CN113709313B

CN113709313B - Intelligent quality inspection method, device, equipment and medium for customer service call data

Info

Publication number: CN113709313B
Application number: CN202111014361.8A
Authority: CN
Inventors: 徐寅磊
Original assignee: Ping An Technology Shenzhen Co Ltd
Current assignee: Ping An Technology Shenzhen Co Ltd
Priority date: 2021-08-31
Filing date: 2021-08-31
Publication date: 2022-10-25
Anticipated expiration: 2041-08-31
Also published as: CN113709313A

Abstract

The application relates to an artificial intelligence technology, and discloses an intelligent quality inspection method, device, equipment and medium for customer service call data, wherein the method comprises the following steps: acquiring customer service call data to be subjected to quality inspection according to the quality inspection instruction; identifying first voice data corresponding to consultants and second voice data corresponding to customer service staff; inputting the first voice data and the second voice data into a voice conversion model for text conversion, and acquiring corresponding first text information and second text information, wherein the first text information and the second text information are both marked with text timestamps corresponding to a time axis; sequencing the first text information and the second text information according to the text timestamp to obtain a call record text so as to obtain a first quality inspection result of the customer service call data; and receiving a secondary quality inspection request of the customer service responding to the first quality inspection result, performing quality inspection on the call data of the customer service according to the secondary quality inspection request, and acquiring a second quality inspection result so as to accurately perform intelligent evaluation on the call data of the customer service staff.

Description

Intelligent quality inspection method, device, equipment and medium for customer service call data

Technical Field

The application relates to the technical field of artificial intelligence, in particular to an intelligent quality inspection method, device, equipment and medium for customer service call data.

Background

Voice services are becoming more common today, for example, service providers may provide voice services to users based on customer service call centers, or via voice bots, etc. In order to further improve the quality of the voice service provided for the user, it is necessary to perform quality inspection on the voice service provided by the customer service staff, and in the prior art, the call records of the customer service staff and the client are identified through a voice recognition technology, so as to perform quality inspection on the voice service provided by the customer service staff, but when the voice recognition occurs, the quality inspection result on the voice service may be inaccurate.

Therefore, how to objectively and truly reflect the work evaluation of the customer service staff is a popular topic that is being researched by those skilled in the art.

Disclosure of Invention

The embodiment of the application provides a method, a device, equipment and a medium for intelligently inspecting customer service call data, which can accurately and intelligently evaluate the call data of customer service personnel.

In a first aspect, an embodiment of the present application provides an intelligent quality inspection method for customer service call data, including:

when a quality inspection instruction is received, acquiring customer service call data to be subjected to quality inspection evaluation according to the quality inspection instruction;

identifying a plurality of first voice data corresponding to consultants in the customer service call data and a plurality of second voice data corresponding to the customer service personnel according to a time axis corresponding to the customer service call data, wherein the first voice data and the second voice data are marked with voice time stamps corresponding to the time axis;

inputting a plurality of first voice data and a plurality of second voice data into a preset voice conversion model for text conversion so as to obtain a plurality of first text messages corresponding to the plurality of first voice data and a plurality of second text messages corresponding to the plurality of second voice messages, wherein the first text messages and the second text messages are marked with text timestamps corresponding to the voice timestamps;

sequencing the plurality of first text messages and the plurality of second text messages according to the text timestamps to obtain call record texts corresponding to the customer service call data;

acquiring a first quality inspection result of the customer service call data according to the call record text;

and receiving a secondary quality inspection request of the customer service responding to the first quality inspection result, and rechecking the call data of the customer service according to the secondary quality inspection request to obtain a second quality inspection result.

In a second aspect, an embodiment of the present application further provides an intelligent quality inspection apparatus for customer service call data, including:

the data acquisition module is used for acquiring customer service call data to be subjected to quality inspection evaluation according to the quality inspection instruction when the quality inspection instruction is received;

the identification module is used for identifying a plurality of first voice data corresponding to consultants in the customer service call data and a plurality of second voice data corresponding to the customer service personnel according to a time axis corresponding to the customer service call data, and the first voice data and the second voice data are marked with voice time stamps corresponding to the time axis;

the text conversion module is used for inputting the first voice data and the second voice data into a preset voice conversion model for text conversion so as to obtain a plurality of first text messages corresponding to the first voice data and a plurality of second text messages corresponding to the second voice messages, and the first text messages and the second text messages are marked with text timestamps corresponding to the voice timestamps;

the sequencing module is used for sequencing the first text information and the second text information according to the text timestamp so as to obtain a call record text corresponding to the customer service call data;

the first quality inspection module is used for acquiring a first quality inspection result of the customer service call data according to the call record text;

and the second quality inspection module is used for rechecking the customer service call data according to the secondary quality inspection request so as to obtain a second quality inspection result.

In a third aspect, an embodiment of the present application further provides a computer device, including a memory and a processor;

a memory for storing a computer program;

and the processor is used for executing the computer program and realizing the intelligent quality inspection query method for the customer service call data during the execution of the computer program.

In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, where a computer program is stored, and when the computer program is executed by a processor, the processor is enabled to implement the above method for intelligently inspecting quality of customer service call data.

The embodiment of the application provides an intelligent quality inspection method, device, equipment and medium for customer service call data, wherein the intelligent quality inspection method for the customer service call data comprises the following steps: when a quality inspection instruction is received, obtaining customer service call data to be subjected to quality inspection evaluation according to the quality inspection instruction; identifying a plurality of first voice data corresponding to consultants in the customer service call data and a plurality of second voice data corresponding to the customer service personnel according to a time axis corresponding to the customer service call data, wherein the first voice data and the second voice data are marked with voice time stamps corresponding to the time axis; inputting a plurality of first voice data and a plurality of second voice data into a preset voice conversion model for text conversion so as to obtain a plurality of first text messages corresponding to the plurality of first voice data and a plurality of second text messages corresponding to the plurality of second voice messages, wherein the first text messages and the second text messages are marked with text timestamps corresponding to the voice timestamps; sequencing the plurality of first text messages and the plurality of second text messages according to the text time stamps to obtain call record texts corresponding to the customer service call data; according to the call record text, obtaining a first quality inspection result of the customer service call data; and receiving a secondary quality inspection request of the customer service responding to the first quality inspection result, and rechecking the call data of the customer service according to the secondary quality inspection request to obtain a second quality inspection result, so that the call data of the customer service staff can be accurately and intelligently evaluated.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.

FIG. 1 is a flowchart illustrating steps of a method for intelligently inspecting customer service call data according to an embodiment of the present application;

FIG. 2 is a flow chart illustrating call data acquisition steps in the intelligent quality inspection method for customer service call data of FIG. 1;

FIG. 3 is a flow chart illustrating the speech data recognition steps of the intelligent quality inspection method for customer service call data of FIG. 1;

FIG. 4 is a flowchart of a text message obtaining step in the intelligent quality inspection method for customer service call data of FIG. 1;

FIG. 5 is a flowchart of a first quality inspection result obtaining step in the intelligent quality inspection method for customer service call data of FIG. 1;

FIG. 6 is a flow chart of a secondary quality inspection step in the intelligent quality inspection method for customer service call data of FIG. 1;

FIG. 7 is a flowchart of a first modified text information obtaining step in a secondary quality inspection step of the intelligent quality inspection method for customer service call data of FIG. 6;

FIG. 8 is a flowchart illustrating a second quality inspection result obtaining step in the secondary quality inspection step of the intelligent quality inspection method for customer service call data of FIG. 6;

fig. 9 is a schematic block structure diagram of an intelligent quality inspection apparatus for customer service call data according to an embodiment of the present disclosure;

fig. 10 is a schematic block diagram of a structure of a computer device according to an embodiment of the present application.

Detailed Description

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some, but not all, embodiments of the present application. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments in the present application without making any creative effort belong to the protection scope of the present application.

The flow diagrams depicted in the figures are merely illustrative and do not necessarily include all of the elements and operations/steps, nor do they necessarily have to be performed in the order depicted. For example, some operations/steps may be decomposed, combined or partially combined, so that the actual execution order may be changed according to the actual situation.

It is to be understood that the terminology used in the description of the present application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used in the specification of the present application and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

It should also be understood that the term "and/or" as used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

The embodiment of the application provides a method, a device, equipment and a medium for intelligently detecting customer service call data. The intelligent quality inspection method for customer service call data can be applied to artificial intelligence equipment, wherein the artificial intelligence equipment can be electronic equipment such as a mobile phone, a computer, an intelligent robot, an independent server or a server cluster and the like, and is not limited herein.

In this embodiment, the intelligent quality inspection method for customer service call data is applied to an independent computer as an example, but the method is not limited to the application to an independent computer.

In the following, some embodiments of the present application will be described in detail with reference to the drawings, and features in the following examples and examples may be combined with each other without conflict.

Referring to fig. 1, fig. 1 is a schematic flowchart illustrating steps of an intelligent quality inspection method for customer service call data according to an embodiment of the present application, where the method specifically includes the following steps S1 to S6.

Step S1, when a quality inspection instruction is received, customer service call data to be subjected to quality inspection evaluation are obtained according to the quality inspection instruction.

When the service quality of the customer service staff needs to be checked, the quality check staff sends a quality check instruction to a computer executing the method through corresponding electronic equipment, the computer acquires corresponding customer service call data from a database according to the quality check instruction, and the customer call data comprises a plurality of first voice data corresponding to consultants and a plurality of second voice data corresponding to the customer service staff.

As shown in fig. 2, in some embodiments, the quality inspection command includes information about the service person and time information corresponding to the service call. In the step S1, the step of obtaining the customer service call data to be subjected to quality inspection evaluation according to the quality inspection instruction specifically includes steps S11 to S12.

Step S11: and determining the data storage address of the customer service call data corresponding to the customer service staff according to the customer service staff information.

Specifically, the quality inspection instruction comprises customer service staff information and time information corresponding to customer service calls, and the customer service call data are classified into a plurality of call data sets in advance according to the customer service staff information and stored in a database, wherein the call data sets correspond to the customer service staff one to one, and the call data sets store the customer service call data corresponding to the customer service staff at different times. And the computer also establishes a matching model of the customer service personnel information and the data storage address of the customer service call data in advance through machine learning. When a quality inspection instruction is received, determining a data storage address of customer service call data corresponding to a customer service person according to customer service person information contained in the quality inspection instruction, so that the customer service call data can be extracted in the subsequent steps according to the data storage address and time information corresponding to the customer service call.

Step S12: and sending the call data request to a corresponding database according to the data storage address and the time information corresponding to the customer service call so as to acquire customer service call data to be subjected to quality inspection evaluation with corresponding customer service personnel.

In some embodiments, the customer service call data is classified into a plurality of call data sets in advance according to the customer service person information, and is stored in a database, wherein the call data sets store the customer service call data corresponding to the customer service persons at different times. And the computer generates a corresponding call data request according to the data storage address and the time information corresponding to the customer service call, and sends the call data request to a corresponding database so as to acquire customer service call data to be subjected to quality inspection evaluation with corresponding customer service personnel.

Illustratively, when quality inspection is performed on customer service call data of customer service staff A in 6.1.6.2020, customer service staff information A and time information corresponding to the customer service call are input into a computer, wherein the computer pre-defines storage addresses of the customer service staff information A and the corresponding customer service call data, the customer service staff information A comprises at least one of a contact telephone, an identity card number and a job number, the time information corresponding to the customer service call can be a specific time point or a time period of the customer service call, the computer determines the storage address of the customer service call data of the customer service staff A according to the identity information of the customer service staff and the time information corresponding to the customer service call, generates a corresponding customer service call data request and sends the corresponding customer service call data request to a database so that the database calls the corresponding customer service call data, and the computer receives the customer service call data of the customer service staff A sent by the database.

Step S2: and identifying a plurality of first voice data corresponding to consultants in the customer service call data and a plurality of second voice data corresponding to customer service personnel according to a time axis corresponding to the customer service call data, and giving voice time stamps corresponding to the first voice data and the second voice data.

According to the time axis of the call time, first voice data corresponding to a plurality of time points or time periods of consultants in the customer service call data and second voice data corresponding to a plurality of time points of the customer service personnel are identified, and time stamps corresponding to the first voice data and the second voice data are given according to the time corresponding to the voice data so as to identify the time sequence corresponding to the corresponding voice data.

As shown in fig. 3, in some embodiments, step S2 specifically includes steps S21-S24.

Step S21: and extracting voiceprint characteristic data in the customer service call data and time information corresponding to the voiceprint characteristic data according to a time axis corresponding to the customer service call data.

It can be understood that, if the voiceprint characteristics corresponding to each person's voice are different, a voiceprint characteristic model can be established by using machine learning. Specifically, the feature data that can be used by the voiceprint feature model include: acoustic feature data; lexical characteristic data; prosodic feature data.

Specifically, the customer service call data corresponds to a time axis, corresponding voiceprint feature data in the customer service call data are extracted according to a preset voiceprint feature model and time information corresponding to corresponding voiceprint features is marked according to the sequence of call time on the time axis. Wherein the voiceprint feature data includes at least one of a pitch spectrum and its contour, an energy of a pitch frame, an occurrence frequency and its trajectory of a pitch formant, a linear prediction cepstrum, a line spectrum pair, an autocorrelation and a log-area ratio, and a perceptual linear prediction.

In some embodiments, the voiceprint feature model is obtained by training with voice data input of the customer service person into a preset neural network model. The preset voiceprint feature model can be obtained by training the voiceprint feature data of the customer service staff, so that the extraction output result output by the voiceprint feature model is more accurate. In practical applications, when voiceprint feature extraction is performed, voiceprint feature data in the recorded customer service call data may be extracted, where the voiceprint feature data includes at least one of a pitch spectrum and its contour, energy of a pitch frame, occurrence Frequency and its trajectory of a pitch formant, a linear prediction Cepstrum, a line spectrum pair, an autocorrelation and log area ratio, mel Frequency Cepstrum Coefficient (MFCC), and perceptual linear prediction.

And screening and classifying the acquired voiceprint features by using a preset voiceprint model so as to divide the voiceprint features into a plurality of first voiceprint feature data corresponding to consultants and a plurality of second voiceprint feature data corresponding to customer service personnel, wherein the preset voiceprint model can be acquired by training by using the voiceprint feature data of the customer service personnel.

Step S22: and classifying the voiceprint characteristic data according to a preset voiceprint characteristic model so as to obtain a plurality of first voiceprint characteristic data corresponding to consultants and a plurality of second voiceprint characteristic data corresponding to customer service staff.

In some embodiments, a preset voiceprint feature model is used to perform screening classification on the obtained voiceprint features, so that the voiceprint features are divided into a plurality of first voiceprint feature data corresponding to counselors and a plurality of second voiceprint feature data corresponding to customer service staff, wherein the voiceprint feature model is obtained by inputting voice data of the customer service staff into a preset neural network model for training.

Step S23: and acquiring a plurality of corresponding first voice data according to the first voiceprint characteristic data, and acquiring a plurality of corresponding second voice data according to the second voiceprint characteristic data.

In some embodiments, based on the preset correspondence, a plurality of corresponding first voice data may be obtained according to the first voiceprint feature data, and a plurality of corresponding second voice data may be obtained according to the second voiceprint feature data, so as to associate the time point on the time axis with the voice object.

Exemplarily, correspondence among voice data corresponding to the customer service person a, the time axis of the customer service call, the voiceprint feature, and the voice object on 1 am 6/2020 is shown in the following table.

Based on the correspondence shown in the above table, a specific speech object corresponding to a specific time point is determined.

Step S24: and giving voice time stamps corresponding to the first voice data and the second voice data according to the time information corresponding to the voiceprint feature data.

After the voiceprint features are classified, a plurality of corresponding first voice data are obtained according to the first voiceprint feature data, a plurality of corresponding second voice data are obtained according to the second voiceprint feature data, and then voice timestamps corresponding to the first voice data and the second voice data are given according to time information corresponding to the voiceprint feature data, so that the sequence of corresponding voice data can be obtained according to the voice timestamps.

And step S3: inputting a plurality of first voice data and a plurality of second voice data into a preset voice conversion model for text conversion so as to obtain a plurality of first text messages corresponding to the plurality of first voice data and a plurality of second text messages corresponding to the plurality of second voice messages, wherein the first text messages and the second text messages are marked with text timestamps corresponding to the voice timestamps.

As shown in fig. 4, in some embodiments, in step S3, a plurality of first voice data and a plurality of second voice data are input into a preset voice conversion model for text conversion, so as to obtain a plurality of first text messages corresponding to the plurality of first voice data and a plurality of second text messages corresponding to the plurality of second voice messages, which specifically includes steps S31-S33.

Step S31: inputting a plurality of first voice data into a preset voice conversion model for text conversion to obtain a plurality of first text messages;

step S32: inputting a plurality of second voice data into a preset voice conversion model for text conversion so as to obtain a plurality of second text messages;

step S33: and marking a text time stamp corresponding to the first text information according to the voice time stamp corresponding to the first voice data, and marking a text time stamp corresponding to the second text information according to the voice time stamp corresponding to the second voice data.

In some embodiments, the preset Speech conversion model may be an Automatic Speech Recognition model (ASR), the first Speech data and the second Speech data are input into the ASR, and the Speech to be detected is converted into a corresponding text through the ASR, so as to obtain a plurality of first text information corresponding to the plurality of first Speech data and a plurality of second text information corresponding to the plurality of second Speech information. And marking corresponding text time stamps for the first text information and the second text information according to the voice time stamps corresponding to the first voice data and the second voice data.

Illustratively, the following table is a correspondence relationship between a time axis on which a customer service person talks, voice data, a voice time stamp, text information, and a text time stamp on 1 am on 6/month in 2020.

Time axis	Voice data	Voice time stamp	Text information	Text time stamp
					10：20	First voice data	Y1020	First text data	W1020
10：22	Second voice data	Y1022	Second text data	W1022
					10：24	First voice data	Y1024	First text data	W1024
10：26	First voice data	Y1026	First text data	W1026
					10：28	Second voice data	Y1028	Second text data	W1028
10：30	Second voice data	Y1030	Second text data	W1030
					……	……	……	……	……

According to the corresponding relation in the table, a plurality of text timestamps corresponding to each time point on the time axis of the customer service call data can be determined, and the time sequence of the timestamps on the time axis is obtained as follows: "W1020-W1022-W1024-W1026-W1024-W1026".

And step S4: and sequencing the plurality of first text messages and the plurality of second text messages according to the text time stamps so as to obtain the call record text corresponding to the customer service call data.

And sequencing the plurality of first text messages and the plurality of second text messages according to the text time stamps corresponding to the obtained text messages so as to obtain the customer service personnel and the call record texts corresponding to the customer service call data, wherein the call record texts are recorded with the call personnel information, the call time information and the call text content information.

Illustratively, the plurality of text timestamps are sorted according to a chronological order, that is, the order of the text timestamps corresponding to the customer service call data is obtained: and W1020-W1022-W1024-W1026-W1024-W1026", splicing the plurality of first text data and the plurality of second text data according to the sequence to obtain a call record text corresponding to the customer service call data.

Step S5: acquiring a first quality inspection result of the customer service call data according to the call record text;

as shown in fig. 5, in some embodiments, step S5 specifically includes steps S51-S52.

Step S51: performing keyword splitting on the first text information and the second text information to obtain a first keyword corresponding to the first text information and a second keyword corresponding to the second text information;

step S52: and judging the accuracy of answering questions by the customer service personnel according to the first keyword and the second keyword so as to generate a first quality inspection result of the customer service call data.

The computer stores a first corresponding relation between the similarity of the first keyword and the second keyword and the first evaluation information, obtains the similarity of the first keyword and the second keyword through machine learning, and obtains the first evaluation information through the first corresponding relation and by using the obtained similarity of the first keyword and the second keyword.

And a second corresponding relation of the similarity between the second keyword and the preset keyword is stored in the computer, the similarity between the second keyword and the preset keyword is obtained through machine learning, and second evaluation information is obtained through the second corresponding relation and the obtained similarity between the second keyword and the preset keyword.

Specifically, a computer executing the method performs keyword splitting on first text information and second text information to obtain a first keyword corresponding to the first text information and a second keyword corresponding to the second text information; acquiring the similarity of a first keyword and a second keyword, and acquiring first evaluation information by using the acquired similarity of the first keyword and the second keyword; and acquiring the similarity of the second keyword and the preset keyword, and acquiring second evaluation information by using the acquired similarity of the second keyword and the preset keyword. And judging the accuracy of the customer service personnel for answering the questions according to the first keyword and the second keyword so as to generate a first quality inspection result of the customer service call data.

And generating a quality evaluation result corresponding to the customer service call data through the first evaluation information and the second evaluation information, and taking the quality evaluation result as a first quality inspection result of the customer service call data, wherein the first quality inspection result is used for representing the accuracy of answering questions by customer service staff.

It can be understood that the first evaluation information is determined based on the similarity between the first keyword and the second keyword, and it is characterized that whether the voice service provided by the customer service staff is consistent with the consultation of the consultant or not, and when the similarity between the first text information corresponding to the consultant and the first text information corresponding to the customer service staff is high, the voice service provided by the customer service staff is considered to be consistent with the consultation of the consultant, and the quality of the voice service is better.

Further, the second evaluation information is determined based on the similarity between the second keyword and a preset keyword, and the representation is whether the voice service provided by the customer service staff includes a dedication, the preset keyword may include common dedications such as "you", "please", "thank you", and "bad breath", and when the similarity between the second text information corresponding to the customer service staff and the preset keyword is high, the voice service provided by the customer service staff can be considered to include the dedication, and the quality of the voice service is good.

Step S6: and receiving a secondary quality inspection request of the customer service responding to the first quality inspection result, and rechecking the customer service call data according to the secondary quality inspection request to obtain a second quality inspection result.

As shown in fig. 6, in some embodiments, in step S6, rechecking the customer service call data according to the secondary quality inspection request to obtain a second quality inspection result includes steps S61 to S63:

step S61: and judging whether the translation of the first text information corresponding to the call record text is wrong or not according to the secondary quality inspection request.

It can be understood that in the voice translation process, the voice of the consultant is often different due to the complexity of the regional distribution of the consultant, so that the accuracy of translation is insufficient when a voice translation model is used for translating some special pronunciations in some contexts. For example, the "sh" flat tongue and "s" of the curly tongue sound, "ping" of the front nasal sound and "pin" of the rear nasal sound, the letters "j" and "z", the letters "L" and "N", and the letters "F" and "H". Therefore, words and short sentences with high error rate when the voice is converted into the text are taken as preset words and sentences, and the wrong words database is collected and formed.

And when the customer service staff disagrees with the first quality inspection result and provides a secondary quality inspection request, the computer responds to the secondary quality inspection request and judges whether the translation of the first text information is wrong or not according to the wrong word database. Specifically, the computer traverses first text information in the call record text, identifies whether the first text information contains preset words and sentences stored in the wrong word database, and judges that translation of the corresponding first text information in the call record text is wrong when the first text information contains the preset words and sentences.

Illustratively, in text conversion under certain contexts, the word "milk" is easily converted into "Liu Lai", and thus "Liu Lai" is stored as a preset word in the wrong-word database.

When the counselor wants to express "milk" but inputs the corresponding first voice data into the preset voice conversion model, the first text information output by the model contains "Liu Lai". The accuracy of answering questions by customer service staff is judged by using a call record text containing 'Liu Lai', so that a first quality inspection result is generated, the accuracy obtained by judgment is reduced, the first quality inspection result is not accurate enough, the computer responds to a secondary quality inspection request, traverses first text information in the call record text, identifies whether the first text information contains preset words and phrases stored in an incorrect word database, and judges that translation of the first text information is wrong when the first text information contains 'Liu Lai' stored in the incorrect word database.

Step S62: when the corresponding translation of the first text information is wrong, the first text information with wrong translation is marked, and the first text information with wrong translation is corrected according to the call record text to obtain first corrected text information.

In some embodiments, step S62 specifically includes: steps S621-S622.

Step S621: marking a first phrase corresponding to the translation error in the first text information with the wrong translation;

step S622: and extracting second text information adjacent to the first text information with wrong translation according to the text timestamp, and correcting the first phrase with wrong translation by using the extracted second text information to obtain first corrected text information.

Illustratively, the phrase with wrong translation in the first text information is analyzed by using the comparison result, and the phrase with wrong translation is marked. And then extracting second text information adjacent to the first text information with wrong translation and first text information adjacent to the first text information with wrong translation from the call record text according to the text timestamp corresponding to the first text information with wrong translation.

And splitting the keywords of the extracted first text information and second text information, acquiring a second phrase corresponding to the extracted second text information and a third phrase corresponding to the first text information, acquiring a second phrase corresponding to the translated wrong phrase according to a preset corresponding relation between the first phrase and the second phrase and the third phrase, and correcting the first phrase by using the second phrase to acquire first corrected text information.

The method comprises the steps that a plurality of first phrases with wrong translation and second phrases with correct corresponding translation are stored in a computer, when the first phrases are marked as the first phrases in first text information and the second phrases are identified in second text information, the first phrases are corrected by using the second phrases, and therefore first corrected text information is obtained.

For example, the call record text after the speech translation is as follows:

w1020 consultant: your activity of 6 months and 17 days is clear to buy all 688 Yuan's say sending a box of ' Liu Lai ', how now there is nothing!

W1022 customer service personnel: do you say that is we sent a box of "milk" for the activity of day 17/6?

W1024 consultant: you can write clearly, send after buying 688, I is good and not easy to gather enough!

The W1026 consultant: for, a box "Liu Lai"!

W1028 customer service staff: good, do you check you here, then reply you can?

W1030 customer service personnel: thank you for the incoming call, here xxx serves you and asks you to answer.

In the above record, in the W1020 sentence corresponding to the text time axis, the voice translation model may have a wrong translation of the word "milk" due to the accent problem of the counselor, and when it is determined that the translation of the word is wrong, the word with the wrong translation is labeled and recorded.

In some embodiments, step S622 specifically includes: the method comprises the steps of firstly determining the relative position of first text information with errors in translation in a time axis, then determining the relative position of second text information adjacent to the first text information with errors in translation in the time axis according to a text timestamp, splitting keywords of the obtained second text information to extract the second text information, and correcting the first phrase with errors in translation by using the extracted second text information to obtain first corrected text information.

For example, when it is determined that the word "Liu Lai" in the first text information is incorrect, second text information adjacent to the first text information corresponding to the first text information containing "Liu Lai" is extracted from the call record text according to the text timestamp, and keyword splitting is performed on the acquired second text information. When the first phrase is extracted to be 'Liu Lai' and the second phrase is extracted to be 'milk', the 'milk' of the second phrase is substituted for the 'Liu Lai' extracted from the first phrase to obtain first corrected text information.

Step S63: and acquiring a second quality inspection result of the customer service call data according to the first corrected text message and the second text message.

As shown in fig. 8, in some embodiments, step S63 specifically includes steps S631-S632.

Step S631: performing keyword splitting on the first corrected text information and the second text information to obtain a first keyword corresponding to the first corrected text information and a second keyword corresponding to the second text information;

step S632: and judging the accuracy of answering questions by the customer service personnel according to the first keyword and the second keyword so as to generate a second quality inspection result of the customer service call data.

It can be understood that the computer stores a first corrected corresponding relationship between the similarity of the first corrected keyword and the second keyword and the first corrected evaluation information, obtains the similarity of the first corrected keyword and the second keyword through machine learning, and obtains the first corrected evaluation information through the first corrected corresponding relationship and using the obtained similarity of the first corrected keyword and the second keyword. Meanwhile, a second corresponding relation of the similarity between the second keyword and the preset keyword is stored in the computer, the similarity between the second keyword and the preset keyword is obtained through machine learning, and second evaluation information is obtained through the second corresponding relation and the obtained similarity between the second keyword and the preset keyword.

Specifically, a computer executing the method performs keyword splitting on first corrected text information and second text information to obtain a first corrected keyword corresponding to the first corrected text information and a second keyword corresponding to the second text information; acquiring the similarity of a first corrected keyword and a second keyword, and acquiring first corrected evaluation information by using the acquired similarity of the first corrected keyword and the second keyword; and acquiring the similarity of the second keyword and the preset keyword, and acquiring second evaluation information by using the acquired similarity of the second keyword and the preset keyword. And judging the accuracy of the customer service personnel for answering the questions according to the first correction keyword and the second keyword so as to generate a first correction quality inspection result of the customer service call data.

And generating a quality evaluation result corresponding to the customer service call data through the first correction evaluation information and the second evaluation information, taking the quality evaluation result as a second quality inspection result of the customer service call data, wherein the second quality inspection result is used for representing the accuracy of answering questions by customer service staff, and after text correction, the accuracy of the second quality inspection result is higher than that of the first quality inspection result.

In other embodiments, step S6 specifically includes: and receiving a secondary quality inspection request of the customer service responding to the first quality inspection result, outputting the first voice data and the second voice data to a quality inspection end according to the secondary quality inspection request, and receiving an artificial quality inspection result generated by the quality inspection end as a second quality inspection result.

Specifically, a computer executing the intelligent quality inspection method for customer service call data generates a first quality inspection result of the customer service call data through first evaluation information and second evaluation information, outputs the first quality inspection result, when a customer service person disagrees with the first quality inspection result and inputs a secondary quality inspection request to the computer, the computer sends the first voice data and the second voice data to corresponding quality inspection terminals according to the secondary quality inspection request, so that the quality inspection terminals perform manual quality inspection on the first voice data and the second voice data to generate manual quality inspection results, and then the computer receives the manual quality inspection results of the quality inspection terminals as second quality inspection results.

The scheme is used for evaluating the voice service of the customer service staff so as to improve the quality of customer service. In the scheme, when a computer executing the intelligent quality inspection method of the customer service call data receives a quality inspection instruction, the customer service call data to be subjected to quality inspection evaluation is obtained according to the quality inspection instruction; identifying a plurality of first voice data corresponding to consultants in the customer service call data and a plurality of second voice data corresponding to the customer service personnel according to a time axis corresponding to the customer service call data, wherein the first voice data and the second voice data are marked with voice time stamps corresponding to the time axis; inputting a plurality of first voice data and a plurality of second voice data into a preset voice conversion model for text conversion so as to obtain a plurality of first text messages corresponding to the plurality of first voice data and a plurality of second text messages corresponding to the plurality of second voice messages, wherein the first text messages and the second text messages are marked with text timestamps corresponding to the voice timestamps; sequencing the plurality of first text messages and the plurality of second text messages according to the text time stamps to obtain call record texts corresponding to the customer service call data; according to the call record text, obtaining a first quality inspection result of the customer service call data; and receiving a secondary quality inspection request of the customer service responding to the first quality inspection result, and rechecking the customer service call data according to the secondary quality inspection request to obtain a second quality inspection result. Therefore, quality inspection misevaluations caused by text conversion errors of text information can be effectively reduced, the accuracy of voice service quality inspection is improved, and the work evaluation of customer service personnel is objectively and truly reflected.

Fig. 9 is a schematic block structure diagram of an intelligent customer service call data quality inspection apparatus according to an embodiment of the present disclosure, and as shown in fig. 9, the intelligent customer service call data quality inspection apparatus 100 includes:

the data acquisition module 101 is used for acquiring customer service call data to be subjected to quality inspection evaluation according to a quality inspection instruction when the quality inspection instruction is received;

the identification module 102 is configured to identify, according to a time axis corresponding to the customer service call data, a plurality of first voice data corresponding to a counselor in the customer service call data and a plurality of second voice data corresponding to the customer service person, where the first voice data and the second voice data are both marked with a voice timestamp corresponding to the time axis;

the text conversion module 103 is configured to input the plurality of first voice data and the plurality of second voice data into a preset voice conversion model for text conversion, so as to obtain a plurality of first text messages corresponding to the plurality of first voice data and a plurality of second text messages corresponding to the plurality of second voice messages, where the first text messages and the second text messages are both marked with text timestamps corresponding to the voice timestamps;

the sorting module 104 is configured to sort the plurality of first text messages and the plurality of second text messages according to the text timestamps to obtain call record texts corresponding to the customer service call data;

the first quality inspection module 105 is used for acquiring a first quality inspection result of the customer service call data according to the call record text;

and the second quality inspection module 106 is configured to perform recheck on the customer service call data according to the secondary quality inspection request to obtain a second quality inspection result.

In an embodiment, the quality inspection instruction includes customer service staff information and time information corresponding to a customer service call, and the data obtaining module 101 obtains the customer service call data to be subjected to quality inspection evaluation according to the quality inspection instruction, specifically including:

determining a data storage address of customer service call data corresponding to a customer service staff according to the customer service staff information;

and sending a call data request to a corresponding database according to the data storage address and the time information corresponding to the customer service call so as to acquire customer service call data to be subjected to quality inspection evaluation with corresponding customer service personnel.

In an embodiment, the identifying module 102 identifies, according to a time axis corresponding to the customer service call data, a plurality of first voice data corresponding to a counselor in the customer service call data and a plurality of second voice data corresponding to the customer service person, and assigns voice timestamps corresponding to the first voice data and the second voice data, specifically including:

extracting voiceprint characteristic data in the customer service call data and time information corresponding to the voiceprint characteristic data according to a time axis corresponding to the customer service call data;

classifying the voiceprint feature data according to a preset voiceprint feature model to obtain a plurality of first voiceprint feature data corresponding to consultants and a plurality of second voiceprint feature data corresponding to customer service staff;

acquiring a plurality of corresponding first voice data according to the first voiceprint characteristic data, and acquiring a plurality of corresponding second voice data according to the second voiceprint characteristic data;

and giving voice time stamps corresponding to the first voice data and the second voice data according to the time information corresponding to the voiceprint characteristic data.

In an embodiment, the text conversion module 103 inputs a plurality of first voice data and a plurality of second voice data into a preset voice conversion model for text conversion to obtain a plurality of first text messages corresponding to the plurality of first voice data and a plurality of second text messages corresponding to the plurality of second voice messages, where the first text messages and the second text messages are both marked with text timestamps corresponding to the voice timestamps, and the text conversion method specifically includes:

inputting a plurality of first voice data into a preset voice conversion model for text conversion so as to obtain a plurality of first text messages;

inputting a plurality of second voice data into a preset voice conversion model for text conversion so as to obtain a plurality of second text messages;

and marking a text time stamp corresponding to the first text information according to the voice time stamp corresponding to the first voice data, and marking a text time stamp corresponding to the second text information according to the voice time stamp corresponding to the second voice data.

In an embodiment, the obtaining, by the first quality inspection module 105, a first quality inspection result of the customer service call data according to the call record text specifically includes:

performing keyword splitting on the first text information and the second text information to obtain a first keyword corresponding to the first text information and a second keyword corresponding to the second text information;

and judging the accuracy of the customer service personnel for answering the questions according to the first keyword and the second keyword so as to generate a first quality inspection result of the customer service call data.

In an embodiment, the second quality inspection module 106 performs recheck on the customer service call data according to the second quality inspection request to obtain a second quality inspection result, which specifically includes:

judging whether the translation of the corresponding first text information in the call record text is wrong or not according to the secondary quality inspection request;

when the translation of the corresponding first text information is wrong, marking the first text information with wrong translation, and correcting the first text information with wrong translation according to the call record text to obtain first corrected text information;

and acquiring a second quality inspection result of the customer service call data according to the first corrected text information and the second text information.

In one embodiment, the marking of the first text information with the translation error by the second quality inspection module 106 and the correcting of the first text information with the translation error according to the call record text to obtain the first corrected text information specifically include:

marking a first phrase corresponding to the translation error in the first text information with the translation error;

and extracting second text information and first text information adjacent to the first text information with wrong translation according to the text timestamp, and correcting the first phrase with wrong translation by using the extracted second text information and the first text information to obtain first corrected text information.

Referring to fig. 10, fig. 10 is a schematic block diagram illustrating a structure of a computer device according to an embodiment of the present disclosure.

As shown in fig. 10, the computer device 200 includes a processor 201 and a memory 202, and the processor 201 and the memory 202 are connected by a bus 203 such as an I2C (Inter-integrated Circuit) bus.

In particular, the processor 201 is used to provide computing and control capabilities, supporting the operation of the entire computer device. The Processor 201 may be a Central Processing Unit (CPU), and the Processor 201 may also be other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic device, discrete hardware component, etc. Wherein a general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

Specifically, the Memory 202 may be a Flash chip, a Read-Only Memory (ROM) magnetic disk, an optical disk, a usb disk, or a removable hard disk.

Those skilled in the art will appreciate that the configuration shown in fig. 10 is a block diagram of only a portion of the configuration associated with the embodiments of the present application, and does not constitute a limitation on the applicability of the embodiments of the present application to computing devices that may include more or less components than those shown, or combine certain components, or have a different arrangement of components.

The processor is configured to run a computer program stored in the memory, and when executing the computer program, implement any one of the intelligent quality inspection methods for customer service call data provided in the embodiments of the present application.

In one embodiment, the processor 201 is configured to run a computer program stored in the memory 202 and when executing the computer program to perform the steps of:

and receiving a secondary quality inspection request of the customer service responding to the first quality inspection result, and rechecking the customer service call data according to the secondary quality inspection request to obtain a second quality inspection result.

In one embodiment, the quality inspection instruction comprises customer service personnel information and time information corresponding to customer service calls; when obtaining the customer service call data to be subjected to quality inspection evaluation according to the quality inspection instruction, the processor 201 includes:

determining a data storage address of customer service call data corresponding to the customer service staff according to the customer service staff information;

In one embodiment, the processor 201, in recognizing a plurality of first voice data corresponding to a counselor and a plurality of second voice data corresponding to a servicer in the service call data according to a time axis corresponding to the service call data, and assigning a voice time stamp corresponding to the first voice data and the second voice data, includes:

In one embodiment, the method for performing text conversion by the processor 201 after inputting a plurality of first voice data and a plurality of second voice data into a preset voice conversion model to obtain a plurality of first text messages corresponding to the plurality of first voice data and a plurality of second text messages corresponding to the plurality of second voice messages, where each of the first text messages and the second text messages is marked with a text timestamp corresponding to the voice timestamp includes:

In one embodiment, when obtaining the first quality inspection result of the customer service call data according to the call record text, the processor 201 includes:

In one embodiment, when the processor 201 rechecks the customer service call data according to the second quality inspection request to obtain the second quality inspection result, the method includes:

when the translation of the corresponding first text information is wrong, marking the first text information with the wrong translation, and correcting the first text information with the wrong translation according to the call record text to obtain first corrected text information;

In one embodiment, when marking the first text message with an error in translation and correcting the first text message with an error in translation according to the call record text to obtain the first corrected text message, the processor 201 includes:

and extracting second text information and first text information adjacent to the first text information with wrong translation according to the text timestamp, and correcting a first phrase corresponding to the wrong translation by using the extracted second text information and the first text information to obtain first corrected text information.

It should be noted that, as will be clear to those skilled in the art, for convenience and brevity of description, the specific working process of the computer device described above may refer to the corresponding process in the foregoing embodiment of the animal identification method, and is not described herein again.

The embodiment of the present application further provides a storage medium, where a computer program is stored, where the computer program can be executed by one or more processors to implement the steps of any one of the customer service call data intelligent quality inspection methods provided in the description of the embodiment of the present application.

The storage medium may be an internal storage unit of the computer device described in the foregoing embodiment, for example, a hard disk or a memory of the computer device. The storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like, provided on the computer device.

It will be understood by those of ordinary skill in the art that all or some of the steps of the methods, systems, functional modules/units in the devices disclosed above may be implemented as software, firmware, hardware, and suitable combinations thereof. In a hardware embodiment, the division between functional modules/units mentioned in the above description does not necessarily correspond to the division of physical components; for example, one physical component may have multiple functions, or one function or step may be performed by several physical components in cooperation. Some or all of the physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application specific integrated circuit. Such software may be distributed on computer readable media, which may include computer storage media (or non-transitory media) and communication media (or transitory media). The term computer storage media includes volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data, as is well known to those of ordinary skill in the art. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital Versatile Disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store the desired information and which can accessed by a computer. In addition, communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media as known to those skilled in the art.

It should be understood that the term "and/or" as used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items. It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or system. Without further limitation, an element defined by the phrase "comprising a … …" does not exclude the presence of another identical element in a process, method, article, or system that comprises the element.

The above-mentioned serial numbers of the embodiments of the present application are merely for description, and do not represent the advantages and disadvantages of the embodiments. The above description is only for the specific embodiment of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily conceive various equivalent modifications or substitutions within the technical scope of the present application, and these modifications or substitutions should be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. An intelligent quality inspection method for customer service call data is characterized by comprising the following steps:

when a quality inspection instruction is received, obtaining customer service call data to be subjected to quality inspection evaluation according to the quality inspection instruction;

identifying a plurality of first voice data corresponding to consultants in the customer service call data and a plurality of second voice data corresponding to customer service personnel according to a time axis corresponding to the customer service call data, wherein the first voice data and the second voice data are marked with voice time stamps corresponding to the time axis;

inputting a plurality of first voice data and a plurality of second voice data into a preset voice conversion model for text conversion so as to obtain a plurality of first text messages corresponding to the plurality of first voice data and a plurality of second text messages corresponding to the plurality of second voice data, wherein the first text messages and the second text messages are marked with text timestamps corresponding to the voice timestamps;

sequencing the first text messages and the second text messages according to the text timestamps to obtain call record texts corresponding to the customer service call data;

receiving a secondary quality inspection request of a customer service responding to the first quality inspection result, and rechecking the customer service call data according to the secondary quality inspection request to obtain a second quality inspection result;

the rechecking the customer service call data according to the secondary quality inspection request to obtain a second quality inspection result includes:

judging whether the translation of the first text information corresponding to the call record text is wrong or not according to the secondary quality inspection request;

when the corresponding translation of the first text information is wrong, marking the first text information with wrong translation, and correcting the first text information with wrong translation according to the call record text to obtain first corrected text information;

2. The method of claim 1, wherein the quality inspection instructions include customer service personnel information and time information corresponding to a customer service call;

the step of obtaining the customer service call data to be subjected to quality inspection evaluation according to the quality inspection instruction comprises the following steps:

and sending the call data request to a corresponding database according to the data storage address and the time information corresponding to the customer service call so as to acquire customer service call data to be subjected to quality inspection evaluation by corresponding customer service staff.

3. The method of claim 1, wherein the recognizing a plurality of first voice data corresponding to counselors and a plurality of second voice data corresponding to service staff in the service call data according to a time axis corresponding to the service call data and assigning voice time stamps corresponding to the first voice data and the second voice data comprises:

extracting voiceprint feature data in the customer service call data and time information corresponding to the voiceprint feature data according to a time axis corresponding to the customer service call data;

and giving voice time stamps corresponding to the first voice data and the second voice data according to the time information corresponding to the voiceprint feature data.

4. The method according to claim 1, wherein the inputting a plurality of the first voice data and a plurality of the second voice data into a preset voice conversion model for text conversion to obtain a plurality of first text messages corresponding to a plurality of the first voice data and a plurality of second text messages corresponding to a plurality of the second voice data, the first text messages and the second text messages being labeled with text time stamps corresponding to the voice time stamps, comprises:

inputting a plurality of first voice data into a preset voice conversion model for text conversion to obtain a plurality of first text messages;

inputting a plurality of second voice data into a preset voice conversion model for text conversion to obtain a plurality of second text messages;

5. The method of claim 1, wherein obtaining a first quality inspection result of the customer service call data according to the call record text comprises:

and judging the accuracy of answering questions by the customer service personnel according to the first keyword and the second keyword so as to generate a first quality inspection result of the customer service call data.

6. The method of claim 1, wherein the marking the first text message with the wrong translation and correcting the first text message with the wrong translation according to the call record text to obtain a first corrected text message comprises:

marking a first phrase corresponding to the translation error in the first text information with the wrong translation;

and extracting the second text information adjacent to the first text information with wrong translation and the first text information according to the text timestamp, and correcting the first phrase with wrong translation by using the extracted second text information and the first text information to obtain first corrected text information.

7. The utility model provides a customer service conversation data intelligence quality control device which characterized in that includes:

the data acquisition module is used for acquiring customer service call data to be subjected to quality inspection evaluation according to a quality inspection instruction when the quality inspection instruction is received;

the text conversion module is used for inputting the plurality of first voice data and the plurality of second voice data into a preset voice conversion model for text conversion so as to obtain a plurality of first text messages corresponding to the plurality of first voice data and a plurality of second text messages corresponding to the plurality of second voice data, wherein the first text messages and the second text messages are marked with text timestamps corresponding to the voice timestamps;

the second quality inspection module is used for receiving a secondary quality inspection request of the customer service responding to the first quality inspection result and rechecking the customer service call data according to the secondary quality inspection request to obtain a second quality inspection result;

the second quality inspection module rechecks the customer service call data according to the secondary quality inspection request to acquire a second quality inspection result, and the second quality inspection module comprises:

and acquiring a second quality inspection result of the customer service call data according to the first corrected text message and the second text message.

8. A computer device, wherein the computer device comprises a memory and a processor;

the memory for storing a computer program;

the processor is used for executing the computer program and realizing the intelligent quality inspection method of the customer service call data according to any one of claims 1 to 6 when the computer program is executed.

9. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program which, when executed by a processor, causes the processor to implement the intelligent quality inspection method for customer service call data according to any one of claims 1 to 6.