For a long time chat remained a bit of Internet whimsy that, while
possibly having tremendous potential as a business tool, still appeared to
function best when applied towards entertainment or "community
building" purposes. Perhaps the current reinvention of the customer
(a component that used to be optional in the typical Internet business
plan) can be credited with chat's recent elevation from lark to compelling
form of customer interaction.
A side effect of this focus on the customer is that we've provided
online consumers with a million ways to contact an agent, without giving
agents a way to manage multiple contacts conducted in multimedia. In
addition to telephone calls, they're now asked to deal with chat and
e-mail. Not only do the beleaguered agents suffer, contributing to a
traditionally high turnover rate within the industry, but customers
receive poor service through impersonal, systemized responses or curt
replies from overtaxed agents.
Enter eShare, who was in the chat and e-mail marketplace long before
their purchase by Melita. The 4.0 release of their NetAgent product brings
a new level of productivity to the multimedia call center, allowing agents
to handle multiple chat visitors and e-mail recipients without losing that
crucial human touch. In addition to the inclusion of bi-directional Web
collaboration, Web callback requests, and assisted form-filling, the new
version brings NetAgent closer to the CRM space by supporting the
leveraging of customer histories through Enterprise Data Access. In
addition to these chat features, a separately licensed component called
"Re:sponse" provides e-mail management through features such as
skills-based routing.
INSTALLATION
NetAgent was installed on a machine running Windows 2000. Its minimum
requirements are Windows NT 4 or higher (this excludes some NT versions
not offering a specific options pack), 96 MB of RAM, and a Pentium II
processor. Additionally it requires a dedicated copy of MS Internet
Information Server and its own ODBC-compliant database: MS SQL Server was
used for the installation.
Requisites aside, server and client installation of NetAgent were
exceptionally quick and simple. Our installation, including many pauses to
explore various features and pose questions to the visiting eShare
representative, was completed in well under an hour. It does not appear
that installation would present much of an interruption to normal
operations in a newly-adopting contact center. Additionally, APIs allow
administrators to integrate NetAgent with existing applications such that,
during training, agents can familiarize themselves with NetAgent while
still utilizing their existing system.
DOCUMENTATION
EShare sent TMC Labs five manuals: Installation, Supervisor's, Agent's,
Reports, and Database Reference. Each was built with a clear, concise
table of contents and contained a logical structure throughout. The
Installation manual provided checkboxes at the end of each procedure,
which is a small but beneficial tool if installation is interrupted and
must be continued at a later time. The Installation manual also provided
detailed screen shots for nearly every step in each procedure.
Both the Supervisor and Agent's manual contained a Glossary of NetAgent
Terms, which we also liked. The NetAgent online Help files are adequate --
providing the standard Contents, Find, and Index tabs -- though it would
have been nice if you could return to the original help dialog box after
selecting a topic. The only other hampering issue we encountered with the
documentation was that we weren't always able to facilitate reference due
to the absence of an index in the manuals.
FEATURES
NetAgent is aimed at maximizing an agent's time and reducing operational
costs, while enhancing a customer's experience. EShare calls the following
features Agent Productivity Maximizers (APMs). Some of these include:
auto-pilot, which is a self-running interactive question-and-answer
session that directs the consumer, according to their response to
predetermined prompts, towards various contents from the host site.
Frequently Used Items (FUI), consist of text, files, and Web pages
compiled by agents and administrators are used to generate a large amount
of material, accommodating customer inquiries in specific areas. A
slide-show presentation consisting of Web pages and automated
instructional text can also be sent to a customer. A real-time spell
checker is also considered an APM.
Other features include:
- Skills-based e-mail routing.
- Integration with PBX/ACD for call queuing and rollover if an agent
is in "work mode."
- Multi-level administrator logins.
- Bi-directional page pushing.
- Enterprise data access (EDA).
- Frequently used items (sayings, Web pages, etc.).
- Web callback requests.
- Consumer-scheduled PSTN callback requests.
- Customer interaction logging and history.
- VoIP via NetMeeting, Lipstream, or HearMe.
- Script with selectable push page as "hold"
("Info-On-Hold").
- Support for outbound e-mail campaigns.
OPERATIONAL TESTING
NetAgent
Chatting through NetAgent was tested both from an agent's and a user's
perspective: the goal was to see how well an agent could interact with
multiple site visitors in a quick and believable manner. Using the
Web-based administration screens -- which, while being a bit Spartan in
appearance, proved exceptional in ease of use -- a few FUIs (greetings or
responses that an agent can automatically trigger when needed) were
created, and scripted content as well as "on hold" Web pages (to
be shown while consumers wait for a net agent) were chosen. It took a very
short time to accomplish all of this, and would obviously be even shorter
for a seasoned NetAgent administrator. Once these configurations were
completed, multiple chat sessions which showed themselves as several
screens opening up on the agent interface were initiated by simulated
consumers.
Because of the various levels of interaction afforded through NetAgent,
it proved simple to engage in live, one-on-one chat with a needier visitor
while pushing canned content to others, allowing agents to deal with a
higher volume of visitors while maintaining a human touch. At the most
intimate level is regular chat, which proved as close to
"real-time" as possible for both agent and consumer in the Labs
environment. The most automated forms of interaction are presentations,
where scripted content is accompanied by pushed Web pages, and Auto Pilot
sequences which push pages according to consumers' responses to
"yes" and "no" prompts. Even while this automated
content was active, our faux agents were able to monitor its progress in
case the visitor should require more help.
Scripts And FUIs
Somewhere in between complete automation and completely personal
interaction, a middle ground is afforded by FUIs and scripted content.
Scripted content (as well as telephony integration) allows chat to be
treated and managed in the same manner as traditional telephone calls to a
contact center, affording a recognizable experience to both agents and
consumers. For example, when acting as consumers, we received an "on
hold" Web page and message pushed to our browser upon initiating chat
-- providing an even more interactive equivalent to on hold telephone
recordings. While this page is presented, an agent is given a
pre-determined pause period to prepare for the oncoming chat. When
telephony integration is in place NetAgent integrates with the ACD such
that calls are bounced past chatting agents to available ones. We used
FUIs, in our tests to field common questions or greet new visitors in a
naturalistic manner while allowing the faux agents to respond to or
monitor other chats. For example, our agents were able to select from
pre-established greetings (such as, "Hi, how can I help you
today?") upon initiation of chats, while responding to questions
about shipping or pricing with pre-scripted answers.
In an ideal scenario, an agent could resultantly provide a mixture of
live and virtually live responses to customer queries, dealing with other
tasks while visitors are occupied. As for how real it is, it would take
real savvy or prior knowledge to realize when an agent is actually
responding as opposed to pushing scripted material.
Web Collaboration
With this release eShare introduced bi-directional Web collaboration
(either the customer or agent can determine which pages are co-browsed)
and assisted form-filling. The response time for both of these features
was exceptionally fast. In a test of form-filling, an "agent"
conducted Yahoo! searches on a "customer's" computer, entering
key words into the search field with minimal delay.
VoIP And Call-back Requests
The default customer interface allows consumers to enter their name,
e-mail address, a description of their issue or question, and to choose
which Service Line they need to be place into (i.e. support, sales, etc.).
It also provides checkboxes which, if selected, will trigger a request
either for an immediate NetMeeting callback or will allow the customer to
enter a specific time at which they'd like to receive a regular PSTN call.
These features were tested in the midst of our simulated customer support
session. PSTN callback requests were queued in the appropriate service
line, NetMeeting requests were announced by an "accept" or
"decline" pop-up window on the agent interface. While individual
Web calls could be placed and received in an efficient manner, it was not
apparent exactly how the system would handle multiple VoIP callback
requests. Although the system integrates with a company's PBX and ACD
systems to distribute PSTN calls to available agents, it appears as though
chatting agents may still be disturbed by NetMeeting callback requests.
Re:sponse
To test the Re:sponse feature, skills-based rules and
categorically-triggered stock responses were created to evaluate how the
system handles multiple e-mails with varying content. Re:sponse performed
exactly as it was configured, directing e-mail with certain keywords in
the subject or body to only one of our two fictional agents (in this case
the keyword was "shocks," as in automobile suspension), and
automatically populating e-mail replies to those e-mails with pre-written
content (i.e. "AutoParts Warehouse has a full line of shocks.").
Our agents were supplied with an e-mail workspace within which they could
choose from various stock phrases while also composing their own content.
ROOM FOR IMPROVEMENT
As readers can probably tell, we were uniformly impressed with NetAgent.
It offers what appears to be an exceptional way for agents to interact
quickly and naturally with customers, applying the proven call center
model to chat as well as e-mail. Equally as important though, is the
ability to build long-term relationships with these same customers,
turning these immediate interactions into customer histories while mining
the results of past interactions. eShare's inclusion of Enterprise Data
Access support with this release and its support for database lookups are
steps in the right direction. What would take the product to an entirely
different level would be the automation of at least some of these
features.
In an ideal scenario, once a new visitor submits a chat request, the
user name or e-mail address is checked against the customer database and
any pertinent information from their history (Are they a huge client? An
investor? What types of products have they purchased in the past? Do they
have any outstanding support issues that have not been addressed?) would
be automatically populated in the NetAgent screen pop. eShare should be
credited for a Chat EDA which presents some visitor login info within the
incoming screen pop-up, but it still does appear that required manual
querying could lesson response time.
Additionally, in customer service or support scenarios, agents require
the ability to log "calls" (in this case chats or e-mail
exchanges) quickly, report problems, and create trouble tickets or assign
action requests all within the same, easily accessible interface -- one
that shows and prioritizes their existing tasks.
Currently eShare's open architecture, which includes software
development kits and APIs, would most likely allow a company to integrate
NetAgent with another CRM application satisfying the requirements above.
All that is being suggested here is that an already powerful product which
handles the first phase of customer contact with great aptitude would be
made even more powerful if eShare themselves were to bundle it with such a
CRM application. Any number of powerful alliances would allow them to do
so.
The only other room for improvement would be uniform treatment of VoIP
and PSTN calls (if it is not already supported), such that multiple
NetMeeting callback requests are queued along those for regular telephone
calls.
CONCLUSION
NetAgent 4.0 is a feature-rich application with far too many components to
list. eShare has done a good job of delivering both the agents and
visitors an array of two-way mediums for collaboration and information
delivery. Agents can more effectively manage their time by choosing how
heavily they need to interact with a customer, when to return calls,
answer e-mail, and not get interrupted by a text chat or phone call when
engaged with another customer. Pushing scripted responses and Web pages to
answer FAQs frees time and manpower to respond to other queries. This
appears to increase agent productivity with minimal effort, and seems to
provide more accurate and comprehensive information in a shorter period of
time.
Ultimately this leads to a golden mean between complete automation on
the one hand and taxing, one-on-one chat on the other. If (or when) eShare
includes even more integration with the back-end part of customer
interaction and expands their VoIP capabilities, their already impressive
product will be no contest for a full A+ rating.
[ Return
To The March 2001 Table Of Contents ] |