CN117057813A - Customer service assisting method and system - Google Patents
Customer service assisting method and system Download PDFInfo
- Publication number
- CN117057813A CN117057813A CN202311014713.9A CN202311014713A CN117057813A CN 117057813 A CN117057813 A CN 117057813A CN 202311014713 A CN202311014713 A CN 202311014713A CN 117057813 A CN117057813 A CN 117057813A
- Authority
- CN
- China
- Prior art keywords
- work order
- customer service
- content
- text
- intention
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000012549 training Methods 0.000 claims abstract description 36
- 238000000605 extraction Methods 0.000 claims abstract description 16
- 239000003795 chemical substances by application Substances 0.000 claims description 14
- 238000003058 natural language processing Methods 0.000 claims description 10
- 238000004140 cleaning Methods 0.000 claims description 7
- 230000011218 segmentation Effects 0.000 claims description 7
- 238000003032 molecular docking Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 abstract description 4
- 238000013518 transcription Methods 0.000 description 4
- 230000035897 transcription Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 230000008909 emotion recognition Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011010 flushing procedure Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000003362 replicative effect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/01—Customer relationship services
- G06Q30/015—Providing customer assistance, e.g. assisting a customer within a business location or via helpdesk
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/34—Browsing; Visualisation therefor
- G06F16/345—Summarisation for human users
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
- G06F16/353—Clustering; Classification into predefined classes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/10—Text processing
- G06F40/166—Editing, e.g. inserting or deleting
- G06F40/186—Templates
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Business, Economics & Management (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Finance (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Accounting & Taxation (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a customer service auxiliary method and a customer service auxiliary system, which belong to the technical field of electric digital data processing. The method for filling the worksheet comprises the following steps: acquiring a first training set of work order data; training the first training set based on a pointer generation network to obtain a work order content extraction model; acquiring a appeal text of a customer; extracting the appeal text through a work order content extraction model to obtain abstract content; acquiring a work order template; and filling in the dynamic content of the work order template according to the abstract content. The abstract of the network complaint text is generated based on the pointer, and the dynamic content of the work order is filled in, so that the content of the work order can be simplified, and the quick reading is facilitated. The customer service is assisted to rapidly and accurately answer; the working efficiency of the seat can be improved, and the labor cost is reduced.
Description
Technical Field
The invention relates to the technical field of electric digital data processing, in particular to a customer service auxiliary method and system.
Background
Customer service auxiliary systems are widely used in various fields, and after a customer service center establishes a connection with a customer, the customer needs to know the customer's appeal and answer and reply the customer's appeal.
The main processing modes of the current customer requirements are as follows:
(1) The customer service seat needs to manually select the reason of the incoming call under the condition of knowing the customer appeal, inquire the knowledge base and search the module for handling the related business for operation; the background worksheets are circulated, processed and fed back, and the workers need to manually process worksheets of different types, fill in worksheets to accept contents, identify a dispatching mode and the like, so that the operation is complicated, the working efficiency is low, a large amount of labor cost is consumed, the conversation duration is high, and the automatic knowledge accumulation and autonomous learning capability are not realized.
(2) After the voice is transcribed into the text, the text is written into the work order. However, the filling content is more, so that the work order has more redundant information and needs to be manually reprocessed.
Therefore, a customer assistance method is needed to be designed to realize the assistance inquiry of high-frequency service and functions, incoming call cause analysis, error-prone point real-time reminding and emotion recognition, assist customer service representatives to answer rapidly and accurately, and improve the working efficiency of telephone operators.
Disclosure of Invention
Aiming at the technical problems in the prior art, the invention provides a customer service auxiliary method and a customer service auxiliary system so as to improve the working efficiency of customer service.
The invention discloses a customer service auxiliary method, which comprises a method for filling a work order, wherein the method for filling the work order comprises the following steps: acquiring a first training set of work order data; training the first training set based on a pointer generation network to obtain a work order content extraction model; acquiring a appeal text of a customer; extracting the appeal text through a work order content extraction model to obtain abstract content; acquiring a work order template; and filling in the dynamic content of the work order template according to the abstract content.
Preferably, the specific method for filling the worksheet comprises the following steps:
after cleaning and word segmentation are carried out on the worksheet data, a data set is obtained;
dividing the data set to obtain a first training set;
training the first training set based on a pointer generation network to obtain a work order content extraction model;
after data cleaning and word segmentation are carried out on the appeal text, inputting a work order extraction model to obtain abstract content;
and filling in the dynamic content of the work order template according to the abstract content.
Preferably, the method for forbidden language identification comprises the following steps:
establishing forbidden regular configuration;
based on a natural language processing algorithm, a negative service intention recognition model is established;
and identifying the forbidden language of the appeal text/customer service seat reply text through the forbidden language regular configuration and the negative service intention identification model.
Preferably, the intention recognition method comprises the following steps:
obtaining a word vector model;
obtaining a first word vector of the work order data through a word vector model, and constructing a second training set according to the work order data and the first word vector;
training the second training set based on a two-way long-short-term memory network to obtain an intention recognition model;
after the complaint text or the abstract content is segmented, a second word vector is obtained through a word vector model;
and obtaining the corresponding intention of the second word vector through the intention recognition model.
Preferably, the method of knowledge point pushing,
establishing a first association of the intent with the knowledge point;
and pushing knowledge points corresponding to the intention to the customer service seat according to the first association.
Preferably, the method for reminding the error-prone point comprises the following steps:
establishing a second association of the intent with the error prone point;
and pushing the error-prone points corresponding to the intentions to customer service agents according to the second association.
Preferably, the method for reminding the phone is as follows:
obtaining context and service content according to the text or abstract content;
obtaining a corresponding conversation according to the context and the service content;
pushing the conversation for customer service agents.
Preferably, the appeal text is obtained by any of the following means:
transcribing the voice or conversation of the engine client into a appeal text through automatic voice recognition;
acquiring a client's appeal text through a short message, chat software, applet or webpage;
the method for converting the voice into the appeal text comprises any one of the following methods:
docking the voice gateway with a platform based on a media resource control protocol, wherein the platform transcribes voice into a appeal text through a voice recognition engine;
forwarding the voice stream of the seat client to a voice recognition engine through a voice proxy, and writing the voice stream into a appeal text through the voice recognition engine;
after the voice stream collected by the sound card collecting equipment and the microphone collecting equipment is divided by the audio stream collecting equipment, the voice stream is forwarded to the voice recognition engine, and the voice stream is transcribed into a appeal text by the voice recognition engine.
The invention also provides a customer service auxiliary system, which comprises a work order filling module,
the work order filling module is used for: extracting the appeal text through a work order content extraction model to obtain abstract content; and filling in the dynamic content of the work order template according to the abstract content.
The system also comprises an intention recognition module, a forbidden language recognition module, a knowledge point pushing module, an error-prone point reminding module and a speaking reminding module;
the intention recognition module is used for obtaining a first word vector of the work order data through the word vector model and constructing a second training set according to the work order data and the first word vector; training the second training set based on a two-way long-short-term memory network to obtain an intention recognition model; after the complaint text or the abstract content is segmented, a second word vector is obtained through a word vector model; obtaining corresponding intention of the second word vector through the consciousness recognition model;
the forbidden term identification module is used for establishing forbidden term regular configuration; based on a natural language processing algorithm, a negative service intention recognition model is established; identifying forbidden languages of the appeal text/customer service seat reply text through the forbidden language regular configuration and the negative service intention identification model;
the knowledge point pushing module is used for pushing the knowledge point corresponding to the intention to the customer service seat according to the first association of the intention and the knowledge point;
the error-prone point reminding module is used for pushing error-prone points corresponding to the intention to the customer service seat according to the second association of the intention and the error prone point;
the speaking reminding module is used for obtaining context and service content according to the text or abstract content; obtaining a corresponding conversation according to the context and the service content; pushing the conversation for customer service agents.
Compared with the prior art, the invention has the beneficial effects that: the abstract of the network complaint text is generated based on the pointer, and the dynamic content of the work order is filled in, so that the content of the work order can be simplified, and the quick reading is facilitated. The customer service is assisted to rapidly and accurately answer; the working efficiency of the seat can be improved, and the labor cost is reduced.
Drawings
FIG. 1 is a flow chart of a work order filling method of the present invention;
FIG. 2 is a flow chart of a method of intent recognition;
FIG. 3 is a logical block diagram of the customer service assistance system of the present invention;
FIG. 4 is a particular logical deployment diagram;
fig. 5 is a specific hardware deployment diagram.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The invention is described in further detail below with reference to the attached drawing figures:
the customer service auxiliary method comprises the steps of work order filling, intention recognition, forbidden language recognition, knowledge point pushing, error-prone point reminding and speaking reminding.
The method for filling the worksheet shown in fig. 1 comprises the following steps:
step 101: a first training set of work order data is obtained.
The method comprises the steps of performing data cleaning and word segmentation on worksheets to obtain a data set; the data set is partitioned into a first training set and a first test set.
Step 102: the first training set is trained based on a pointer generation network (Poi nter-Generator Networks) to obtain a work order content extraction model.
Step 103: and acquiring the appeal text of the client.
The appeal text can be subjected to data cleaning, word segmentation and preprocessing operations.
Step 104: and extracting the appeal text through a work order content extraction model to obtain abstract content.
Step 105: and obtaining a work order template. The work order template typically includes fixed information and dynamic information.
Step 106: and filling in the dynamic content of the work order template according to the abstract content.
The abstract of the text of the request is generated based on the pointer, and the dynamic content of the work order is filled in, so that the content of the work order can be simplified, and the quick reading is facilitated; the pointer generator network can help accurately replicate information by pointing to replicating words from the source text while preserving the ability to generate a vocabulary; attention may also be continually updated, preventing text from continually repeating.
Step 107: according to the abstract content, key elements such as business types and the like required by customers are matched; these key requirements can also be automatically filled in on the worksheet. The seat does not need to be manually input or selected, and the working efficiency of the seat can be effectively improved.
The method for forbidden language identification comprises the following steps:
step 201: a forbidden regular configuration, also called forbidden rules, is established.
Step 202: based on natural language processing algorithms, a negative service intention recognition model (NLP) is established.
Step 203: and identifying the forbidden language of the appeal text/customer service seat reply text through the forbidden language regular configuration and the negative service intention identification model.
The use of either a forbidden regular configuration alone or a passive service intention recognition model has poor accuracy. The NLP identifies the negative service intention, the regular configuration identifies the service forbidden language, the NLP hit result can be analyzed regularly based on the NLP generalized identification capability, the rules are adjusted, recall is increased, and the identification accuracy is improved. There is also a correlation of forbidden words with intent.
As shown in fig. 2, the intention recognition method:
step 301: a word vector model is obtained. The word vector model can be retrained according to the work order data.
Step 302: and obtaining a first word vector of the work order data through the word vector model, and constructing a second training set according to the work order data and the first word vector.
Wherein, single data can also be preprocessed: data cleaning, word segmentation, data set division, and the like.
Step 303: training the second training set based on a two-way long-short term memory network (BI-LSTM) to obtain an intention recognition model.
BI-LSTM has better text feature extraction efficiency and performance than a single LSTM network.
Step 304: and after the complaint text or the abstract content is segmented, obtaining a second word vector through a word vector model.
Step 305: and obtaining the corresponding intention of the second word vector through the intention recognition model.
The intention of the customer is extracted based on the intention recognition model, and reminding information such as knowledge points, error prone points and the like can be associated through the intention, so that the reply efficiency of the seat customer service can be improved.
In one particular embodiment, the intent includes: telephone charge inquiry, integral inquiry, main package inquiry, flow direct flushing and the like. For example, the complaint text is: i want to ask about that own number, after the money is paid in the last month, then the cost of looking up the money is reduced by 15 pieces, and i do not know which of the 15 pieces is going; the recognition intent is: and inquiring the deduction reason. For another example, the appeal text is: you listen me say that o me is that all old people do not use the intelligent machine, and the oldest mobile phone cannot surf the internet; the intent to identify is: the non-intelligent machine cannot subscribe.
The knowledge point pushing method comprises the following steps: establishing a first association of the intent with the knowledge point; and pushing knowledge points corresponding to the intention to the customer service seat according to the first association. The knowledge points comprise detailed information of the knowledge points, knowledge base addresses associated with the knowledge points and the like.
The error-prone point reminding method comprises the following steps: establishing a second association of the intent with the error prone point; and pushing the error-prone points corresponding to the intentions to customer service agents according to the second association.
The speaking reminding method comprises the following steps: obtaining context and service content according to the text or abstract content; obtaining a corresponding conversation according to the context and the service content; pushing the conversation for customer service agents. The call operation can be pushed to customer service agents in real time, customer appeal is solved, and working efficiency is improved.
The appeal text is obtained in step 103 by: transcribing the voice or conversation of the engine client into a appeal text through automatic voice recognition; and obtaining the appeal text of the client through a short message, chat software, an applet or a webpage.
Specifically, the transcription of speech into prosecution text includes any of the following methods:
the voice gateway interfaces with a media resource control protocol (Med ia Resource Contro l Protoco l, MRCP) based platform that transcribes speech into prosecution text through a speech recognition engine (ASR).
And forwarding the voice stream of the seat client to a voice recognition engine through the voice proxy, and writing the voice stream into a appeal text through the voice recognition engine. More specifically, the agent responds and sends a call start instruction to the voice agency through the agent control interface, and after initialization preparation is completed, the intelligent voice recognition capability platform transmits voice streams of the client and the agent to the voice agency, and the voice agency transmits the voice streams to the real-time voice transcription engine ASR for real-time voice stream transcription.
After the voice stream collected by the sound card collecting equipment and the microphone collecting equipment is divided by the audio stream collecting equipment, the voice stream is forwarded to the voice recognition engine, and the voice stream is transcribed into a appeal text by the voice recognition engine.
Through three voice transcription modes, the method is applicable to different docking scenes.
The invention also provides a customer service auxiliary system, as shown in fig. 3, which comprises a work order filling module 1, an intention recognition module 2, a forbidden language recognition module 3, a knowledge point pushing module 4, an error-prone point reminding module 5 and a speaking reminding module 6.
The work order filling module 1 is used for: extracting the appeal text through a work order content extraction model to obtain abstract content; and filling in the dynamic content of the work order template according to the abstract content.
The intention recognition module 2 is used for obtaining a first word vector of the work order data through a word vector model and constructing a second training set according to the work order data and the first word vector; training the second training set based on a two-way long-short-term memory network to obtain an intention recognition model; after the complaint text or the abstract content is segmented, a second word vector is obtained through a word vector model; and obtaining the corresponding intention of the second word vector through the consciousness recognition model.
The forbidden term identification module 3 is used for establishing forbidden term regular configuration; based on a natural language processing algorithm, a negative service intention recognition model is established; and identifying the forbidden language of the appeal text/customer service seat reply text through the forbidden language regular configuration and the negative service intention identification model.
The knowledge point pushing module 4 is configured to push, to a customer service agent, a knowledge point corresponding to an intention according to a first association between the intention and the knowledge point.
The error-prone point reminding module 5 is used for pushing error-prone points corresponding to the intention to the customer service seat according to the second association of the intention and the error-prone points.
The speaking reminding module 6 is used for obtaining context and service content according to the text or abstract content; obtaining a corresponding conversation according to the context and the service content; pushing the conversation for customer service agents.
The system of the invention also comprises a voice processing module 11, a model management module, a seat end 13 and the like.
The seat end 13 is used for: receiving the voice of a customer; receiving push speaking operation, knowledge points and error prone points; displaying the automatically filled work order; and recording calls that do not meet the specification or are problematic (forbidden), and alerting customer service representatives of non-specification phrases, notes, etc. The model management module is used for managing each model.
In one particular deployment, as in FIG. 4, the logical deployment includes an interaction layer, a capability layer, a service layer, and a data layer. The work order filling module 1, the intention recognition module 2, the forbidden language recognition module 3, the knowledge point pushing module 4, the error prone point reminding module 5 and the speaking reminding module 6 can be deployed on a service layer. As in fig. 5, the hardware deployment includes: DCN networks, switches, firewalls, load balancers, and servers.
The service and the application thereof in the product are stateless, and for stateless application, the linear expansion of the application can be realized by only building a service cluster and establishing load balance; by load balancing configuration, namely modifying the proportion of old and new version applications in the service cluster proxied by load balancing, the method can gradually and stably realize that the version is lifted without stopping.
The voice is converted into the appeal text in real time, the context and the service content can be analyzed, and the related content is intelligently pushed to the seat representative; auxiliary inquiry of high-frequency service and functions, incoming call cause analysis, error-prone point real-time reminding and emotion recognition are realized, and quick and accurate response of customer service is assisted; the working efficiency of the seat can be improved, and the labor cost is reduced.
The above is only a preferred embodiment of the present invention, and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. The customer service assisting method is characterized by comprising a method for filling a work order, wherein the method for filling the work order comprises the following steps of:
acquiring a first training set of work order data;
training the first training set based on a pointer generation network to obtain a work order content extraction model;
acquiring a appeal text of a customer;
extracting the appeal text through a work order content extraction model to obtain abstract content;
acquiring a work order template;
and filling in the dynamic content of the work order template according to the abstract content.
2. The customer service assistance method according to claim 1, wherein the specific method of job ticket filling comprises:
after cleaning and word segmentation are carried out on the worksheet data, a data set is obtained;
dividing the data set to obtain a first training set;
training the first training set based on a pointer generation network to obtain a work order content extraction model;
after data cleaning and word segmentation are carried out on the appeal text, inputting a work order extraction model to obtain abstract content;
and filling in the dynamic content of the work order template according to the abstract content.
3. The customer service assistance method as claimed in claim 1, further comprising a method of forbidden speech recognition:
establishing forbidden regular configuration;
based on a natural language processing algorithm, a negative service intention recognition model is established;
and identifying the forbidden language of the appeal text or the customer service seat reply text through the forbidden language regular configuration and the negative service intention identification model.
4. The customer service assistance method as claimed in claim 1, further comprising a method of intention recognition:
obtaining a word vector model;
obtaining a first word vector of the work order data through a word vector model, and constructing a second training set according to the work order data and the first word vector;
training the second training set based on a two-way long-short-term memory network to obtain an intention recognition model;
after the complaint text or the abstract content is segmented, a second word vector is obtained through a word vector model;
and obtaining the corresponding intention of the second word vector through the intention recognition model.
5. The customer service assistance method as claimed in claim 4, further comprising a knowledge point pushing method,
establishing a first association of the intent with the knowledge point;
and pushing knowledge points corresponding to the intention to the customer service seat according to the first association.
6. The customer service assistance method as claimed in claim 4, further comprising a method of error prone point alerting:
establishing a second association of the intent with the error prone point;
and pushing the error-prone points corresponding to the intentions to customer service agents according to the second association.
7. The customer service assistance method as claimed in claim 1, further comprising a speaking alert method:
obtaining context and service content according to the text or abstract content;
obtaining a corresponding conversation according to the context and the service content;
pushing the conversation for customer service agents.
8. A customer service assistance method as claimed in claim 1, wherein the appeal text is obtained by any one of:
transcribing the voice or conversation of the engine client into a appeal text through automatic voice recognition;
acquiring a client's appeal text through a short message, chat software, applet or webpage;
the method for converting the voice into the appeal text comprises any one of the following methods:
docking the voice gateway with a platform based on a media resource control protocol, wherein the platform transcribes voice into a appeal text through a voice recognition engine;
forwarding the voice stream of the seat client to a voice recognition engine through a voice proxy, and writing the voice stream into a appeal text through the voice recognition engine;
after the voice stream collected by the sound card collecting equipment and the microphone collecting equipment is divided by the audio stream collecting equipment, the voice stream is forwarded to the voice recognition engine, and the voice stream is transcribed into a appeal text by the voice recognition engine.
9. A customer service assistance system for implementing the customer service assistance method as claimed in any one of claims 1 to 8, said customer service assistance system comprising a work order filling module,
the work order filling module is used for: extracting the appeal text through a work order content extraction model to obtain abstract content; and filling in the dynamic content of the work order template according to the abstract content.
10. The customer service assistance system of claim 9, further comprising an intent recognition module, a forbidden language recognition module, a knowledge point pushing module, an error prone point reminding module, and a speaking reminding module;
the intention recognition module is used for obtaining a first word vector of the work order data through the word vector model and constructing a second training set according to the work order data and the first word vector; training the second training set based on a two-way long-short-term memory network to obtain an intention recognition model; after the complaint text or the abstract content is segmented, a second word vector is obtained through a word vector model; obtaining corresponding intention of the second word vector through the consciousness recognition model;
the forbidden term identification module is used for establishing forbidden term regular configuration; based on a natural language processing algorithm, a negative service intention recognition model is established; identifying forbidden languages of the appeal text/customer service seat reply text through the forbidden language regular configuration and the negative service intention identification model;
the knowledge point pushing module is used for pushing the knowledge point corresponding to the intention to the customer service seat according to the first association of the intention and the knowledge point;
the error-prone point reminding module is used for pushing error-prone points corresponding to the intention to the customer service seat according to the second association of the intention and the error prone point;
the speaking reminding module is used for obtaining context and service content according to the text or abstract content; obtaining a corresponding conversation according to the context and the service content; pushing the conversation for customer service agents.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311014713.9A CN117057813A (en) | 2023-08-11 | 2023-08-11 | Customer service assisting method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311014713.9A CN117057813A (en) | 2023-08-11 | 2023-08-11 | Customer service assisting method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117057813A true CN117057813A (en) | 2023-11-14 |
Family
ID=88652986
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311014713.9A Pending CN117057813A (en) | 2023-08-11 | 2023-08-11 | Customer service assisting method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117057813A (en) |
-
2023
- 2023-08-11 CN CN202311014713.9A patent/CN117057813A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112804400B (en) | Customer service call voice quality inspection method and device, electronic equipment and storage medium | |
| EP1602102B1 (en) | Management of conversations | |
| US10789943B1 (en) | Proxy for selective use of human and artificial intelligence in a natural language understanding system | |
| US7907705B1 (en) | Speech to text for assisted form completion | |
| CN101662549B (en) | Customer evaluation system and customer evaluation method based on voice | |
| CN103714813A (en) | Phrase spotting systems and methods | |
| CN108156824A (en) | Association centre's virtual assistant | |
| CN111681653A (en) | Call control method, device, computer equipment and storage medium | |
| WO2006031609A2 (en) | Machine learning | |
| CN111739519A (en) | Dialogue management processing method, device, equipment and medium based on voice recognition | |
| CN115665325B (en) | Intelligent outbound method, device, electronic equipment and storage medium | |
| CN112131358A (en) | Scene flow structure and intelligent customer service system applied by same | |
| CN115643341A (en) | Artificial intelligence customer service response system | |
| CN110677540A (en) | Intelligent voice recognition management system for consultation telephone of medical institution | |
| CN111429157A (en) | Method, device and equipment for evaluating and processing complaint work order and storage medium | |
| CN114818649A (en) | Service consultation processing method and device based on intelligent voice interaction technology | |
| WO2022260794A1 (en) | Reducing biases of generative language models | |
| CN113095852B (en) | Self-service distribution and service system through intelligent voice | |
| CN116340482A (en) | Multi-skill customer service auxiliary product based on enterprise WeChat combined with NLP engine | |
| CN117057813A (en) | Customer service assisting method and system | |
| CN116501844A (en) | Voice keyword retrieval method and system | |
| CN114707515A (en) | Method and device for judging dialect, electronic equipment and storage medium | |
| CN114171063A (en) | Real-time telephone traffic customer emotion analysis assisting method and system | |
| Zou et al. | Design of intelligent customer service report system based on automatic speech recognition and text classification | |
| CN110784603A (en) | Intelligent voice analysis method and system for offline quality inspection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication |