ATT DEFINITY Generic 3 Call Vectoring/Expert Agent Instructions Manual

							Basic Call Vectoring
4-4Issue  4 September 1995
Delay Announcements
Here is an example of a delay announcement:
Figure 4-1. Delay Announcement
If the caller does as suggested but ends up waiting an a ppreciable amount of 
time without receiving further feedback, he or she may tire of waiting and hang 
up.  To keep the caller on the p hone at least a little longer, a supplementary d elay 
announcement similar to the one following might be used:
Figure 4-2. Supplementary Delay Announcement
A d elay announcement is usually coupled with a delay ste p, which is p rovided by 
the 
wait-time command (discussed later).
The customer should incorporate as many supplementary d elay announcements 
as he or she deems necessary, given the resourc es available.
Forced Announcements
There are times when the customer may find it advantageous to have the agents 
not answer calls.  Usually, this option is exercised whenever the customer 
anticipates a barrage of calls concerning an emergency or a service problem of 
which the customer is already aware.  Accordingly, the customer can incorporate 
an appropriate announcement as the very first step in the vector.  Such an 
announcement is referred to as a 
forced announcement.  Here’s an examp le.
Figure 4-3. Forced Announcement
      announcement 2556 (‘‘All our agents are busy.
      Please hold.’’)
      announcement 2557 (‘‘Thanks for holding.  All
      our agents are still busy.  Please hold.’’)
      announcement 1050 (‘‘We are aware of the current
      situation and are working to rectify the problem.  If your
      call is not urgent, please call back later.’’) 
						

							Functions and Examples
Issue  4 September 1995
4-5
Information Announcements
Under certain circumstances, the customer may find it necessary to provide the 
caller with recorded information that, by its very content, resolves a problem with 
such finality that the caller feels no need to follow up on his or her call.  Such a 
recorded message is referred to as an 
information announcement.  An example 
follows:
Figure 4-4. Information Announcement
Note that the 
disconnect command is used with the announcement. After the 
announcement, the caller is disconnected, since he or she need not stay on the 
line any longer.
Delays with Audible Feedback
In presenting an example of a delay announcement earlier in this chapter, we 
mentioned that this type of announcement is usually coupled with a delay ste p.  
A delay step is provided by the 
wait-time command, which allows the caller to 
remain on hold for at least the number of seconds indicated in the command.
Let’s take another look at our delay announcement.  However, this time, let’s 
couple the announcement with a delay step:
Figure 4-5. Delay with Audible Feedback
Here, the caller is allowed to wait at least 20 seconds for the call to be answered 
by an agent.  During this wait period, the caller is provid e d with system music, 
which is one ty pe of feed b ack available via the 
wait-time command.
If the d elay step is the final effective step in the vector, the aud i ble feed back 
continues beyond the specified duration. (A “final effective step” in a vector is 
either the last vector step or a vector step that is followed by a 
stop step.) Under 
normal circumstances, the audible feed back continues until the call is either 
answered or abandoned. However, if the call is not q ueued when vector 
      disconnect after announcement 2918 (‘‘Today has
      been declared a snow day.  Please report for work tomorrow
      at 8 A.M.’’)
      announcement 2556 (‘‘All of our agents are busy.
      Please hold.’’)
      wait-time 20 seconds hearing music 
						

							Basic Call Vectoring
4-6Issue  4 September 1995
processing stops, the call is dropped. Feed back also continues while a call is 
queued to a converse split, that is, any split routed to by a converse-on split 
command, an d while data is being passed to a Voice Response Unit (VRU) (see 
the Voice Response Scripts section later in this chapter). Finally, feedback also 
continues during the wait period before the connection of an announcement 
and/or a Touch-Tone Receiver (TTR). (TTRs are used in conjunction with the Call 
Promp ting feature and are discussed in Chapter 5.)
Multiple Audio or Music Sources
With G3V4 a n d later releases, you can specify an alternative audio or music 
source for a ve ctor 
wait-time step. This alternative source can be any extension 
number that is administered on the Announcements/Audio Sources form. For 
instructions for entering an audio or music source on this form see 
DE FI NITY  
Communication System Generic 3 Version 4 Implementation
, 555-230-655 or 
DEFI NI TY Com munication System Generic 3 V2/V3 Imp lementation, 
555-230-653.
With the Multiple Audio/Music Sources feature, you can tailor the 
wait-time 
feed b ack to the interests, tastes, or requirements of the audience. You can 
provide specific types of music or music with overlays of advertising that relate to 
the service provided by the splits or skills that the vector serves. Or, additional 
advertising messages can be heard  by the caller as they wait for an available 
agent.
Let’s look again at our d elay announcement. However, this time let’s assign an 
alternate audio/music source to the 
wait-time ste p.
Figure 4-6. Delay with Multiple Audio/Music Source Feedback
When the wait-time step is encountered, the caller is connected to extension 
55558 for 20 seconds. At the end of 20 seconds the next vector step is executed. 
The “then” option in the wait-time step specifies what the caller will hear if the 
caller cannot b e connected to the sp ecified source. Or, when the call is waiting in 
queue, the “then” option specifies what the call will hear if the call is not 
answered in 20 seconds. In this example, if the call is not answered in 20 
seconds, the caller will hear system music until a subsequent 
announcement, 
busy, collect, converse-on, disconnect 
or wait-time step is encountered.
You c a n specify 
music (system music), ringback, silence, or continue for the 
“then” option. When continue is specified, the caller continues to hear the 
      announcement 2556 (‘‘All of our agents are busy.
      Please hold.’’)
      wait-time 20 seconds hearing 55558 then music 
						

							Functions and Examples
Issue  4 September 1995
4-7
alternative audio or music  source until it is replaced by a subsequent vector step 
regardless of the time specified in the 
wait-time step.
Busy Tone
A busy tone and subsequent termination of vector processing are produced via 
the 
b usy command. An exception to this occurs on CO trunks where answer 
supervision has not been sent. Callers on such trunks do not hear the busy tone 
from the switch. Instead, these callers continue to hear rin g back from the CO. 
The 
busy command eventually times out and drops the call after 45 seconds. 
With ISDN PRI, b usy tone can be provided from the network switch.
The customer usually forces a busy tone to process a call that arrives at a time 
when there is a large numb er of calls queued in the main split, or when the call 
center is out of service or closed.
The following vector illustrates how you can use the 
b usy command:
Figure 4-7. Providing Busy Tone
In this vector, the 
g oto step command in Step 1 sends c all c ontrol to busy in Step 
6 if the conditions in the former command are met.  Specifically, if the number of 
calls queued at a high priority is greater than 30, the 
b usy command is 
accessed.
Disconnect
The customer can opt to have a call disconnected by incorporating the 
disconnect command. However, the customer should use good judgement in 
this regard lest he or she unnecessarily infuriate the caller. Also, as a courtesy to 
the caller, an announcement should be given to the caller before he or she is 
disconnected under any circumstances.
      1. goto step 6 if calls-queued in split 1 pri h > 30
      2. queue-to main split 1 pri h
      3. announcement 4000
      4. wait-time 2 seconds hearing music
      5. stop
      6. busy 
						

							Basic Call Vectoring
4-8Issue  4 September 1995
The disconnect command itself has a b uilt-in announcement option. We saw an 
example of the command when we were discussing information announcements 
earlier in this chapter.  Here’s the example again:
Figure 4-8. Disconnecting a Call
This example presents an ideal use of the 
disconnect command. The caller is 
given record ed information that, b y its very content, resolves a problem with such 
finality that the caller feels no need to follow up on his or her call.
Voice Response Scripts
Voice Response Integration (VRI) is designed to enhance the integration of 
AT&T’s DEFINITY  call center systems an d to integrate DEFINITY  Ca ll Vectoring 
with the capabilities of voice response units (VRUs), particularly AT&T’s 
CONVERSANT Voice Information System.
VRI can do the following:
nExecute a VRU script while retaining control of the call in DEFINITY  vector 
processing.
nExecute a VRU script while the call remains in the split queue and retains 
its p osition in the queue.
nPool CONVERSANT ports for multiple ap plications (which previously was 
possible only when ASAI was present.)
nUse a VRU as a flexible external announcement device.
nPass data between the DEFINITY switch and a VRU.
nTandem VRU d ata through the DEFINITY  switch to an ASAI host.
The c a pabilities in the previous list are provid e d by the 
converse-on sp lit 
command, which is an enhancement to the Basic Call Vectoring customer 
option. The 
converse-on split Call Vectoring ste p is specifically designed to 
integrate a VRU with the DEFINITY  A ut om ati c  Call Distribution (ACD). VRI allows 
VRU capabilities to be used while keeping control of the call in the DEFI NI TY 
ACD. The inclusion of VRUs with vector processing provides the following 
advantages:
nAccess to local and host databases.
nValidation of  caller information.
      disconnect after announcement 2918 (‘‘Today has
      been declared a snow day.  Please report for work tomorrow
      at 8 P.M.’’) 
						

							Functions and Examples
Issue  4 September 1995
4-9
nText to speech capabilities.
nSpeech recognition.
nIncreased recorded announcement c a pacity.
nAudiotex a p plications.
nInteractive Voice Response (IVR) a p plications.
nTransaction processing applications.
One of the advanta ges of VRI is that it allows users to make more productive use 
of queuing time. For example, while the call is waiting in queue, the caller can 
listen to product information by using an audiotex application or by comp leting 
an interactive voice response transaction. In some cases, it may even be 
possible to resolve the customer’s questions while the call is in queue.  This can 
help reduce the queuing time for all other callers during peak intervals.
In a ddition, when Advanced Vector Routing is enabled, the Expected Wait Time 
for a call can be passed to the VRU. In this way, the caller c an be told how much 
longer they can expect to wait before their call will be answered. See Expected 
Wait Time (EWT) on p a ge 6-2 for a comp lete description of the EWT feature.
During the execution of a VRU script, if the caller previously queued to an ACD 
split, the caller retains his/her position in queue. If an agent on the DEFI NI TY 
switch becomes available to service the call, the line to the VRU is imme diately 
dropped, and the calling party is connected to the available agent.
Here’s an example of a vector that can access voice response scripts from a 
VR U .
NOTE:
Recall from Chapter 1 that one or more VDNs can access the same vector.  
This c a pability is appropriate for the following example.
Figure 4-9. Accessing Voice Response Scripts
VDN (extension=1040   name=‘‘car loans’’     vector=40)
VDN (extension=1041   name=‘‘equity loans’’  vector=40)
Vector 40
     1. goto step 10 if calls-queued in split 1 pri h > 30
     2. queue-to main split 1 pri h
     3. announcement 4000
     4. goto step 7 if calls-queued in split 1 pri h < 5
     5. wait-time 0 seconds hearing music
     6. converse-on split 11 pri h passing vdn and none
     7. wait-time 20 seconds hearing music
     8. announcement 4001
     9. goto step 7 if unconditionally
    10. busy 
						

							Basic Call Vectoring
4-10Issue  4 Septemb er 1995
For this example, let’s suppose first that a caller would like to hear information 
concerning car loans. Let’s also assume the call is queued to split 1 (Step 2) and 
that vector processing proceeds to Step 6. In such a case, the 
c onverse-on split 
command in this step delivers the call to the converse sp lit if there is a queue for 
the split and the queue is not full, or if a VRU port is available. (Otherwise, vector 
processing continues at the next vector step.) When the VRU port responds, the 
step then outpulses VDN 1040 to the VRU via the 
p assing vdn p arameter 
included in the command. In turn, the VRU executes the “car loans”  voice 
response script for the caller. Note that it is imp ortant to provide a feedback step 
prior to the converse-on step in case there is a d elay in reaching an available 
converse sp lit port. In this example ste p 5 provides music for this purpose.
Now, let’s suppose another caller wants information concerning equity loans. In 
such a case, if everything proceeds according to form, VDN 1041 is outpulsed to 
the VRU, which in turn executes the ‘‘equity loan’’ voice response script for the 
caller.
In either case, while interaction with the VRU is taking place, the call remains in 
the appropriate split’s queue (split 1 in this example).  If an agent answers the 
call while the voice resp onse script is being executed, the voice response script 
is interrupted, the line to the VRU is dropped, and the caller is connected to the 
available agent. Once a voice response script starts, no further vector steps are 
executed until the voice response sc ript is completed.
NOTE:
Refer to Appendix H for a detailed explanation of the call flow for 
converse—VRI calls.
Besides VDN extensions, the 
converse-on sp lit command can outpulse to the 
VRU calling party extensions, collected (inputted) caller d igits (if Call Promp ting 
is enabled), Exp ected Wait Time (if Advanced Vector Routing is enabled) call 
queue positions, a string of a maximum of six digits or asterisks, a pound sign 
(#), or nothing. Further d etails are included in Chapter 5, Call Promp ting, 
Chapter 6, Advanced Vector Routing  and  in  Ap pendix  A, Call Vectoring 
Command s.
NOTE:
In the previous example, the c alls-queued condition in the second goto 
step (Step  4) in effect serves as a checkpoint for determining whether or 
not there is enough time for the voice response script (activated by the 
converse-on step) to be executed. Specifically, if five or more calls are 
queued to split 1, it is considered feasible to execute the voice response 
script.
Routing Calls
Basic Call Vectoring offers several means of routing telephone calls, as follows:
nQueuing calls to ACD splits 
						

							Functions and Examples
Issue  4 September 1995
4-11
nLeaving recorded messages
nSending c alls to a vector-programmed number (that represents an internal 
or external d estination)
nRouting calls to a Feature Access Code that will activate Service 
Observing
nAdjunct Routin g
The following sections discuss these routing procedures (except for adjunct 
routing, which is fully described in Chapter 9).
Queuing Calls to ACD Splits
Calls that come into the Call Vectoring system can be queued to a maximum of 
three ACD splits.  Two commands are used to queue calls to splits.
The 
queue-to main split command queues a call unconditionally. The command  
sends a call to a split and assigns a queuing priority level to the call in case all 
agents are busy.
The 
check-backup split command conditionally checks the status of a split for 
possible termination of the call to that split. The command either connects the 
call to an agent in the split or puts the call into the split’s queue (at the specified 
priority level) if the condition specified as part of the command is met.
Multiple Split Queuing
The term multiple split queuing refers to the q ueuing of a call to more than one 
split at the same time. The following vector helps to illustrate this process.
Figure 4-10. Multiple Split Queuing
To avoid completing vector processing without queuing the call to a split, it is 
always good practice to check a split’s queue before queuing to that sp lit.  If the 
queue is full, alternate treatment (such as queuing to an alternate split) should b e 
provided. In this vector, if the main split’s queue (which has 10 queue slots) is 
full, the 
goto ste p command in Step 1 skips the main split and goes directly to 
      1. goto step 4 if calls-queued in split 1 pri l >= 10
      2. queue-to main split 1 pri t
3. wait-time 12 seconds hearing ringback
4. check-backup split 2 pri m if calls-queued < 5
5. check-backup split 3 pri m if calls-queued < 5
6. announcement 3001
7. wait-time 50 secs hearing music
8. goto step 4 if unconditionally 
						

							Basic Call Vectoring
4-12Issue  4 Septemb er 1995
step 4 to check the backup sp lits. Although calls are queued in Step 2 at a top 
priority, a low priority is specified in step 1 so that calls in queue at all priority 
levels are counted. If there are 10 or fewer calls in the main split, control is 
passed to Step 2, where the 
q ueue-to main split command queues the call to 
split 1. Once the call is queued, vector processing continues at the next step. 
Step 4 c ontains a 
c heck-backup split  command.  (Recall that in the last 
paragraph we mention that this step is branched to if the main split queue is 
holding 10 or more calls.)  If the call is not answered by the time Step 4 is 
reached, the 
c heck-backup split in the step attempts to queue the call to a 
second split.  Specifically, the command first d etermines whether there are fewer 
than 5 calls queued  to sp lit 2.  If so, the command then attempts to connect the 
call to an agent in the split. If such a connection cannot be ma de, the command 
puts the call into the split’s queue (at the specified priority level). Vector 
processing then continues at the next ste p. On the other hand, if there are 5 or 
more calls queued to split 2, the command fails, and vector processing 
continues at Step 5.
Step 5 c ontains another 
c heck-backup sp lit  command and, accordingly, the 
process described in the previous paragraph is repeated, with one difference: 
the queuing attempt is made to split 3 instead of to sp lit 2. 
Except for the condition check, the circumstances under which the 
check-
backup split
 command cannot queue a call are i dentical to those for the q ueue-
to main split
 command.
Finally, note that whenever a call is queued to a backup split, the call remains 
queued to the main split and/or to another backup split (if already queued to 
either or b oth of these splits). Once the call is answered in a split to which it is 
queued, the call is automatically removed from all the other split(s) to which it is 
also queued.
NOTE:
The check-backup split, q ueue-to main sp lit, and converse-on c ommands 
can access 
only those splits that are “vector-controlled.” A split is 
considered “vector-controlled” if 
yes is entered in the Vector field of the 
Hunt Group Administration Screen. 
						

							Functions and Examples
Issue  4 September 1995
4-13
Leaving Recorded Messages
Basic Call Vectoring allows the caller to leave a message for the customer if the 
agents at the customer site are not available to take telephone c alls.  This is done 
with the help of the 
messaging split command. Let’s take a look at an example.
Figure 4-11. Leaving Recorded Message
In this vector, the 
goto step command in Step 1 checks to see if the office is 
open, and branches to Step 8 if the office is closed.  This is done to 
accommodate calls that are made during non-working hours, when there are no 
agents available to take telephone calls. Accordingly, Step 8 provides the caller 
with an appropriate announcement and an opportunity to leave a record ed 
message.
Step 2 checks to see if split 47’s queue (which has 20 queue slots) is full, and 
branches to ste p 10 if it is. Ste ps 3 to 7 queue the call to sp lit 47 and then give 
audible feedback to the caller.
If the caller chooses to leave a message, the 
messaging split command in Step 
11 is executed. Sp lit 18 in the command is the Audio Information Exchange 
(AUDIX) split.  AUDIX is a voice mail adjunct that allows a customer to record, 
edit, store, forward, and retrieve voice messages to and/or from callers.  
Extension 2000 is the mailbox for sp lit 47 (from Step 2), which represents a group 
of daytime agents.
Upon execution of the 
messaging s plit command, an attempt is made to connect 
the caller to AUDIX so he or she can leave a recorded message. If the split 
queue is full, or if the AUDIX link is d own, termination to AUDIX is unsuccessful, 
and vector processing continues at the next vector step, which (as is the case 
      1. goto step 8 if time-of-day is all 16:30 to all 7:30
      2. goto step 10 if calls-queued in split 47 pri l >= 20
      3. queue-to main split 47 pri m
4. wait-time 12 secs hearing ringback
5. announcement 4001
      6. wait-time 60 secs hearing music
7. goto step 5 if unconditionally
8. announcement 4111(‘‘We’re sorry, our office
         is closed.  If you’d like to leave a message, please
         do so after the tone.  Otherwise, please call back
         weekdays between 7:30 A.M. and 4:30 P.M.  Thank you.’’)
9. goto step 11 if unconditionally
10. announcement 4222 (“We’re sorry, all of our agents are busy, 
please leave a message after the tone and we will return your 
call.”)
11. messaging split 18 for extension 2000
     12. disconnect after announcement 4333 (‘‘We’re sorry, we are
         unable to take your message at this time.  Please
         call back at your convenience weekdays between
         7:30 A.M. and 4:30 P.M.  Thank you.’’)