JP2006345181A - Outbound campaign system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To allow a call center which uses an IP network to decide whether an answering side is a person or FAX equipment, an automatic answering telephone set, etc., other than a person. <P>SOLUTION: An automatic calling device 61 automatically dials telephone numbers previously recorded in a storage section. A speech stream connecting device 64 analyzes the frequency of a speech stream obtained in response to automatic dialing, decides the continuance of the speech, and discriminates between a busy state and FAX answering based upon the result of the frequency analysis and the result of the continuance decision. Further, whether a speech is received is checked by speech packets of the speech stream and patterned and it is decided whether the speech stream is an automatic answering telephone voice or talkie voice based upon the obtained patterns and patterns previously stored in the storage section. Based upon those decision results, it is specified whether the answering side is a person or not. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an out-of-band campaign system used in a call center using an IP (Internet Protocol) network.

Conventionally, an out-of-band campaign system is used in a call center using an analog line. In this system, an automatic calling device automatically makes a call based on a customer list, and when the called party answers, it determines whether the other party is a person, whether it is a FAX, an answering machine, or a talky using a voice recognition device, Only calls that require connection are connected to the operator for efficient operation of the call center.
“Avaya Predictive Dialing System (PDS) Enhanced Call Progress Analysis (ECPA)” Avaya Japan, Inc. March 2001

On the other hand, in a call center using an IP network, call control is performed using a SIP (Session Initiation Protocol) protocol, and the state of a response partner is determined based on the content of the SIP response code at the time of connection. However, from the response code, it can only be determined whether it is connected to the other party, and if the other party is a person, a fax, an answering machine, or a talkie, it does not know what the contents are. There was a problem of having to connect the call to the operator.
The present invention has been made in consideration of the above circumstances, and its purpose is to be able to determine whether the answering party is a person, a fax other than a person, an answering machine, or the like in a call center using an IP network. To provide a band campaign system.

  The present invention has been made in order to solve the above-mentioned problems. The invention according to claim 1 is an out-of-band campaign system using an IP network, and makes a call based on a telephone number recorded in a storage unit in advance. Automatic call means for performing the analysis, and analyzing the voice stream obtained in response to the call made by the automatic call means to detect whether the response partner is a person or a person, An out-of-band campaign system comprising audio stream analysis means connected to an operator terminal.

According to a second aspect of the present invention, in the out-of-band campaign system according to the first aspect, the voice stream analyzing means performs frequency analysis of the voice stream, and a response partner is a person based on the frequency detected by the analysis. It is characterized by detecting whether it is a person or a person other than a person.
According to a third aspect of the present invention, in the out-of-band campaign system according to the second aspect, the audio stream analyzing means performs frequency analysis on the audio stream, then determines an intermittent time of the audio, and results of the frequency analysis And detecting whether the response partner is a person or a person based on the result of the intermittent time determination.

According to a fourth aspect of the present invention, in the out-of-band campaign system according to the first aspect, the voice stream analyzing means checks the presence / absence of voice for each voice packet of the voice stream and patterns the obtained pattern and It is characterized in that it is detected whether a response partner is a person or a person other than a person based on a pattern stored in advance in a storage unit.
The invention according to claim 5 is an out-of-band campaign system using an IP network, wherein an automatic call is made based on a telephone number recorded in advance in a storage unit, and an audio stream obtained in response to the automatic call Frequency analysis, and then determining a voice intermittent time, and determining whether the response partner is a person or a person based on the result of the frequency analysis and the result of the intermittent time determination. This is a response partner determination method in the band campaign system.

  According to a sixth aspect of the present invention, in the response partner determination method in the out-of-band campaign system according to the fifth aspect, the presence / absence of voice is examined and patterned for each voice packet of the voice stream, and the obtained pattern and stored in advance Based on the pattern stored in the unit, it is determined whether the response partner is a person or a person other than the person, and based on the result of the voice packet in addition to the result of the frequency analysis and the result of the intermittent time determination It is characterized by determining whether the response partner is a person or a person other than a person.

   According to the seventh aspect of the present invention, a procedure for making an automatic call based on a telephone number recorded in advance in a storage unit to a computer that performs an out-of-band campaign using an IP network, and a method obtained according to the automatic call are obtained. A procedure for frequency analysis of the received audio stream, a procedure for determining the intermittent time of the audio, and determining whether the response partner is a person or a person based on the result of the frequency analysis and the result of the intermittent time determination The program for executing the procedure to perform.

  According to the present invention, in a call center using an IP network, it is possible to determine whether a response partner is a person, a fax other than a person, an answering machine, or the like. As a result, it is possible to prevent the answering party from connecting a call other than a person to the operator, and it is possible to efficiently operate the call center by eliminating the operator's wasteful operation. Further, according to the present invention, there is an advantage that the response partner can be determined only by frequency analysis or voice packet analysis without using an expensive voice recognition device.

  Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of an out-of-band campaign system according to an embodiment of the present invention. In this figure, 1 to 3 are a customer telephone, customer FAX, customer answering machine, 4 is a PSTN (Public Switched Telephone Networks) network (or IP network), and 5 is a missing number guide and number change provided in the public network. A talkie sending device 6 for performing guidance and the like is an out-of-band campaign system according to an embodiment of the present invention.

  In the out-of-band campaign system 6, reference numeral 61 denotes an automatic calling device that automatically makes a call based on a customer list. Reference numeral 62 denotes CTI (Computer Telephony Integration), which integrates telephone calls and computer customer information. 63 is a call control device that controls connection and disconnection of calls. Reference numeral 64 denotes an audio stream connection device that determines whether the answering party is a person, a fax other than a person, an answering machine, or the like based on an audio stream (voice packet) of a call. Is connected to the operator terminal 65. Reference numeral 66 denotes a network connection device that connects a call to the PSTN network 4.

  In such a configuration, the calling instruction to the customer issued by the automatic calling device 61 reaches the call control device 63 via the CTI 62 as shown in the sequence diagram of FIG. Driven and called to the PSTN network 4. When the receiving side answers, the call control device 63 connects the voice stream between the network connection device 66 and the voice stream connection device 64, and the other party is the customer telephone 1, customer FAX 2, customer answering machine 3 at the voice stream connection device 64. The PSTN network 4 is detected as various talkies 5. As will be described in detail below, this detection uses a simple algorithm for recognizing the presence / absence of a voice and a voice pattern, and only when the customer telephone 1 responds (that is, a person answers), the call control device 63 transmits a voice stream. Connect to the operator terminal 65.

  In the prior art, since there is no portion indicated by symbol A in FIG. 2, even an invalid call such as an answering machine or a fax is connected to the operator 65.

Next, details of processing performed in the audio stream connection device 64 will be described with reference to FIGS.
FIG. 3 is a diagram showing the configuration of software executed in the audio stream connection device 64. The audio stream connection function program 12 is provided on the operating system 11, and the busy detection function program 13, FAX is provided on the operating system 11. A detection function program 14, an answering machine detection function program 15, and a talkie detection function program 16 are provided.

  FIG. 4 is a flowchart of the busy detection function program. First, frequency analysis is performed on the voice stream from the other party by FFT (Fast Fourier Transform), and the frequency included in the voice stream is detected (step Sa1). Then, it is checked whether or not the detected frequency is the business tone frequency (400 Hz) (step Sa2). If the detected frequency is the business tone frequency, then a voice intermittent time determination is performed (step Sa3). That is, in the case of a visit tone, since 0.5 sec on / 0.5 sec off is repeated, it is determined whether or not the voice is intermittently repeated at this timing (step Sa4). If it has been repeated (the determination at step Sa4 is “YES”), it is determined as “busy”.

On the other hand, if the determination in step Sa2 is “NO”, and if the determination in step Sa4 is “NO”, it is determined that it is not “busy”.
FIG. 5 is a flowchart of the FAX detection function program. First, similarly to the above, the audio stream from the other party is subjected to frequency analysis by FFT (Fast Fourier Transform), and the frequency included in the audio stream is detected (step Sb1). Then, it is checked whether or not the detected frequency is a response sound frequency (beep sound) in the case of FAX (step Sb2). If the frequency is different, it is determined that it is not a FAX.

  FIG. 6 is a flowchart of the answering machine detection function program. First, the presence / absence of voice is checked for each voice packet of the voice stream and patterned (step Sc1). In this case, when the voice level of the packet is higher than a predetermined threshold, the voice is present, and when the voice level is low, the voice is absent. In order to eliminate noise, it is assumed that the voice continues for a certain period of time, and the silence at the word breaks is treated as if the word continues. If no voice continues, no voice. Next, it is checked whether or not the time when the voice is continuous and the time when there is no subsequent voice is within the range stored in the internal memory in advance (step Sc2). If it is within the range, it is determined that there is no answering machine (step Sc3 is “YES”), and if it is not within the range, it is determined that it is not the answering machine (step Sc3 is “NO”).

  FIG. 7 is a flowchart of the talkie detection function program. First, in the same manner as described above, the presence / absence of voice is checked for each voice packet of the voice stream and patterned (step Sd1). Next, the obtained pattern is collated with a plurality of talkie sound patterns stored in advance in the internal memory (step Sd2). If there is a matching pattern, it is determined that the pattern is a talkie (step Sd3 is “YSE”). If there is no matching pattern, it is determined that the pattern is not a talkie (step Sd3 is “NO”). ").

  The above is the audio determination process performed in the audio stream connection device 64. The audio stream connection device 64 sequentially performs the four determinations described above, and determines that the audio stream is a human voice when none of the audio streams corresponds to the audio stream, and connects the audio stream to the operator 65. In any case, as shown in FIG. 2, the detection result is notified to the call control device 63, the call control device 63 sends a call disconnection request to the network connection device 66, and the network connection device 66 sends a PSTN network. A disconnection request is output to 4, and the call is disconnected by the PSTN network 4. The call disconnection notification is notified from the PSTN network 4 to the network connection device 66, and the network connection device 66 sends a disconnection notification to the call control device 63. The call control device 63 receives this notification and sends the fact that the call has been disconnected and the reason to the automatic call device 61 via the CTI 62. The automatic calling device 61 records data related to the disconnection and the reason in the internal storage unit.

  The present invention is applied when the call center uses an IP network, and the connection partner may be an IP network or a PSTN.

  The present invention is used in a call center that conducts an out-of-band campaign.

It is a block diagram which shows the structure of the out-of-band campaign system by one Embodiment of this invention. It is a sequence diagram for demonstrating operation | movement of the embodiment. It is a figure which shows the structure of the software performed in the audio | voice stream connection apparatus 64 in FIG. 4 is a flowchart of software executed in the audio stream connection device 64; 4 is a flowchart of software executed in the audio stream connection device 64; 4 is a flowchart of software executed in the audio stream connection device 64; 4 is a flowchart of software executed in the audio stream connection device 64;

Explanation of symbols

1 ... Customer phone 2 ... Customer FAX
3 ... Customer answering machine 4 ... PSTN network 6 ... Out-of-band campaign system 61 ... Automatic calling device 62 ... CTI
63 ... Call control device 64 ... Voice stream connection device 65 ... Operator terminal 66 ... Network connection device

Claims (7)

In an out-of-band campaign system using an IP network,
Automatic calling means for making a call based on a telephone number recorded in the storage unit in advance;
Analyzing a voice stream obtained in response to a call made by the automatic calling means to detect whether the response partner is a person or a person other than the person, and if it is a person, a voice stream analysis for connecting to an operator terminal And an out-of-band campaign system.   2. The audio stream analysis unit according to claim 1, wherein the audio stream analysis unit performs frequency analysis on the audio stream and detects whether a response partner is a person or a person other than the person based on the frequency detected by the analysis. Out-of-band campaign system.   The audio stream analyzing means frequency-analyzes the audio stream, then determines an audio intermittent time, and the response partner is a person or a person other than the person based on the result of the frequency analysis and the result of the intermittent time determination The out-of-band campaign system according to claim 2, wherein:   The audio stream analysis means examines and patterns the presence or absence of audio for each audio packet of the audio stream, and based on the obtained pattern and the pattern stored in advance in the storage unit, the response partner is a person or a person other than the person The out-of-band campaign system according to claim 1, wherein it is detected. In an out-of-band campaign system using an IP network,
Make an automatic call based on the telephone number recorded in advance in the storage unit,
Frequency analysis of the audio stream obtained in response to the automatic call,
Next, determine the voice interruption time,
A response partner determination method in an out-band campaign system, wherein it is determined whether a response partner is a person or a person other than a person based on the result of the frequency analysis and the result of the intermittent time determination. Check the presence or absence of voice for each voice packet of the voice stream, determine whether the response partner is a person or a person based on the obtained pattern and the pattern stored in advance in the storage unit,
6. The method according to claim 5, wherein, in addition to the result of the frequency analysis and the result of the intermittent time determination, it is determined whether a response partner is a person or a person other than the person based on the determination result by the voice packet. Response partner determination method in the out-of-band campaign system. To a computer that conducts an out-of-band campaign using an IP network,
A procedure for making an automatic call based on a telephone number recorded in advance in the storage unit;
A frequency analysis of the audio stream obtained in response to the automatic call;
A procedure for determining the audio intermittent time;
A program for executing a procedure for determining whether a response partner is a person or a person other than a person based on the result of the frequency analysis and the result of the intermittent time determination.

Images