Caller adaptive voice response system

• US 5,493,608 A
• Filed: 03/17/1994
• Issued: 02/20/1996
• Est. Priority Date: 03/17/1994
• Status: Expired due to Term

• Current Assignee: Contact Solutions Incorporated
• Original Assignee: Alpha Logic, Incorporated
• Inventors: O'Sullivan, Daniel
• Primary Examiner(s): Hofsass, Jeffery A.
• Assistant Examiner(s): WEAVER, SCOTT LOUIS
• Application Number:
  US08/210,031
• Time in Patent Office: 705 Days
• Field of Search: 379/67, 379/88, 379/89, 379/72-77, 379/79-80, 381/43-44, 381/110, 381/51, 395/2.6, 395/2.84, 395/2.85, 395/2.79
• US Class Current: 379/88.04
• CPC Class Codes: H04M 3/493 Interactive information services, e.g. directory enquiries ; Arrangements therefor, e.g. interactive voice response [IVR] systems or voice portals

Abstract

A interactive voice response system that automatically adapts to suit the speed at which the caller interacts with the system. The caller interacts with the system utilizing the keypad buttons on a DTMF-type telephone or by speaking verbal instructions over any type of telephone, and hears spoken instructions and information over the telephone in response to those entries. The system is programmed to measure the response times of the caller and adjust the playing speed of the application dialogue'"'"'s voice messages accordingly using a caller adaptive response algorithm incorporated into the application software of the voice response system. If the caller is responding relatively fast and without error to the voice message prompts, the caller adaptive response algorithm software will gradually speed up subsequent voice message prompts to a preset, reasonable maximum speech rate. If the caller is responding more slowly to the voice message prompts or is making errors in their responses, the caller adaptive response algorithm software will slow down subsequent voice message prompts to a reasonable, preset minimum speech rate. This process continues throughout the application dialogue such that said system is always closely approximating the natural conversational pace of the caller.

Claims

1. A caller adaptive voice response system which adapts to a conversational pace of a caller, comprising:

(a) means for measuring an amount of time indicative of the time required for said caller to respond to a voice message spoken by said system at a predetermined speaking rate; and

(b) means for adjusting a subsequent speaking rate of a subsequent voice message to be spoken by said system, said adjusting means being responsive to said measuring means so as to increase said subsequent speaking rate of said subsequent voice message if said amount of time was shorter in duration than a first predetermined amount of time.

2. The apparatus according to claim 1 further comprising:

means for determining if a caller response is an expected response and thereby a valid response and determining if a caller response is not a valid response and therefore an erroneous response; and

means for decreasing said speaking rate of said subsequent voice message if said caller performs an action selected from a group consisting of no response and an erroneous response to said voice message preceding said subsequent voice message.

3. The apparatus according to claim 1, wherein said system speaks a plurality of voice messages in a predetermined order, and wherein said measuring means measures said amount of time required for said caller to respond to each one of said plurality of voice messages spoken by said system, and wherein said adjusting means adjusts the speaking rate of each one voice message of said plurality of voice messages in accordance with said amount of time required for said caller to respond to the one voice message which immediately preceded the one voice message to be next spoken by said system.

4. The apparatus of claim 1, further including means for storing a plurality of speech file database portions, each speech file database portion of said plurality being recorded at different speaking rates, and wherein said adjusting means adjusts said subsequent speaking rate by accessing a speech file to be spoken by the system from a selected one of said speech file database portions of said plurality in accordance with said amount of time measured by said measuring means.

5. The apparatus of claim 1, wherein said caller is prompted, by a prompt spoken by the system and selected from a group comprising said voice message and said subsequent voice message, to respond to said system by depressing at least one button on a DTMF-type telephone keypad associated with a telephone being used by said caller to call said system.

6. The apparatus of claim 1, wherein said caller is prompted, by a prompt spoken by the system and selected from a group comprising said voice message and said subsequent voice message, to respond to said system by speaking at least one word into a telephone being used by said caller to call said system.

7. The apparatus according to claim 1, further comprising means for decreasing said subsequent speaking rate of said subsequent voice message if said amount of time was longer in duration than a second predetermined amount of time.

8. A method of adapting a voice response system to a conversational pace of a caller, comprising the steps of:

measuring an amount of time indicative of the time required for said caller to respond to a voice message spoken by the system at a predetermined speaking rate; and

adjusting, in accordance with the caller response time measurement obtained during said measuring step, a subsequent speaking rate of a subsequent voice message to be subsequently spoken by the system.

9. The method according to claim 8, wherein said adjusting step includes increasing said subsequent speaking rate of said subsequent voice message if said caller response time measured in said measuring step was shorter in duration than a first predetermined amount of time.

10. The method according to claim 9, wherein said adjusting step includes decreasing said subsequent speaking rate of said subsequent voice message if said caller response time measured in said measuring step was longer in duration than a second predetermined amount of time.

11. The method according to claim 8, wherein said adjusting step includes decreasing said subsequent speaking rate of said subsequent voice message if said caller response time measured in said measuring step was longer in duration than a predetermined amount of time.

12. The method according to claim 8, wherein said voice message of said measuring step is the first prompt spoken by said system.

13. The method according to claim 8, wherein said measuring step is repeated for each caller response to a plurality of voice messages, and a subsequent speaking rate for each of a plurality of subsequent voice messages is adjusted by said adjusting step in response to the immediately preceding caller response time measurement performed during said measuring step.

14. The method according to claim 8, further including the steps of:

determining if a caller response is an expected response and is thereby a valid response;

determining if a caller response is not a valid response and is therefore an erroneous response; and

decreasing said subsequent speaking rate of said subsequent voice message if said caller performs an action selected from a group consisting of no response or an erroneous response to a voice message which precedes said subsequent voice message.

15. A caller adaptive telephonic voice response system capable of interacting via a telephone line with a caller by generating voice messages recognizable by said caller as audible speech spoken at a predetermined speaking rate and by detecting a response from said caller to said voice message within a variable response time interval, the system comprising:

(a) means for generating over said telephone line at least one voice message recognizable by said caller as audible speech spoken at an initial predetermined speaking rate and for generating over said telephone line at least one subsequent voice message recognizable by said caller as audible speech spoken at a subsequent predetermined speaking rate, said subsequent voice message being generated by said system subsequent to said voice message;

(b) means for detecting over said telephone line a response from said caller;

(c) means for measuring said variable response time interval, said variable response time interval being indicative of the elapsed time between the end of a voice message most recently spoken by said system and the detection of said response from said caller to said most recently spoken voice message; and

(d) means for adjusting said initial predetermined speaking rate to said subsequent predetermined speaking rate, said adjusting means being responsive to said measuring means so as to adjust said subsequent predetermined speaking rate to a rate higher than said initial predetermined speaking rate if said variable response time interval is shorter in duration than a first predetermined amount of time.

16. An apparatus according to claim 15, wherein said adjusting means adjusts said subsequent predetermined speaking rate to a rate lower than said initial predetermined speaking rate if said variable response time interval is longer in duration than a second predetermined amount of time.

17. An apparatus according to claim 15 further comprising means for determining if a caller response is an expected response and thereby a valid response and determining if a caller response is not a valid response and therefore an erroneous response, and wherein said adjusting means is further responsive to said detecting means so as to adjust said subsequent predetermined speaking rate to a rate lower than said initial predetermined speaking rate if said caller performs an action selected from a group consisting of no response and an erroneous response.

18. An apparatus according to claim 16, wherein said first predetermined amount of time and said second predetermined amount of time are equal.

19. An apparatus according to claim 15, further comprising means for measuring a second variable response time indicative of the elapsed time between the end of said subsequent voice message and the detection of a response from said caller to said subsequent voice message, and wherein said adjusting means adjusts the speaking rate of a further subsequent voice message spoken after said subsequent voice message in accordance with said second variable response time interval.

20. An apparatus according to claim 15, wherein said adjusting means comprises:

(a) a plurality of pre-recorded speech file database portions containing a plurality of pre-recorded speech files having verbal contents and from which a voice message may be generated by said generating means at a database portion-specific preselected speaking rate, the verbal content of each of said pre-recorded speech file database portion being recorded at different database portion-specific preselected speaking rates; and

(b) means for controllably selecting from said plurality of database portions a pre-recorded speech file database portion which has a database portion-specific speaking rate closest in value to said subsequent predetermined speaking rate; and

wherein said generating means generates said subsequent voice message from said pre-recorded speech file database portion selected by said selecting means.

21. An apparatus according to claim 15, wherein said caller is instructed by at least one of said voice messages to respond to said system by depressing at least one button on a DTMF-type telephone keypad of a telephone being used by said caller to interact with said system.

22. An apparatus according to claim 15, wherein said caller is instructed by at least one of said voice messages to respond to said system by speaking at least one verbal utterance into a telephone being used by said caller to interact with said system.

23. An apparatus according to claim 15, wherein said adjusting means further comprises means for determining said subsequent predetermined speaking rate, said determining means comprising:

(a) means for increasing said subsequent predetermined speaking rate to a rate higher than said initial predetermined speaking rate, said increasing means comprising;

(i) means for adding the numerical value one to the numerical value of said variable response time interval to arrive at a sum;

(ii) means for modifying said sum if said sum is a fractional value by increasing the value of said sum to the next highest whole number to obtain a modified sum;

(iii) means for dividing a first predetermined fixed number by a divisor selected from a group consisting of said sum and said modified sum to arrive at a numerical value to be added; and

(iv) means for adding said numerical value to be added to a numerical value equal to an average number of words per minute spoken by said system at said initial predetermined speaking rate to arrive at a second numerical value;

(v) means for setting an average number of words per minute spoken by said system at said subsequent predetermined speaking rate to said second numerical value; and

(vi) means for limiting said subsequent predetermined speaking rate to a maximum speaking rate.

24. An apparatus according to claim 23, further comprising means for decreasing said subsequent predetermined speaking rate to a rate lower than said initial predetermined speaking rate, said decreasing means comprising:

(i) means for reducing said initial predetermined speaking rate by a predetermined fixed number of words per minute to arrive at said subsequent predetermined speaking rate; and

(ii) means for limiting said subsequent predetermined speaking rate to a minimum speaking rate.

25. An apparatus according to claim 23, wherein said determining means is a software program operating in a general purpose digital computer.

26. An apparatus according to claim 24, wherein said initial predetermined speaking rate is approximately 200 words per minute, said first predetermined amount of time is approximately 5 seconds, said second predetermined amount of time is approximately 5 seconds, said maximum speaking rate is approximately 300 words per minute, said first predetermined fixed number is approximately 20, said predetermined fixed number of words per minute is approximately 20, and said minimum speaking rate is approximately 100 words per minute.

27. A method of adapting a caller adaptive telephonic voice response system capable of interacting via a telephone line with a caller by generating voice messages recognizable by the caller as audible speech spoken at a predetermined speaking rate and by detecting a response from said caller to said voice message within a variable response time interval, the method comprising:

(a) generating over said telephone line a first voice message recognizable by said caller as audible speech spoken at an initial predetermined speaking rate;

(b) detecting over said telephone line a response from said caller to said first voice message;

(c) measuring as said variable response time interval a time interval indicative of the elapsed time between the end of said first voice message generated in step (a) and the detection of step (b);

(d) generating over said telephone line at least one subsequent voice message recognizable by said caller as audible speech spoken at a subsequent predetermined speaking rate, said subsequent voice message being generated by said system subsequent to said first voice message; and

(e) adjusting said initial predetermined speaking rate to said subsequent predetermined speaking rate in response to said variable response time interval measured in step (c) so as to adjust said subsequent predetermined speaking rate to a rate higher than said initial predetermined speaking rate if said variable response time interval is shorter in duration than a first predetermined amount of time.

28. A method according to claim 27 wherein said adjusting step further adjusts said subsequent predetermined speaking rate to a rate lower than said initial predetermined speaking rate if said variable response time interval is longer in duration than a second predetermined amount of time.

29. A method according to claim 28 further comprising the steps of determining if said caller response is an expected response and thereby a valid response, and determining if said caller response is not a valid response and therefore an erroneous response, and wherein said adjusting step further comprises adjusting said subsequent predetermined speaking rate to said rate lower than said initial predetermined speaking rate if said caller performs an action selected from a group consisting of no response and an erroneous response.

30. A method according to claim 27, wherein said adjusting step adjusts said subsequent predetermined speaking rate of said subsequent voice message in accordance with said variable response time interval of a caller response measured immediately prior to said system speaking said subsequent voice message.

31. A method according to claim 27, wherein said step (e) comprises:

(i) storing a plurality of pre-recorded speech file database portions containing a plurality of pre-recorded speech files having verbal contents and from which a voice message may be generated at a database portion-specific preselected speaking rate, the verbal contents of each of said pre-recorded speech file database portions being recorded at different database portion-specific preselected speaking rates;

(ii) controllably selecting from said plurality of pre-recorded speech file database portions a pre-recorded speech file database portion having a database portion-specific speaking rate closest in value to said subsequent predetermined speaking rate; and

wherein said step (d) comprises generating said subsequent voice message from said pre-recorded speech file database portion selected in said step (ii).

32. A method according to claim 27, further comprising the step of instructing said caller, by at least one of said voice messages, to respond to said system by depressing at least one button on a DTMF-type telephone keypad associated with a telephone being used by said caller to call said system.

33. A method according to claim 27, further comprising the step of instructing said caller, by at least one of said voice messages, to respond to the system by speaking at least one verbal utterance into a telephone being used by said caller to call said system.

34. A method according to claim 27, wherein said adjusting step further comprises:

(i) increasing said subsequent predetermined speaking rate to a rate higher than said initial predetermined speaking rate by;

(1) adding a predetermined numerical value to the numerical value of said variable response time interval to arrive at a sum;

(2) modifying said sum if said sum is a fractional value by increasing the value of said sum to the next highest whole number to obtain a modified sum;

(3) dividing a first predetermined fixed number by a divisor selected from a group consisting of said sum and said modified sum to arrive at a numerical value to be added;

(4) adding said numerical value to be added to a numerical value equal to an average number of words per minute spoken by said system at said initial predetermined speaking rate to arrive at a second numerical value;

(5) setting an average number of words per minute to be spoken by said system at said subsequent predetermined speaking rate to said second numerical value; and

(6) limiting said subsequent predetermined speaking rate to a maximum speaking rate.

35. A method according to claim 28, wherein said adjusting step further comprises:

(i) decreasing said subsequent predetermined speaking rate to a rate lower than said initial predetermined speaking rate by;

(1) reducing said initial predetermined speaking rate by a predetermined fixed number of words per minute to arrive at said subsequent predetermined speaking rate; and

(2) limiting said subsequent predetermined speaking rate to a minimum speaking rate.

36. A method according to claim 34, wherein said steps (1) through (5) are performed by a software program operating in a general purpose digital computer.

37. A method according to claim 35, wherein said steps (1) and (2) are performed by a software program operating in a general purpose digital computer.

38. A method according to claim 34, wherein said initial predetermined speaking rate is approximately 200 words per minute, said predetermined value is 1, said first predetermined amount of time is approximately 5 seconds, said maximum speaking rate is approximately 300 words per minute, and said first predetermined fixed number is approximately 20.

39. A method according to claim 35, wherein said initial predetermined speaking rate is approximately 200 words per minute, said predetermined expected amount of time is approximately 5 seconds, said predetermined fixed number of words per minute is approximately 20, and said minimum speaking rate is 100 words per minute.