CN116485409A - Intelligent voice technology-based power service scene application method - Google Patents
Intelligent voice technology-based power service scene application method Download PDFInfo
- Publication number
- CN116485409A CN116485409A CN202211492450.8A CN202211492450A CN116485409A CN 116485409 A CN116485409 A CN 116485409A CN 202211492450 A CN202211492450 A CN 202211492450A CN 116485409 A CN116485409 A CN 116485409A
- Authority
- CN
- China
- Prior art keywords
- outbound
- business
- branch
- voice
- task
- 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 59
- 238000005516 engineering process Methods 0.000 title claims abstract description 21
- 238000007689 inspection Methods 0.000 claims abstract description 13
- 230000005611 electricity Effects 0.000 claims description 22
- 238000012545 processing Methods 0.000 claims description 21
- 238000004458 analytical method Methods 0.000 claims description 14
- 230000008569 process Effects 0.000 claims description 12
- 238000004140 cleaning Methods 0.000 claims description 10
- 230000009471 action Effects 0.000 claims description 6
- 238000004891 communication Methods 0.000 claims description 5
- 230000006870 function Effects 0.000 claims description 5
- 230000011218 segmentation Effects 0.000 claims description 5
- 238000011835 investigation Methods 0.000 claims description 4
- 230000002194 synthesizing effect Effects 0.000 claims description 3
- 238000012546 transfer Methods 0.000 claims description 3
- 238000012795 verification Methods 0.000 claims description 3
- 230000001133 acceleration Effects 0.000 abstract description 3
- 230000006872 improvement Effects 0.000 abstract description 2
- 238000007781 pre-processing Methods 0.000 abstract description 2
- 238000007405 data analysis Methods 0.000 description 5
- 238000010276 construction Methods 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 230000018109 developmental process Effects 0.000 description 3
- 230000003993 interaction Effects 0.000 description 3
- 238000012216 screening Methods 0.000 description 3
- 230000008676 import Effects 0.000 description 2
- 230000035515 penetration Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000002996 emotional effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000000977 initiatory effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000003908 quality control method Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
- 230000009466 transformation Effects 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
-
- 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
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
-
- 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
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/26—Speech to text systems
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Economics (AREA)
- Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Databases & Information Systems (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- General Engineering & Computer Science (AREA)
- General Business, Economics & Management (AREA)
- Data Mining & Analysis (AREA)
- Human Computer Interaction (AREA)
- Artificial Intelligence (AREA)
- Water Supply & Treatment (AREA)
- Tourism & Hospitality (AREA)
- Human Resources & Organizations (AREA)
- General Health & Medical Sciences (AREA)
- Development Economics (AREA)
- Accounting & Taxation (AREA)
- Primary Health Care (AREA)
- Public Health (AREA)
- Finance (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Telephonic Communication Services (AREA)
Abstract
The invention discloses an electric power business scene application method based on intelligent voice technology, which comprises the following steps: the method comprises the steps of designing a voice call operation scene, constructing an incoming call card, preprocessing outbound data, establishing an outbound task, executing the outbound task, failing strategy of the outbound task, voice quality inspection and the like; the method is applied to the fee-forcing business and business expansion return visit business of the power industry; the input of the voice outbound is accessed to develop the collect-urging work and the return visit work by the standardized service, so that the service quality of the collect-urging business and the business expansion business is improved, and the comprehensive improvement of the marketing service level of enterprises is promoted; the problems of manual work repeatability, low work efficiency, incomplete return visit coverage, occupation of a large amount of human resources and the like existing in the conventional business work technologies such as fee acceleration, business expansion return visit and the like in the existing power industry are solved.
Description
Technical Field
The invention belongs to the technical field of intelligent voice, and particularly relates to an electric power business scene application method based on an intelligent voice technology.
Background
The current power industry has a large number of repeated electric charge balance deficiency and payment-induced works and arrears and has great difficulty in development. Meanwhile, the access rate of the new industrial expansion is low in a traditional user return visit mode, and the return visit work quality control standard is often difficult to reach the standard, so that the return visit purpose is difficult to achieve. On the other hand, the management of the basic file cleaning work does not form complete closed-loop management, so that the management is performed manually on links of file problem discovery, correction, result evaluation and the like, telephone number misnumber, blank number and shutdown data statistics are time-consuming and labor-consuming, and the quality of related work is difficult to ensure. Because the customer file is inaccurate when sending short messages such as electricity charge notification, power failure notification, arrearage notification and the like, the cost is wasted, and the risk exists that the user is affected due to the fact that information cannot be timely transmitted to the user.
The conventional fee-forcing mode and the conventional return visit mode currently have the defects that: a large amount of manpower resource investment is needed, manual collection is promoted and revisited under the influence of holiday time, and the working time quantity cannot meet the service requirement; the repeatability of manual work is high, and the emotional office is easy to cause due to various electricity utilization user groups, so that the risk of complaints is high; the manual collection and return visit work efficiency is low, and only 150-200 service calls can be dialed by a person every day according to the traditional work feedback; the traditional working mode does not have an information tool to store and record business work records, and is not beneficial to later data rectification and data analysis.
Disclosure of Invention
The invention aims to solve the technical problems that: the utility model provides an electric power business scene application method based on intelligent voice technology, which solves the problems of manual work repeatability, low work efficiency, incomplete return visit coverage, occupation of a large amount of human resources and the like in the existing business work technologies such as the traditional fee-forcing, the business expansion return visit and the like in the existing electric power industry.
The technical scheme of the invention is as follows:
an electric power business scene application method based on intelligent voice technology, the method comprising:
step 1, newly-built business talk scenes: according to the electric power industry checking and collecting business, business expansion business and customer service business, designing a fee-forcing business scene, business expansion call-back business scene, customer service satisfaction investigation business scene and file check business scene;
step 2, constructing an incoming call name card: according to the region of the power enterprise, the power enterprise applies for the calling card from the local communication operator;
step 3, outbound data cleaning: exporting basic file information of a business scenario outbound target electricity user from an electric power enterprise marketing system, and cleaning the information;
step 4, newly building an outbound task: the method comprises the steps of importing intelligent voice outbound data in batches in a manual input or interface input mode, setting telephone lines according to power enterprises, associating intelligent voice outbound resources, selecting a voice outbound operation service scene, and setting starting time, outbound time interval and failure recall operation of outbound tasks;
step 5, executing outbound tasks: according to the configured outbound target electricity user basic file information, outbound appointed users; matching business speaking process branches according to the customer intention recognition result, returning the relevant branch result to be synthesized into voice broadcast for the electricity user to listen;
step 6, outbound failure processing: for the telephone with the failure of the outbound call, according to a secondary outbound call strategy configured by the task, dialing the telephone to the target electricity user again or marking according to an outbound call failure prompt;
step 7, quality inspection of the outbound result: after the outbound task is finished, quality inspectors perform quality inspection on outbound details in the intelligent voice outbound system.
The business telephone scene comprises more than one telephone sub-node, each telephone sub-node comprises more than one telephone branch node, and various telephone functions are formed by constructing the relationship among the telephone branch nodes.
The fixed telephone of the power enterprise molecular company unit is used as the calling card.
The method for cleaning the data of the information comprises the following steps: removing problem user information by using electricity user basic file information in a repeated judgment and telephone number regular judgment mode; and meanwhile, simplifying the address information to remove redundant address information.
The branch types of the speaking operation flow include a common branch, a jump branch, a key collection branch, a key branch, a condition judgment branch and an information query branch, the speaking operation branch sets the matching condition of the branch and the branch, and when the client intention recognition result is matched with the current branch, the flow jumps to the next branch.
The voice text escape is to detect, extract and reduce noise of the voice of the customer by using the processing mode of the signal, obtain the voice for the recognition engine to process, and transfer the voice into text and return; semantic recognition is to return the true intent of the text by first word segmentation of the text and then matching with a pre-configured key intent or utterance.
Defining 1 or more than 1 branches for each service node according to actual service requirements, wherein each service node at least comprises a default processing branch; the node branch matching adopts a keyword, phrase and regular matching mode, and when the intention recognition result cannot be matched with the vocabulary entry or the regular condition configured by the node branch, the intention recognition result is matched with a built-in knowledge question-answering system; returning to a default branch for processing when the intention of the client cannot be understood or identified; when the intention of the customer cannot be understood more than once, the call is ended and marked by adopting a set of ending language according to the flow setting.
When the outbound fails, marking according to the outbound failure feedback category of the operator, and if the outbound failure type is busy, in a call, unable to be connected or shut down, re-initiating the outbound again according to the outbound failure re-call setting interval for a certain time; and marking information if the failure type is a null number or a shutdown.
The method for detecting the quality of the outbound details comprises the following steps: and checking the intelligent voice application semantic analysis and semantic recognition according to the semantic analysis and the recording, marking error analysis and error recognition, manually correcting semantic matching, marking problem data which is not self-feedback of the power utilization user, feeding back to marketing business personnel of an electric power enterprise, and correcting after verification by the business personnel.
An intelligent voice outbound platform for intelligent voice outbound, the platform comprising:
and (3) a new module for voice operation: the method is used for creating a corresponding robot BOT operation flow according to service requirements;
and a task new module: the method comprises the steps of inputting outbound task data in batches through a manual input or interface input mode, setting related parameter information, associating a robot BOT (boil off procedure), and setting starting time of the outbound task, a time range of the outbound and a strategy of failed retry;
the task execution module is used for starting and executing a task after the task setting is completed, calling a specific client according to the configuration data, receiving user voice and performing text escape by the robot after the call is connected, performing semantic understanding analysis on the escape text to obtain a client intention recognition result, matching specific branches through a speaking process, realizing corresponding business operation, returning the result and synthesizing the result into voice to be played for the client to listen;
the exception handling module is used for dialing the customer call again according to the retry strategy of the task configuration under the condition of not being normally connected;
the central processing module is used for sending instructions to the modules and completing related actions;
the process newly-built module, the task execution module and the exception handling module are all electrically connected with the central processing module.
The invention has the beneficial effects that:
the invention provides an electric power business scene application method based on intelligent voice technology, which uses intelligent voice outbound technology, combines with telephone special lines of communication operators, automatically outbound through intelligent voice robot resources according to established voice operation and target user groups, completes the telephone scene application of related fee-forcing business, business expansion return visit business, satisfaction investigation business and the like, replaces the original manual business development work, realizes business scenes such as electric power industry notification business (such as electric fee prompt, power failure notification and the like), return visit business (business expansion return visit, customer service complaint return visit and the like), policy penetration business (electric power policy penetration, official product (south network online, public number and the like) introduction and the like) and the like through an intelligent voice outbound mode, comprehensively clearing basic file information of an electric user, positioning problem data, and assisting an electric power enterprise to complete data quality improvement, and assisting the electric power enterprise to advance digital development; meanwhile, the invention is beneficial to reducing the manual customer service cost of enterprises, improving the service efficiency of the enterprises, improving the service quality of clients and improving the satisfaction degree of the clients.
Compared with the prior art, the invention has the following advantages: the intelligent voice outbound technology is applied to business scenes such as fee acceleration, business expansion return visit, satisfaction survey and the like in the power industry, is widely applicable to the business fields, and is generally applicable to business scenes such as notification, return visit, marketing and the like; the efficiency of outbound service can be greatly improved, and the labor cost is effectively reduced; the outbound strategy is efficient, flexible and customized, and the outbound result data is complete and standard and uniform; the method can restore natural and smooth real person conversation scenes by supporting modes of multi-turn conversation, intelligent breaking, silent voice recognition, context association, knowledge base inquiry and the like, and is worthy of popularization and use.
The problems of manual work repeatability, low work efficiency, incomplete return visit coverage, occupation of a large amount of human resources and the like existing in the conventional business work technologies such as fee acceleration, business expansion return visit and the like in the existing power industry are solved.
Drawings
FIG. 1 is a schematic flow chart of the present invention;
fig. 2 is a schematic diagram of the intelligent voice outbound platform.
Detailed Description
The invention solves the technical problems through the following technical proposal, and the invention comprises the following steps:
f1 new business session scenario
According to the electric power industry checking and collecting business, business expansion business and customer service business, a fee-forcing business scene, business expansion call-back business scene, customer service satisfaction investigation business scene, file check business scene and other business scenes are designed, and the business scenes are composed of business-inherent business sub-nodes.
The business telephone scene comprises a plurality of telephone sub-nodes, each telephone sub-node comprises a plurality of telephone branch nodes, and various telephone functions are formed by constructing the relationship among the telephone branch nodes.
F2 construction of calling card
According to the region of the power enterprise, the power enterprise applies for the calling business card from the local communication carrier, and when the power user receives the telephone of the related service scene, the telephone can be seen to originate from the power enterprise, so that the credibility of the telephone source can be increased, and the answering rate of the service telephone is improved.
The fixed telephone of the power enterprise molecular company unit is used as the calling card.
F3 outbound data cleaning
According to the business needs of the power enterprises, the basic file information of the outgoing call target electricity utilization users of relevant business scenes is derived in the marketing system of the power enterprises, the information comprises basic information such as the numbers, names and addresses of the electricity utilization users, repeated calls and wrong calls are cleaned once according to the content of the basic file information of the users, and meanwhile, the repeated information of the addresses is simplified, for example, the information such as XX area/county of XX City and XX is omitted, keywords are refined, the voice operation is guaranteed to be accepted by the electricity utilization users, and the answering will is higher.
The method comprises the steps of performing primary screening on basic archives of electricity users led out by an electric power enterprise marketing system, and removing problem user information in modes of repeatability judgment, regular telephone number judgment and the like; meanwhile, simplifying the address information, removing redundant address information, and guaranteeing that the address information is simplified and critical, so that voice interaction is more available and recognizable, the voice interaction effective rate is improved, and the intelligent voice external call completing rate is improved.
F4 outbound task new construction
The intelligent voice outbound data is imported in batches in a manual entry or interface entry mode, relevant telephone lines are set according to power enterprises, intelligent voice outbound resources (robots) are associated, a voice outbound operation service scene is selected, the starting time of outbound tasks, outbound time intervals and failed recall operation are set.
The outbound task is provided with a plurality of outbound call operation service scenes; when the user newly establishes an outbound task, a corresponding business telephone operation scene is selected according to business requirements, for example, a fee-urging business selection fee-urging outbound telephone operation scene and a business expansion return visit business selection business expansion outbound telephone operation scene.
F5 executing outbound tasks
After the setting of the outbound task is completed, the outbound task is executed, the task designates the user according to the configured outbound target electricity user basic file information, after the call is connected, the robot receives user voice according to the selected voice outbound operation service scene and carries out semantic recognition, and simultaneously carries out voice text escape to obtain a customer intention recognition result, matches a business operation flow branch according to the customer intention recognition result, returns a relevant branch result and synthesizes voice broadcasting to the electricity user for listening.
The call flow branch types include: the normal branch, the jump branch, the key collection branch, the key branch, the condition judgment branch and the information inquiry branch, the operation branch can set the matching condition of the branches and the branches, and when the client intention recognition result matches with the current branch, the flow jumps to the next branch of the specific branch.
The voice text escape is to detect, extract the characteristics and reduce the noise of the customer voice by using the signal processing mode, obtain the voice which can be processed by the recognition engine, and transfer the voice into text for return.
Semantic recognition is to return the true intent of the text by first word segmentation of the text and then matching with a pre-configured key intent or utterance.
Further, in the step F5, the business session procedure defines 1 or more branches for each business node according to the actual business requirement, and each business node includes at least one default processing branch; the node branch matching adopts a keyword, phrase and regular matching mode, and when the intention recognition result cannot be matched with the vocabulary entry or the regular condition configured by the node branch, the intention recognition result is matched with a built-in knowledge question-answering system.
Returning to a default branch for processing when the intention of the client cannot be understood or identified; when the customer intention cannot be understood correctly for many times, the call is ended by a set of ending words according to the process setting condition, and the call is marked.
F6 outbound failure handling
For the telephone with failure (busy or refused) of the outbound call, dialing the telephone again to the target electricity user according to the secondary outbound call strategy configured by the task; or marking according to the outbound failure (idle sign, stop) prompt.
Further, in the step F6, when the outbound fails, marking is performed according to the outbound failure feedback category of the operator, if the outbound failure type is busy, in a call, unable to be connected, shut down, etc., then the outbound is initiated again at a certain time interval according to the outbound failure re-call setting; if the failure type is a null number, a stop, etc., the information is marked for subsequent data analysis.
F7 outbound quality inspection
After the outbound task is finished, quality inspectors perform quality inspection on outbound details in an intelligent voice outbound system, intelligent voice application semantic analysis and semantic recognition are checked according to semantic analysis and recording, error analysis and error recognition are marked, semantic matching is corrected manually, meanwhile, the marked electricity users feed back problem data which is not personal, and the problem data is fed back to marketing business personnel of an electric power enterprise in a later period, and the business personnel verify and correct the problem data.
The manual foreign call detail quality inspection of the quality inspector can improve the semantic recognition capability of the intelligent voice robot, and comprises dialect screening matching, bilingual screening matching and the like.
The invention also provides an intelligent voice outbound platform, which completes voice outbound service by adopting the method, and comprises the following steps:
the conversation creation module is used for creating a corresponding robot BOT conversation process according to the service requirement;
the task creation module is used for importing outbound task data in batches in a manual input or interface import mode, setting related parameter information, associating a robot BOT operation flow, setting the starting time of the outbound task and the time range of the outbound, and failing retry strategies;
the task execution module is used for starting and executing a task after the setting is completed, calling a specific client according to configuration data, receiving user voice and performing text escape by the robot after the call is connected, performing semantic understanding analysis on the escape text to obtain a client intention recognition result, matching specific branches through a speaking flow, realizing corresponding business operation, returning the result and synthesizing the result into voice to be played for the client to listen;
the exception handling module is used for dialing the customer call again according to the retry strategy of the task configuration under the condition of not being normally connected;
the central processing module is used for sending instructions to each module to complete related actions;
the process newly-built module, the task execution module and the exception handling module are all electrically connected with the central processing module.
The following describes in detail the examples of the present invention, which are implemented on the premise of the technical solution of the present invention, and detailed embodiments and specific operation procedures are given, but the scope of protection of the present invention is not limited to the following examples.
Example one (fee-forcing scenario)
F1 new business session scenario
In this embodiment, the fee-based operation flow in step F1 includes a plurality of fee-based operation sub-flows; each speaking sub-flow comprises a plurality of flow branch nodes, and a specific service function is formed by constructing the relationship among the flow branch nodes
F2 construction of calling card
In this example, the calling card in step F2 selects a fixed telephone of the power enterprise, and provides related data and a telephone directory according to the requirements of the local communication operator, and establishes the calling card, and when the user receives the call, there is a display of the calling card "XX power supply office".
F3 outbound data cleaning
In this example, the outbound data cleaning in step F3 includes performing data quality inspection on the information such as the user name, the user phone, the user arrearage amount, the user address, etc. in the arrearage list, and checking whether the user is arrearage currently, so as to ensure the accuracy of the voice outbound target, and reduce unnecessary resource consumption during outbound.
F4 outbound task new construction
According to the service requirement of the charge-urging scene, the charge-urging scene outbound application is developed in the charge-urging period, outbound task data is imported in batches in a manual input or interface import mode, related parameter information such as outbound lines and the like is set according to power units, the distributed intelligent voice outbound robots are associated, the charge-urging operation flow is associated, the starting time of the outbound tasks, the outbound time, the failed retry strategy and the like are set.
F5 initiating execution of outbound tasks
After the related configuration is completed, a task is started and executed, the task calls a specific client according to the configuration data, after the call is connected, the robot receives user voice and carries out text escape, semantic understanding analysis is carried out on the text after escape to obtain an intention recognition result, the operation flow is matched with a specific branch, operation such as service information inquiry is realized, and the result is returned and synthesized into voice to be played for the client to listen. Returning to a default branch for processing when the intention of the client cannot be understood or identified; in addition, in the conversation process of the intelligent voice robot and the user, generated conversation data, service data and the like can be recorded by the system for subsequent voice quality inspection and data analysis.
In this embodiment, the voice text escape in step F5 is a voice recognition function, and the voice processed by the most suitable recognition engine is obtained by performing preprocessing such as detection, feature extraction, noise reduction, etc. on the voice of the speaker by using the signal processing method, and the voice is transcribed into text and returned.
In this embodiment, the semantic understanding analysis in step F5 is to firstly segment the text, then match the text with a pre-configured key intention or description, and return the true intention of the text;
and in the specific matching, the text word segmentation result is precisely matched with the keywords in the configured branches (namely whether the content of the keywords are matched or not) or the regular expression description of the keywords is subjected to text matching (namely, each word segmentation result is subjected to text rule matching through the regular expression).
In this embodiment, the speaking flow defines 1 or more branches for each service node according to actual service requirements, where each service node at least includes a default processing branch; the node branch matching supports the modes of keywords, phrases, regular matching and the like, when the intention recognition result cannot be matched with terms or regular conditions configured by the node branch, the intention recognition result is matched with a built-in knowledge question-answering system, if the matching is to a certain knowledge, the corresponding answer is answered, and then the current node is returned to or the conversation sub-flow is turned; in addition, for cases where multiple intents do not match, further processing actions can be done by configuring dialog context rules. The method has the main advantages that the efficiency of the intended matching branch is high, the efficiency is high, the processing modes of the situation which cannot be matched are multiple and comprehensive, and the interaction between people is simulated more truly.
Each speaking flow has a related question-answer knowledge base, when the client intention result cannot be matched with the node branch, the matched questions are searched in the question-answer knowledge base, if the questions can be correctly matched, the answers of the questions are returned, the synthesized audio is played to the incoming client for listening, the previous node is returned, and the next flow is continued.
F6 outbound failure handling
In this embodiment, for the case of abnormal connection, the client phone may be dialed again according to the retry strategy configured by the task, and meanwhile, relevant data such as a null sign and a shutdown are marked for subsequent data analysis and use.
F7 outbound quality inspection
In this embodiment, after the outbound is finished, outbound quality inspection personnel perform outbound recording quality inspection, firstly correct semantic recognition errors in related telephone operation, record related data which is fed back by the electricity users and is not the person, give marketing service personnel of the electric power enterprises in subsequent data analysis, and correct and modify basic file information after verification by the service personnel, so as to improve basic data quality of the electricity users, thereby providing a data basis for digital transformation of the electric power enterprises.
In summary, in the method for applying the intelligent voice outbound to the power service scene according to the embodiment, the client call is translated into the call text through voice recognition, the intention recognition is performed on the call text, the result of the intention recognition is matched with the call flow branch, the next flow action is performed, for the node action itself, operations such as service information interface query, short message sending, manual conversion and the like may exist according to the call flow setting, and for the answer result or content of the robot, voice is synthesized through voice synthesis and fed back to the client.
Claims (10)
1. An electric power business scene application method based on intelligent voice technology is characterized in that: the method comprises the following steps:
step 1, newly-built business talk scenes: according to the electric power industry checking and collecting business, business expansion business and customer service business, designing a fee-forcing business scene, business expansion call-back business scene, customer service satisfaction investigation business scene and file check business scene;
step 2, constructing an incoming call name card: according to the region of the power enterprise, the power enterprise applies for the calling card from the local communication operator;
step 3, outbound data cleaning: exporting basic file information of a business scenario outbound target electricity user from an electric power enterprise marketing system, and cleaning the information;
step 4, newly building an outbound task: the method comprises the steps of importing intelligent voice outbound data in batches in a manual input or interface input mode, setting telephone lines according to power enterprises, associating intelligent voice outbound resources, selecting a voice outbound operation service scene, and setting starting time, outbound time interval and failure recall operation of outbound tasks;
step 5, executing outbound tasks: according to the configured outbound target electricity user basic file information, outbound appointed users; matching business speaking process branches according to the customer intention recognition result, returning the relevant branch result to be synthesized into voice broadcast for the electricity user to listen;
step 6, outbound failure processing: for the telephone with the failure of the outbound call, according to a secondary outbound call strategy configured by the task, dialing the telephone to the target electricity user again or marking according to an outbound call failure prompt;
step 7, quality inspection of the outbound result: after the outbound task is finished, quality inspectors perform quality inspection on outbound details in the intelligent voice outbound system.
2. The power business scenario application method based on the intelligent voice technology as claimed in claim 1, wherein the method comprises the following steps: the business telephone scene comprises more than one telephone sub-node, each telephone sub-node comprises more than one telephone branch node, and various telephone functions are formed by constructing the relationship among the telephone branch nodes.
3. The power business scenario application method based on the intelligent voice technology as claimed in claim 1, wherein the method comprises the following steps: the fixed telephone of the power enterprise molecular company unit is used as the calling card.
4. The power business scenario application method based on the intelligent voice technology as claimed in claim 1, wherein the method comprises the following steps: the method for cleaning the data of the information comprises the following steps: removing problem user information by using electricity user basic file information in a repeated judgment and telephone number regular judgment mode; and meanwhile, simplifying the address information to remove redundant address information.
5. The power business scenario application method based on the intelligent voice technology as claimed in claim 1, wherein the method comprises the following steps: the branch types of the speaking operation flow include a common branch, a jump branch, a key collection branch, a key branch, a condition judgment branch and an information query branch, the speaking operation branch sets the matching condition of the branch and the branch, and when the client intention recognition result is matched with the current branch, the flow jumps to the next branch.
6. The power business scenario application method based on the intelligent voice technology as claimed in claim 1, wherein the method comprises the following steps: the voice text escape is to detect, extract and reduce noise of the voice of the customer by using the processing mode of the signal, obtain the voice for the recognition engine to process, and transfer the voice into text and return; semantic recognition is to return the true intent of the text by first word segmentation of the text and then matching with a pre-configured key intent or utterance.
7. The power business scenario application method based on the intelligent voice technology as claimed in claim 1, wherein the method comprises the following steps: defining 1 or more than 1 branches for each service node according to actual service requirements, wherein each service node at least comprises a default processing branch; the node branch matching adopts a keyword, phrase and regular matching mode, and when the intention recognition result cannot be matched with the vocabulary entry or the regular condition configured by the node branch, the intention recognition result is matched with a built-in knowledge question-answering system; returning to a default branch for processing when the intention of the client cannot be understood or identified; when the intention of the customer cannot be understood more than once, the call is ended and marked by adopting a set of ending language according to the flow setting.
8. The power business scenario application method based on the intelligent voice technology as claimed in claim 1, wherein the method comprises the following steps: when the outbound fails, marking according to the outbound failure feedback category of the operator, and if the outbound failure type is busy, in a call, unable to be connected or shut down, re-initiating the outbound again according to the outbound failure re-call setting interval for a certain time; and marking information if the failure type is a null number or a shutdown.
9. The power business scenario application method based on the intelligent voice technology as claimed in claim 1, wherein the method comprises the following steps: the method for detecting the quality of the outbound details comprises the following steps: and checking the intelligent voice application semantic analysis and semantic recognition according to the semantic analysis and the recording, marking error analysis and error recognition, manually correcting semantic matching, marking problem data which is not self-feedback of the power utilization user, feeding back to marketing business personnel of an electric power enterprise, and correcting after verification by the business personnel.
10. Be applied to intelligent pronunciation outbound platform that intelligent pronunciation outbound called, its characterized in that: the platform comprises:
and (3) a new module for voice operation: the method is used for creating a corresponding robot BOT operation flow according to service requirements;
and a task new module: the method comprises the steps of inputting outbound task data in batches through a manual input or interface input mode, setting related parameter information, associating a robot BOT (boil off procedure), and setting starting time of the outbound task, a time range of the outbound and a strategy of failed retry;
the task execution module is used for starting and executing a task after the task setting is completed, calling a specific client according to the configuration data, receiving user voice and performing text escape by the robot after the call is connected, performing semantic understanding analysis on the escape text to obtain a client intention recognition result, matching specific branches through a speaking process, realizing corresponding business operation, returning the result and synthesizing the result into voice to be played for the client to listen;
the exception handling module is used for dialing the customer call again according to the retry strategy of the task configuration under the condition of not being normally connected;
the central processing module is used for sending instructions to the modules and completing related actions;
the process newly-built module, the task execution module and the exception handling module are all electrically connected with the central processing module.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211492450.8A CN116485409A (en) | 2022-11-25 | 2022-11-25 | Intelligent voice technology-based power service scene application method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211492450.8A CN116485409A (en) | 2022-11-25 | 2022-11-25 | Intelligent voice technology-based power service scene application method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116485409A true CN116485409A (en) | 2023-07-25 |
Family
ID=87220157
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211492450.8A Pending CN116485409A (en) | 2022-11-25 | 2022-11-25 | Intelligent voice technology-based power service scene application method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116485409A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117041430A (en) * | 2023-10-09 | 2023-11-10 | 成都乐超人科技有限公司 | Method and device for improving outbound quality and robustness of intelligent coordinated outbound system |
-
2022
- 2022-11-25 CN CN202211492450.8A patent/CN116485409A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117041430A (en) * | 2023-10-09 | 2023-11-10 | 成都乐超人科技有限公司 | Method and device for improving outbound quality and robustness of intelligent coordinated outbound system |
| CN117041430B (en) * | 2023-10-09 | 2023-12-05 | 成都乐超人科技有限公司 | Method and device for improving outbound quality and robustness of intelligent coordinated outbound system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7907705B1 (en) | Speech to text for assisted form completion | |
| CN111192060B (en) | Full-channel self-service response implementation method based on power IT service | |
| US11095773B2 (en) | Coaching in an automated communication link establishment and management system | |
| CN105592237B (en) | A kind of method, apparatus and intelligent customer service robot of session switching | |
| CN110263144A (en) | A kind of answer acquisition methods and device | |
| US7136814B1 (en) | Syntax-driven, operator assisted voice recognition system and methods | |
| US9258423B1 (en) | Contextual lead generation in an automated communication link establishment and management system | |
| US8605868B2 (en) | System and method for externally mapping an interactive voice response menu | |
| US20080159495A1 (en) | System and Method of Use for Indexing Automated Phone Systems | |
| US20110022386A1 (en) | Speech recognition tuning tool | |
| TWI770461B (en) | System and method of instant-messaging bot supporting human-machine symbiosis | |
| US8948372B1 (en) | Contextual lead generation in an automated communication link establishment and management system | |
| CN108428148A (en) | Active smart phone marketing method and system | |
| CN113067950A (en) | Intelligent call platform | |
| CN116485409A (en) | Intelligent voice technology-based power service scene application method | |
| CN104580762A (en) | Method and device for realizing call forwarding according to language | |
| CN113779217A (en) | Intelligent voice outbound service method and system based on human-computer interaction | |
| CN113282732A (en) | Intelligent voice interaction comprehensive management platform and method | |
| US9986076B1 (en) | Closed loop calling process in an automated communication link establishment and management system | |
| US20230239402A1 (en) | Coaching in an automated communication link establishment and management system | |
| CN110298543A (en) | Business tracking method, device, computer equipment and storage medium | |
| US9876886B1 (en) | System and method for automatic update of calls with portable device | |
| CN113542509B (en) | Emergency processing method, device, storage medium and equipment | |
| CN111988476B (en) | Automatic voice cooperative working method of customer service system | |
| US11778090B1 (en) | Communication monitoring systems and methods |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |