AT&T DEFINITY Generic 3 Manuale Utente
Advanced Vector Routing
6-4
Issue 4 September 1995
The VRU can also announce expected wait time to a caller periodically
throughout the time that a call is in queue. In this way, the caller can observe his
or her progress up the queue. However, this approach should be used with
caution. Circumstances such as a reduction in the number of agents or a sudden
influx of higher priority calls could cause the caller’s expected wait time to
increase from one announcement to the next.
throughout the time that a call is in queue. In this way, the caller can observe his
or her progress up the queue. However, this approach should be used with
caution. Circumstances such as a reduction in the number of agents or a sudden
influx of higher priority calls could cause the caller’s expected wait time to
increase from one announcement to the next.
If the call is not queued or if it is queued only to splits that are unstaffed or splits
where all agents are in AUX work mode, the end-of-string character “#” is the
only data item outpulsed.
where all agents are in AUX work mode, the end-of-string character “#” is the
only data item outpulsed.
The EWT Algorithm
EWT is calculated using an algorithm that is based on the number of calls in a
queue at a particular priority level and the rate of service of calls from the queue
at that priority level. It adjusts for many other factors such as multiple split
queuing, call handling times, and the impact of direct agent calls on the wait time
of other calls to the split. The algorithm adjusts EWT immediately for changes in
staffing, such as agents logging in or taking breaks in AUX work mode.
queue at a particular priority level and the rate of service of calls from the queue
at that priority level. It adjusts for many other factors such as multiple split
queuing, call handling times, and the impact of direct agent calls on the wait time
of other calls to the split. The algorithm adjusts EWT immediately for changes in
staffing, such as agents logging in or taking breaks in AUX work mode.
Since changes occur constantly in a call center, and since EWT cannot predict
the future, the accuracy of the EWT predictions will be in proportion to the rate at
which call s are serviced from the queue and the level of stability achieved in the
call center between the time that the prediction is made and the time that the call
is serviced from queue.
the future, the accuracy of the EWT predictions will be in proportion to the rate at
which call s are serviced from the queue and the level of stability achieved in the
call center between the time that the prediction is made and the time that the call
is serviced from queue.
When to Use Wait Time Predictions
Wait time predictions are best suited for medium or high volume call scenarios.
In general, the potential accuracy of a wait time predictor increases as the rate of
removal from queue increases. It is recommended that EWT be used when the
rate of removal from queue at a given split priority level is at least one call every
30 seconds.
In general, the potential accuracy of a wait time predictor increases as the rate of
removal from queue increases. It is recommended that EWT be used when the
rate of removal from queue at a given split priority level is at least one call every
30 seconds.
Predictions can be made for a split with multiple priority levels in use as long as
the majority of calls are delivered to the lower priority levels. If the majority of calls
are queued at the higher priority levels, any predictions made for the lower
priority levels may not be accurate.
the majority of calls are delivered to the lower priority levels. If the majority of calls
are queued at the higher priority levels, any predictions made for the lower
priority levels may not be accurate.
The following list describes circumstances that will limit the accuracy of the wait
time predictions.
time predictions.
■
Immediately after a system restart or when a new split is administered.
The EWT algorithm uses a combination of historical and real-time
information to make predictions. When no historical information exists,
such as when a new split is added or a reset system 3 or 4 is completed,
there is the potential for inaccuracies.
information to make predictions. When no historical information exists,
such as when a new split is added or a reset system 3 or 4 is completed,
there is the potential for inaccuracies.