U.S. patent application number 10/783739 was filed with the patent office on 2004-08-19 for adaptive voice recognition menu method and system.
This patent application is currently assigned to SBC Technology Resources, Inc.. Invention is credited to Bushey, Robert R., Joseph, Kurt M., Knott, Benjamin A., Martin, John M..
Application Number | 20040161078 10/783739 |
Document ID | / |
Family ID | 29249390 |
Filed Date | 2004-08-19 |
United States Patent
Application |
20040161078 |
Kind Code |
A1 |
Knott, Benjamin A. ; et
al. |
August 19, 2004 |
Adaptive voice recognition menu method and system
Abstract
A method and system for automated speech-enabled responses to
caller requests for information prompts caller requests for
information from a menu that lists options based upon the frequency
of requests for information and for automated collection and
updating of content based upon the feedback provided by customers.
For instance, a task analyzer classifies information requested by
caller opening statements with the classifications based on the
task the caller intends to accomplish. For a menu that provides
information on telephone services, classifications may include
instructions on how to use a service, pre-purchase information,
identification of the services that the caller subscribes to, the
cost of services, post-purchase information and identification of
the caller's long distance carrier.
Inventors: |
Knott, Benjamin A.; (Round
Rock, TX) ; Bushey, Robert R.; (Cedar Park, TX)
; Martin, John M.; (Austin, TX) ; Joseph, Kurt
M.; (Austin, TX) |
Correspondence
Address: |
BAKER BOTTS L.L.P.
PATENT DEPARTMENT
98 SAN JACINTO BLVD., SUITE 1500
AUSTIN
TX
78701-4039
US
|
Assignee: |
SBC Technology Resources,
Inc.
|
Family ID: |
29249390 |
Appl. No.: |
10/783739 |
Filed: |
February 20, 2004 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
10783739 |
Feb 20, 2004 |
|
|
|
10135143 |
Apr 30, 2002 |
|
|
|
6697460 |
|
|
|
|
Current U.S.
Class: |
379/88.01 ;
379/88.22; 379/88.24 |
Current CPC
Class: |
H04M 3/5166 20130101;
Y10S 379/917 20130101; H04M 3/493 20130101 |
Class at
Publication: |
379/088.01 ;
379/088.22; 379/088.24 |
International
Class: |
H04M 001/64 |
Claims
What is claimed is:
1. A method for effecting a speech-enabled menu, comprising:
defining a menu of options to respond to a caller request for
information, the options in the menu each having a corresponding
information classification; presenting the menu of options to the
caller; prompting for selection of a menu option via a caller voice
utterance; analyzing the caller voice utterance to identify a
selected menu option; associating one or more terms from the caller
voice utterance with the information classification corresponding
to the selected menu option; and updating a voice utterance
glossary with one or more terms associated with the information
classification corresponding to the selected menu option.
2. The method of claim 1, further comprising comparing one or more
aspects of the caller voice utterance with the voice utterance
glossary to identify the selected menu option.
3. The method of claim 1, further comprising defining the menu of
options based on a likelihood of caller selection from the
presented menu of options.
4. The method of claim 3, further comprising defining the menu of
options likely to be selected by the caller based on a frequency of
information classification selection.
5. The method of claim 1, further comprising: determining whether
the selected menu option has associated with it a sub-menu of
options; and presenting the sub-menu of options to the caller for
selection therefrom, the options in the sub-menu each having a
corresponding information classification.
6. The method of claim 5, further comprising comparing one or more
aspects of a caller voice utterance in response to presentation of
the sub-menu of options with the voice utterance glossary to
identify a caller selected sub-menu option.
7. The method of claim 5, further comprising defining the sub-menu
of options presented based on a likelihood of caller selection from
the sub-menu of options.
8. The method of claim 7, further comprising defining the sub-menu
of options likely to be selected by the caller based on a sub-menu
option information classification request frequency.
9. The method of claim 1, further comprising: prompting the caller
for feedback regarding caller experience with the menu of options;
and updating one or more aspects of menu content based on caller
feedback.
10. An information delivery system, comprising: a selection
prediction module operable to predict one or more menu options
likely to be selected by a user; a menu generation module operable
to generate a menu of options including the one or more options
likely to be selected by a user; an interactive voice response
system operable to communicate the menu of options to a user for
selection, await user selection of an option via DTMF input or user
utterance and connect the user to an information module associated
with the user option selection; and an analyzer module operable to
ascertain an option selection in a user utterance, record user
option selections in one or more option selection histories and
consider the one or more option selection histories in predicting
the one or more options likely to be selected by a user.
11. The system of claim 10, further comprising the selection
prediction module operable to predict the one or more options most
likely to be selected based on a selection frequency value in the
option selection histories.
12. The system of claim 10, further comprising: the menu generation
module operable to determine whether user selection from the menu
of options suggests presentation of a sub-menu of options; and if
presentation of a sub-menu of options is suggested, the selection
prediction module operable to predict one or more sub-menu options
likely to be selected by the user.
13. The system of claim 12, further comprising the analyzer module
operable to record user selection from the sub-menu of options in
one or more sub-menu selection histories and consider the one or
more sub-menu option histories in predicting the one or more
sub-menu options likely to be selected by the user.
14. The system of claim 10, further comprising the analyzer module
operable to track user utterance option selections and associate
the user utterances with the selected options.
15. The system of claim 14, further comprising the analyzer module
operable to adapt a user utterance selection glossary to recognize
expected utterances in accordance with the tracked and associated
user utterances.
16. Software for maintaining an option selection menu, the software
embodied in computer readable media and when executed operable to:
ascertain a selection frequency for a plurality of menu options;
generate a current menu of options for presentation to a user based
on menu option selection frequency; present the menu of options for
user selection therefrom; identify an option selection of the user;
record the selection of the user in a selection frequency record
associated with the user selection; and update a user utterance
option selection glossary with one or more terms from a user
utterance menu option selection.
17. The software of claim 16, further operable to route the user to
a system service operable to provide the user with support
associated with the user selected menu option.
18. The software of claim 16, further operable to update the
current menu of options when the current menu of options fails to
identify menu options most likely to be selected by a user.
19. The software of claim 16, further operable to: associate one or
more terms of a user utterance menu option selection with a menu
option task classification; and update the user utterance option
selection glossary with selected terms associated the menu option
task classification.
20. The software of claim 16, further operable to: ascertain a user
selection frequency for a plurality of sub-menu options associated
with the menu option selection of the user; and generate a current
sub-menu of options for presentation to a user based on sub-menu
option selection frequency.
Description
CROSS REFERENCE TO RELATED APPLICATION
[0001] This application is a continuation of U.S. patent
application Ser. No. 10/135,143 entitled "Adaptive Voice
Recognition Menu Method and System" filed on Apr. 30, 2002 by
Benjamin A. Knott et al., now U.S. Pat. No. ______.
TECHNICAL FIELD OF THE INVENTION
[0002] The present invention relates in general to the field of
menus and content for automated responses to callers, and more
particularly relates to a method and system for adapting voice
recognition and touch-tone menus based on the frequency of caller
requests for information and for adapting a voice recognition and
touch-tone content for information retrieval systems.
BACKGROUND OF THE INVENTION
[0003] Telephone call centers handle millions of calls from
consumers each year. One significant expense for call centers is
the expense of training and maintaining operators to answer inbound
calls and interact with callers to help callers complete their
objectives. To manage and reduce costs, call centers typically rely
on automated responses to caller inquiries so that fewer operators
are needed for a given volume of inbound calls. For instance,
inbound calls to a call center are answered by interactive voice
response units that provide information to callers without operator
intervention by interacting with callers through a DTMF tone-based
menu. If callers are able to complete their objectives by
interacting with automated systems, calling centers typically enjoy
substantial cost savings.
[0004] One difficulty with the use of automated responses to caller
inquiries is that the nature of caller inquiries to a calling
center may vary substantially. Thus, automated menus tend to
present a broad range of options for callers to select from.
However, the greater the variety of possible caller objectives, the
more complex the automated menu tends to grow. Complex menu options
tend to discourage callers from pursuing their initial objectives
due to the length of time typically needed to navigate through the
menus. Often, callers end up with their objective unmet, giving up
on the automated system in favor of waiting for an operator. Thus,
complex navigation menus may fail in reducing operator workload
while producing dissatisfied customers.
[0005] One solution that simplifies caller interaction with
automated interactive voice response units is the ability to apply
voice recognition for determining caller inputs. Voice recognition
allows callers to state a query which is analyzed to determine the
desired information. Voice recognition enabled interactive response
units are typically simpler to use than DTMF tone based systems
since a caller is able to simply state a request without having to
select from a defined list of menu items. However, voice
recognition accuracy suffers when the glossary of terms to be
recognized becomes large or when communication of voice utterances
is unclear, such as with a bad telephone connection or an unclear
speaker. Failure of voice recognition often leads to frustrated
callers who are unable to obtain desired objectives or to even
navigate through menus that require voice recognition.
[0006] Another difficulty with the use of automated responses to
caller inquiries is that the content of the information provided to
the caller is rarely updated and tailored to the callers needs over
time. The content of the information provided has a high degree of
validity and applicability on the day of implementation. However,
over time, the callers needs, the technology, and other related
systems change, thus establishing the requirement to update this
content. Out-of-date content will discourage callers from using the
automated systems. Thus, the automated systems will fail to reduce
operator workload while producing dissatisfied customers.
BRIEF DESCRIPTION OF THE DRAWINGS
[0007] The present invention may be better understood, and its
numerous objects, features and advantages made apparent to those
skilled in the art by referencing the accompanying drawings. The
use of the same reference number throughout the several figures
designates a like or similar element.
[0008] FIG. 1 depicts a block diagram of a system for presenting
voice prompt menu options based on the frequency of caller requests
for information;
[0009] FIG. 2 depicts a flow diagram of an opening menu for
providing frequently requested information on telephone
services;
[0010] FIG. 3 depicts a flow diagram of a subsequent menu for
providing frequently requested information on how to use telephone
services;
[0011] FIG. 4 depicts a flow diagram of a subsequent menu for
providing frequently requested information on the cost of telephone
services;
[0012] FIG. 5 depicts a flow diagram of a subsequent menu for
providing frequently requested information on services subscribed
by a caller;
[0013] FIG. 6 depicts a flow diagram of a subsequent menu for
providing frequently requested information on a caller's long
distance carrier; and
[0014] FIG. 7 depicts a flow diagram for updating menu content with
caller feedback.
DETAILED DESCRIPTION OF THE INVENTION
[0015] A speech-recognition enabled, interactive voice response
system presents an adaptable menu to callers to obtain information
over a telephone with speech or touch-tone DTMF inputs. The
customer-centric menu arranges nodes based on call frequency for
classifications of information. The frequency of requests for
information based on the classification of the information adapts
menu options to be presented in an order that relates to the
likelihood of a request for information by a caller, thus providing
a more rapid location of desired information. Analysis of opening
statements made by callers identifies tasks of callers upon initial
contact and relates the tasks to information classifications. The
adaptable menu nodes allow callers to navigate quickly to desired
information by applying voice recognition to caller inputs
responsive to an initial prompt for the caller's task. Menu nodes
list menu options based on the frequency of requests for
classifications of information and include DTMF touch-tone input
options for each information classification. In one embodiment, a
menu adapted to provide frequently requested information on
telephone services provides a reduction in the navigation time of
callers and a reduction in caller reliance on operators.
[0016] An automated system presents adaptable content to callers to
obtain current and appropriate information. Customer-centric
content is updated based on frequency of customer feedback. The
frequency of customer feedback provides a basis for updating the
content, thus providing more current and usable information. The
adaptable content allows callers to retrieve current and usable
information by automatically updating the content based on customer
feedback. This updated content is the sequence of information and
the content itself.
[0017] Referring to FIG. 1, a block diagram depicts a system for
analyzing call request topics and setting a call menu for improved
automated response in accomplishing caller tasks. An interactive
voice response unit (IVR) 10 provides speech and touch-tone enabled
interactive voice responses that allow callers interfaced through a
communication network 12 and telephones 14 to obtain information.
The embodiment depicted by FIG. 1 uses telephone communication
through the Public Switched Telephone Network (PSTN), although
alternative embodiments include mobile or computer telephony
devices communicated through the Internet or other suitable
networks. If a caller is unable to receive information from the
automated responses of IVR 10, then the caller is forwarded to an
operator 16 interfaced with IVR 10 for individual handling.
[0018] Callers establish communication with IVR 10 by, for
instance, placing a telephone call with telephones 14 through PSTN
12 to a predetermined telephone number associated with IVR 10. As
one example, callers request information about their telephone
service by dialing a number provided on their bill or advertised on
their telephone book. IVR 10 provides a voice menu that directs
callers to input information requests through either a voice
response or a touch-tone response. A menu 18 interfaces with IVR 10
to provide callers with menu nodes that guide callers to accomplish
specific desired tasks. Menu 18 has a data-driven design with dual
speech and touch-tone input options. Menu options are presented to
aid callers in completing specific tasks with the order and
selection of menu option presentation based on an analysis of the
frequency of caller requests for specific tasks. A task analyzer 20
interfaces with IVR 10 to analyze the frequency of requested caller
tasks from caller requests to IVR 10. A menu graphical user
interface (GUI) 22 provides access to the task analysis and
interfaces with menu 18 to arrange menu nodes according to the
analysis of the frequency of task requests, thus allowing users to
more efficiently accomplish desired tasks.
[0019] In one embodiment, menu 18 provides an automated service
that provides callers with information about current telephone
services, product or service availability and pricing, and
operating instructions for how to use services. Menu 18 presents
voice responses to callers to request input of and to address the
specific caller task, such as the specific service and associated
information requested by the caller, based upon the analysis of the
frequency of requests for the specific task. By ordering menu
options to address tasks by their requested frequency, menu 18
expedites the identification of caller tasks and helps ensure
accurate and timely transfer of requested information, thus
reducing the likelihood that a caller will request or need operator
intervention. Task analyzer 20 identifies requested information by
frequency and provides the task frequency to menu GUI 22 for
organizing menu 18. In one alternative embodiment, task analyzer 20
automatically updates menu content and order based on frequency
analysis, either on a periodic basis or when significant frequency
changes are noted. Task analyzer 20 automatically implements menu
updates and identifies the updates to system administration through
menu GUI 22.
[0020] In one specific example, task analyzer 20 analyzed the
"opening statements" of callers to determine the caller's
intention, i.e., the task that the caller sought to complete in the
call, and arranged menu 18 to respond to caller requests according
to task frequency. Analysis of 3000 opening statements identified
15.6% of calls as requests for information about telephone products
and services that are efficiently handled by automated responses.
For instance, the opening statement may be a caller utterance to an
operator or a speech-enabled IVR main menu request in which the
user states the purpose of the call. These requests for information
are classified into six high-level user task categories:
1 1 "What is" or "How to use" a feature or 33% 2 Pre-purchase
information. 18% 3 "What service do I have?" or "Do I have a 14%
service?" 4 "How much does a service cost?" 13% 5 Post-purchase
information. 10% 6 "Who is my Long Distance provider?" 6% 7 Other
6%
[0021] The most frequent request for information is for
instructions describing how to use a telephone service, such as
Caller ID, call blocking or call forwarding. Next in frequency are
requests for information about services available for purchase,
such as Internet DSL service, new phone service or wireless phone
service. Third in frequency are requests for information about the
services that the caller already has. Fourth in frequency are
requests for the cost of services, such as adding a second
telephone line or Caller ID. Fifth in frequency are requests for
post-purchase information, such as changing or updating a service.
Sixth in frequency are requests for identification of the caller's
long distance carrier.
[0022] Based upon this frequency analysis, menu 18 provides six
primary functions associated with the identified caller tasks so
that caller requests for information are more rapidly identified
and addressed. Referring now to FIG. 2, a flow diagram depicts the
initial menu node presented to a caller to provide information for
the six frequent tasks identified by the task analyzer. The process
begins at step 24, for instance, with the calling of a
predetermined phone number by a customer to establish communication
with IVR 10, and proceeds to a general greeting at step 26. The
general greeting identifies the purpose of the automated voice and
touch-tone service as providing information in response to caller
requests and instructs the caller how to make an opening statement
regarding the purpose of the call or to wait for a list of menu
items. For instance, the general greeting explains that the caller
has the option of relying on voice recognition or touch-tone inputs
to navigate the menu by stating an utterance or inputting a DTMF
tone at any time.
[0023] After the greeting, the process proceeds to step 28 where
the caller is instructed to "Please identify your task," followed
by a pause to allow an opening statement utterance. IVR 10 then
lists the menu options and an associated DTMF tone in a
predetermined order, such as in the order of frequency of the
requests by callers. For instance, an example script read by IVR 10
at step 28 is:
[0024] "Please identify your task."
[0025] "One. How to use a service."
[0026] "Two. Sales information about a service."
[0027] "Three. What services do I currently have."
[0028] "Four. How much does a service cost."
[0029] "Five. Get information about one of my services."
[0030] "Six. Who is my long distance carrier."
[0031] The order for the list of menu options corresponds to the
frequency with which requests for each type of information are made
as determined by task analyzer 20. The caller may select a type of
information by an appropriate utterance or by inputting a DTMF
tone.
[0032] At step 30, IVR 10 accepts the caller input, determines the
menu node selected by the caller input and advances the input to
task analyzer 20 for classification of the task requested by the
caller. Analysis of user tasks is thus updated to maintain a
current list of frequently requested tasks. For instance, automated
analysis may be performed to determine caller tasks and associated
utterances or, alternatively, opening statements may be recorded
for later play back and analysis. Frequent analysis allows prompt
and efficient updates to menu 18 to improve response times and
flows for caller requests for information by allowing adjustments
to menu 18 in a timely manner. For instance, if a network problem
arises which causes a surge in caller requests for post-purchase
information, such as a surge in requests for Internet service
status when the network has difficulty, task analyzer 20 provides
notification through menu GUI 22 so that menu 18 is updated to an
order corresponding with the increased frequency of requests for
post-purchase information. Thus, callers are presented with the
option to request post-purchase information more quickly, reducing
the time needed by callers to accomplish a task. This reduces
caller frustration for the most frequent call information
classifications and also reduces the resource demands on the call
center equipment, such as voice recognition and text to speech
processors, thereby reducing the overall cost of the
application.
[0033] IVR 10 determines the category of the task associated with
the caller's request for information and forwards the caller to a
menu node associated with the task. For instance, at step 34 the
caller is advanced to a menu node for information on how to use a
service with either an appropriate utterance or the selection of
DTMF tone 1. Examples of utterances that result in advancing to a
menu node on how to use a service are: "How do I set up a service?"
"How do I use a service?" and "How does a service work?" Some
examples of services include call blocking, Caller ID, message
center, call forwarding and long distance. The likelihood of
accurate recognition of utterances by IVR 10 is improved by
narrowing the glossary of recognized terms while including a
variety of terms that may be associated with a user request for
information. Thus, for instance, examples of other types of
utterances that result in advancing to the "how to use" menu node
associated with step 34 include: "Can I?" "What ways can I?" "Why
can't I" and "Why has?" In one embodiment, task analyzer 20 tracks
terms uttered by callers to associate terms with task categories,
resulting in more accurate application of voice recognition to
determine tasks.
[0034] From step 30, IVR 10 advances callers to menu nodes
associated with tasks based on either the DTMF tone or utterance
input of the user. For instance, at step 36, a caller is advanced
to a pre-purchase menu node upon selection of DTMF tone 2 or
recognition of an utterance associated with a request for
pre-purchase information, such as: "I want a service." "Is a
service available?" "Can I get a new service?" "Can you send?" "I
did not get information." and "What does a service cover?" At step
38, a caller is advanced to a current services menu node upon
selection of DTMF tone 3 or recognition of an utterance associated
with a request for information about the caller's current services,
such as: "Do I have a service?" or "What services do I have?" At
step 40, a caller is advanced to a cost menu node upon selection of
DTMF tone 4 or recognition of an utterance associated with a
request for information about service costs, such as: "How much
does it cost to add a service?" "How much is a service?" and "How
much does it cost to get a new service?" At step 42, a caller is
advanced to a post-purchase menu node upon selection of a DTMF tone
of 5 or recognition of an utterance associated with a request for
post-purchase information, such as: "I want to change a service."
"Who is my service provider?" "Is my service still the same?" "My
service connection is bad." and "Can I password protect my
service?" At step 44, a caller is advanced to a long distance menu
node upon selection of a DTMF tone of 6 or recognition of an
utterance associated with a request for information about a
caller's long distance provider, such as: "What long distance
company do I have?" or "Why do I have this company as my long
distance provider?" If a caller fails to select a recognized DTMF
tone or utterance, the process proceeds to step 46 to allow the
caller to select interaction with an operator.
[0035] Referring now to FIG. 3, a flow diagram depicts the process
for presenting "how to use" information. A caller request for "how
to use" information at step 34 of FIG. 2 initiates the process at
step 48 of FIG. 3. At step 50, the caller is provided with a
greeting that explains the purpose of the menu node, the services
for which information is available, and the availability of either
voice recognition or touch-tone inputs for the caller to request
information. At step 52, the caller is requested to input the
service of interest and the caller's area code or state. For
instance, a script instructs the caller to state a service for
which information is available, such as call blocking, caller ID,
message center and call forwarding, or to input a DTMF tone of 1
through 4 with each tone associated with a service. The caller is
also instructed to input an area code or state since service
offerings may vary by calling area. The area code and state
information may be input by a caller utterance or by using the
phone keypad.
[0036] At step 54, IVR 10 determines the service and service area
input by the caller and, at step 56, forwards the input to task
analyzer 20. Analysis of caller inputs is useful in several ways.
For instance, relating caller inputs at step 54 with opening
statements at 30 allows more specific identification of a caller's
initial task and the adapting of voice recognition nodes and
glossaries so that opening statements have an increased likelihood
of resulting in advancement to an appropriate menu node. Analysis
of caller inputs at step 56 also aids in tailoring the "how to use"
menu node so that callers are advanced to desired information in a
rapid and accurate manner. For instance, the glossary for the menu
node is adapted to recognize expected terms, thus improving
recognition speed and accuracy. In addition, the menu items are
arranged in terms of the frequency in which they are chosen, such
that the more frequently selected menu items are listed higher on
the menu.
[0037] At step 58, IVR 10 retrieves the service information
available for the area input by the caller. For instance, service
offerings and functionality may vary based on service areas. At
step 60, a script presents the caller with the types of information
available for the selected service and instructs the caller to
select desired information by either a voice utterance or DTMF
selection. For instance, the caller is presented with four types of
information for the identified service. One type of information is
a description of service functionality such as the definition of
the service and its objectives. Another type of information is set
up information, such as instructions for turning the service on and
off. Another type of information is usage information, such as
instructions on how to use the service. Another type of information
is how to change service functions. The script for each service
varies dependent upon the type of information available for the
service, and includes DTMF inputs as well as voice recognition
inputs for the caller to select desired information.
[0038] At step 62, the caller inputs a DTMF tone or speech
utterance to navigate for the four types of identified information
and the caller's input is forwarded to task analyzer 20. For
instance, if the caller inputs a DTMF tone of 1 or requests a
definition for a service, the process proceeds to step 64 to
provide explanations of definitions and functions for the
identified service. If a caller selects a DTMF tone of 2 or
requests instructions for set up of a service, the process proceeds
to step 66 for instructions on the set up of the identified
service. If a caller selects a DTMF tone of 3 or requests
instructions for how to use the identified service, the process
proceeds to step 68 for an explanation of how to use the service.
If a caller selects a DTMF tone of 4 or requests changes to the
settings of an identified service, the process proceeds to step 70
for instructions to perform the requested change.
[0039] Referring now to FIG. 4, a flow diagram depicts the process
for providing cost information for identified services. The process
begins at step 72 with the identification of a request for cost
information at step 40 of FIG. 2. The process proceeds through a
greeting at step 74 with an explanation of the information
available at the node through the selection of either a DTMF tone
or voice utterance. At step 76, the caller is requested to input
the service for which cost information is sought and the area code
or state of the caller. For instance, frequent requests for cost
information include requests for information on the cost of a
second telephone line, caller ID, or other new services that are
listed along with DTMF tone selection options.
[0040] At step 78, the caller identifies the service of interest
with a DTMF tone or utterance to IVR 10, which forwards the
selection to task analyzer 20 at step 80 as previously described.
At step 82, the cost information for the identified service or
package of services is retrieved for the identified state or area
code and, at step 84 presented to the caller. Although the initial
task of the caller was to obtain cost information, such requests
for cost information are often followed by requests to learn more
about or purchase identified services. At step 86, the caller is
presented with options to find the cost of another service,
purchase the identified service of terminate the call. At step 88,
the caller inputs a DTMF tone or utterance, which is forward to
task analyzer 20. If the caller selects a DTMF tone of 1 or states
a request for cost information of another service, the process
proceeds to step 90 and returns to step 82 to allow the user to
select another service. If the caller selects a DTMF tone of 2 or
states a request for purchase information, the process proceeds to
step 92 to advance the caller to a menu node of service purchase.
If the caller selects a DTMF tone of 3 or requests termination of
the call, the process proceeds to step 94 at which the call is
terminated.
[0041] Referring now to FIG. 5, a flow diagram depicts the process
for providing "what service do I have" information for a caller.
The process begins at step 96, such as from a request for such
information at step 38 of FIG. 2. At step 98, the caller is
provided a greeting that explains the purpose of the menu node and
the availability of DTMF or voice utterance selection. At step 100,
the caller is requested to provide the area code and telephone
number to identify the services. At step 102, the caller inputs the
area code and telephone number, either with an utterance or DTMF
tones, and the input is provided to task analyzer 20 at step 104.
At step 106, the requested information is retrieved and, at step
108, the caller is presented with a list of service names that are
provided to the identified phone number.
[0042] When a caller requests information regarding existing
services, requests related to the existing services frequently
follow. At step 110, the caller is presented with menu options that
address information requests or tasks that frequently follow
requests for information on existing services. For instance, the
caller is presented with options to select DTMF tones or state
utterances to obtain instructions for use of services, billing
costs or questions, handling moves by the caller, purchase of other
services or terminating the call. At step 112, the caller inputs a
DTMF tone or statement to accomplish a desired task, and the
caller's input is forwarded to task analyzer 20. If the caller
requests information on how to use a service, at step 114 the
caller is advanced to a "how to use" menu node, such as at steps 48
or 68 of FIG. 3. If the caller requests information on billing or
costs, at step 116, the caller is advanced to a menu node that
provides billing or cost information. If the caller requests
information on handling a move, then at step 118 the caller is
advanced to a menu node that provides information on moving
telephone service and change of address. If the caller requests
information on purchasing additional services, then at step 120 the
caller is advanced to a menu node to allow service purchase. If the
caller requests termination of the call, then at step 122, the
caller is advanced to a menu node for call termination.
[0043] Referring now to FIG. 6, a flow diagram depicts the process
for providing identification of long distance provider information.
The process begins at step 124, such as from a caller request for
long distance provider information at step 44 of FIG. 2, and
proceeds to step 126 at which the caller is greeted with an
explanation of the menu node. At step 128, the caller is requested
to provide the area code and telephone number either with DTMF
tones or voice utterances and, at step 132 the caller input is
provided to task analyzer 20. At step 134, the long distance
carrier information for the input area code and telephone number
are retrieved and, at step 136, the long distance carrier
information is provided to the caller.
[0044] Requests by callers for long distance carrier information
frequently result in additional requests for related information,
such as reporting of "slam" changes, billing inquiries, and
requests to change the long distance carrier. At step 138, the
caller is presented with menu options for requests related to long
distance carrier information, including DTMF tone or voice
utterance input options. At step 140, the caller inputs a selection
which is forwarded to task analyzer 20. If the caller requests to
report a slamming event, then at step 142 the caller is advanced to
a menu node for reporting slamming complaints. If the caller
requests billing or cost information, then at step 144 the caller
is advanced to a menu node for providing cost or billing
information. If the caller requests a change to the long distance
provider, then at step 146 the caller is advanced to a menu node
for changing the long distance provider. If the caller requests to
terminate the call, then at step 148 the caller is advanced to a
menu node for terminating the call.
[0045] Referring now to FIG. 7, a flow diagram depicts one
embodiment for updating menu content with caller feedback. The
process begins at step 750, such as upon the completion by a caller
of retrieval of How To Use at step 68 of FIG. 3. At step 752, the
caller is asked to provide feedback, such as comments about the
quality, validity and usefulness of the information provided. A
content analyzer 754 receives, records and analyzes the feedback.
For instance, content analyzer 754 may record caller comments for
playback by system administrators, or may evaluate comments based
on speech recognition, resulting in updates to the content of the
associated menu node.
[0046] Although the present invention has been described in detail,
it should be understood that various changes, substitutions and
alterations can be made hereto without departing from the spirit
and scope of the invention as defined by the appended claims.
* * * * *