Oesterreichisches Forschungsinstitut fuer Artificial Intelligence (OeFAI)
Schottengasse 3, A-1010 Wien
Tel.+43-1-53532810,5336112
-------------------------------------------------------------------------
VORTRAG
*******
Prof.Dr.Vasant DHAR
New York University
New York, N.Y.
The Intelligent Customer Support Function:
Minimizing Customer Downtime and Dissatisfaction
------------------------------------------------
A common challenge facing the customer support function in business
organizations is the effective allocation of personnel to tasks so
that the loss of value by the firm due to customers' "down time" is
minimized. But this alone won't insure that customers are happy. We
need to be concerned with minimizing dissatisfaction due to the
timeliness of support. In practice, this is complicated by numerous
issues. The customer support administrator needs to consider the
priority of the various tasks in the queue. Is a problem a severe one
that seriously prevents a user from doing her job, or is it a less
serious problem that is more of an inconvenience? The administrator
must also consider how long will it take to resolve a particular task
and how this impacts the support function's ability to service other
users. The ability of the various customer service representatives to
perform the various tasks will also impact schedule design. All
things being equal, it makes little sense to have a highly experienced
technician perform a relatively simple task while a more complex task
remains undone because the other (idle) technicians do not have the
skills to perform it. On top of all of this, the administrator also
needs to consider the amount of time that a given task has been
outstanding. The longer a problem, even a minor one, is outstanding,
the more dissatisfied a user will be resulting a loss of goodwill.
I shall present an "intelligent" solution to the problem that combines
a genetic algorithm as the task assignment engine with a standard
problem tracking system. The system is currently in routine use at
Moody's Investor's Service where it supports the help desk function.
It has been a resounding success, showing how Artificial Intelligence
techniques can be integrated with conventional technologies to yield
powerful and practical systems.
Zeit: Dienstag, 23.Mai 1995, 18:30 Uhr pktl.
Ort: OeFAI, Schottengasse 3, 1010 Wien 1.