Intelligent outbound platform based on AI technology

Abstract

The invention discloses an intelligent outbound platform based on an AI technology, and belongs to the technical field of outbound supervision. The method and the device are used for solving the technical problem of poor historical call data processing and sharing utilization management effect between different intelligent outbound platforms in the existing scheme. By monitoring, counting, processing and analyzing outbound shared data associated with different outbound objects by an outbound platform in the earlier stage of outbound implementation and dynamically managing subsequent outbound implementation of the different outbound objects according to an analysis result, targeted processing and analysis of historical outbound states of the outbound objects in the earlier stage of outbound implementation are realized; according to the method, the outbound call process implemented by different outbound call object sets subjected to early-stage screening processing is supervised, data analysis is performed on monitoring statistical data corresponding to the different outbound call object sets, different outbound call objects are dynamically marked according to an analysis result, and outbound call shared data are uploaded and shared for the marked different outbound call objects.

Description

Technical Field

[0001] The present invention relates to the technical field of outbound call supervision, and specifically relates to an intelligent outbound call platform based on AI technology. Background Art

[0002] An intelligent outbound call platform is a system that uses artificial intelligence technologies, especially technologies such as speech recognition, natural language processing, and machine learning, to automatically make phone calls and interact with humans.

[0003] When existing intelligent outbound call platforms are implemented, most of them still stay at the statistics display of single call and answer data, and cannot supervise and share the historical automatic call data of different call subjects, conduct multi-dimensional data processing and analysis on different call data, and dynamically manage the calls of different call objects according to the analysis results. There are poor management effects in the processing and shared utilization of historical call data between different intelligent outbound call platforms. Summary of the Invention

[0004] The purpose of the present invention is to provide an intelligent outbound call platform based on AI technology to solve the technical problem of poor management effects in the processing and shared utilization of historical call data between different intelligent outbound call platforms in the existing solutions.

[0005] The purpose of the present invention can be achieved through the following technical solutions: An intelligent outbound call platform based on AI technology includes an early-stage supervision module for outbound calls, which is used to monitor and count the outbound shared data associated with different outbound call objects during the early stage of outbound call implementation by the outbound call platform, process and analyze it, and dynamically manage the subsequent outbound call implementation of different outbound call objects according to the analysis results; the analysis results include the first outbound call object set, the third outbound call object set, the fourth outbound call object set, and the fifth outbound call object set obtained by processing combinations; A late-stage supervision module for outbound calls, which is used to monitor and count the outbound call processes implemented by the outbound call platform for the first outbound call object set, the fourth outbound call object set, and the fifth outbound call object set, conduct data analysis on the monitoring and statistical data corresponding to different outbound call object sets, and dynamically mark different outbound call objects according to the analysis results, and upload and share the outbound shared data of the marked different outbound call objects.

[0006] Preferably, obtain the outbound call numbers corresponding to different outbound call objects during the early stage of outbound call implementation by the outbound call platform, and input the outbound call numbers corresponding to different outbound call objects into the outbound shared database for traversal and retrieval; If the input outbound call number does not exist in the outbound shared database, mark the outbound call object to which the outbound call number belongs as the first outbound call object; If the called number entered exists in the outbound shared database, mark the outbound object to which the called number belongs as the second outbound object, perform data analysis on the historical outbound effect coefficient associated with the second outbound object, and dynamically control the implementation of subsequent outbound calls for the second outbound object according to the analysis results.

[0007] Preferably, the steps for obtaining the historical outbound effect coefficient corresponding to different second outbound objects include: Obtain the total number of historical outbound calls and the total number of answered historical outbound calls corresponding to the second outbound object, as well as the corresponding reply data for each answer, and use the formula to calculate and obtain the historical outbound effect coefficient HJ of the second outbound object; in the formula, N0 and N1 are respectively the total number of historical outbound calls and the total number of answered historical outbound calls corresponding to the second outbound object; α is the answer effect factor.

[0008] Preferably, the steps for obtaining the answer effect factor α include: Obtain the corresponding reply data for each answer of the second outbound object, and classify according to the reply behavior in the reply data; If the second outbound object does not reply, mark the historical outbound call as the first outbound call; If the second outbound object replies, mark the historical outbound call as the second outbound call, and count the outbound communication duration corresponding to the second outbound call; Count and analyze the total number of all first outbound calls corresponding to the second outbound object. If the total number of first outbound calls is the same as its total number of historical inbound calls, update the mark of the corresponding second outbound object to the third outbound object, and set the corresponding answer effect factor to 0.

[0009] Preferably, if the total number of first outbound calls is different from its total number of historical inbound calls, then use the formula for the supervision data corresponding to all its first outbound calls and second outbound calls to calculate and obtain the answer effect factor of the second outbound object; in the formula, N2 is the total number of first outbound calls; NZ is the total number of first outbound calls and second outbound calls; i is different second outbound calls, i = 1, 2, 3,..., n; n is a positive integer, which is the total number of second outbound calls; Ti is the outbound communication duration corresponding to different second outbound calls; T0 is the standard outbound communication duration; A is the first reply standard value; B is the second reply standard value; Perform data analysis on the historical outbound effect coefficient associated with the second outbound object through the outbound supervision recognition model, and output the corresponding outbound supervision evaluation value WP.

[0010] Preferably, the expression of the outbound supervision recognition model is ; in the formula, HJ0 is the historical outbound supervision standard value; An outbound supervision evaluation value of 0 indicates that the historical outbound call effect of the second outbound call object is slightly effective, and the corresponding second outbound call object label is updated to the fourth outbound call object; An outbound supervision evaluation value of 1 indicates that the historical outbound call effect of the third outbound call object is highly effective, and the corresponding second outbound call object label is updated to the fifth outbound call object.

[0011] Preferably, according to the labels, the same first outbound call objects, third outbound call objects, fourth outbound call objects, and fifth outbound call objects are sorted and combined respectively to obtain the corresponding first outbound call object set, third outbound call object set, fourth outbound call object set, and fifth outbound call object set. And implement the first outbound supervision plan for all the first outbound call objects in the first outbound call object set, and implement the second outbound supervision plan for all the fourth outbound call objects in the fourth outbound call object set and all the fifth outbound call objects in the fifth outbound call object set.

[0012] Preferably, when conducting outbound supervision on all the first outbound call objects in the first outbound call object set according to the first outbound supervision plan, the labeled first outbound call objects are updated to the third outbound call objects, fourth outbound call objects, or fifth outbound call objects according to the unanswered, answered but not replied, or answered and replied call behaviors corresponding to the first outbound call objects.

[0013] Preferably, when conducting outbound supervision on all the fourth outbound call objects in the fourth outbound call object set and all the fifth outbound call objects in the fifth outbound call object set according to the second outbound supervision plan, the corresponding historical outbound call effect coefficients are calculated and updated according to the unanswered, answered but not replied, or answered and replied call behaviors corresponding to the fourth outbound call objects and the fifth outbound call objects, and data analysis is performed on the calculated and updated historical outbound call effect coefficients. According to the analysis results, the corresponding fourth outbound call objects and fifth outbound call objects are dynamically labeled.

[0014] Preferably, the dynamic labeling results corresponding to all the first outbound call objects, fourth outbound call objects, and fifth outbound call objects are used to upload and share-update the corresponding outbound call shared data.

[0015] Compared with the existing solution, the beneficial effects achieved by the present invention are: The present invention monitors and statistically analyzes the outbound call shared data associated with different outbound call objects in the early stage of outbound call implementation on the outbound call platform, and dynamically manages the subsequent outbound call implementation of different outbound call objects according to the analysis results. It realizes targeted processing and analysis of the historical outbound call status of outbound call objects in the early stage of outbound call, so as to implement targeted outbound supervision plans for different outbound call objects subsequently, improving the autonomous supervision and analysis classification effect of different outbound call objects in the early stage of outbound call, and the targeted service effect in the process of outbound call implementation.

[0016] The present invention monitors the outbound calling process implemented for different outbound calling object sets screened and processed in the early stage, analyzes the monitoring and statistical data corresponding to different outbound calling object sets, dynamically marks different outbound calling objects according to the analysis results, and uploads and shares the outbound calling shared data for the marked different outbound calling objects, thereby realizing targeted traceability and expansion management for outbound calling objects with different classification marks in the early stage of calling, dynamically updating and sharing the calling marks corresponding to different calling objects, and improving the management effect of historical calling data processing and shared utilization between different intelligent outbound calling platforms. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The present invention will be further described below with reference to the accompanying drawings.

[0018] Figure 1 It is a flow block diagram of the operation of an intelligent outbound calling platform based on AI technology according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] The technical solutions in 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. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0020] As Figure 1 shown, the present invention is an intelligent outbound calling platform based on AI technology, including a pre-outbound-calling supervision module and a post-outbound-calling supervision module; The pre-outbound-calling supervision module is used to monitor and statistically analyze the outbound calling shared data associated with different outbound calling objects by the outbound calling platform in the early stage of outbound calling, and dynamically manage the subsequent outbound calling implementation of different outbound calling objects according to the analysis results; it includes: Obtain the outbound calling numbers corresponding to different outbound calling objects by the outbound calling platform in the early stage of outbound calling, and input the outbound calling numbers corresponding to different outbound calling objects into the outbound calling shared database for traversal and retrieval; Among them, the outbound calling platform is a multi-channel intelligent robot product launched based on the microservice architecture, NLP technology, and voice technology, aiming at the problems of outbound calling, voice customer service, and online customer service in the traditional customer service field; Utilize advanced artificial intelligence technologies and robotic neural network architectures to replace manual operations with AI outbound robots for telephone-based business marketing, customer follow-up, etc. Provide functions such as automatic call dialing, speech recognition, intelligent conversation, and customer intention feedback according to business scenarios, target customer groups for marketing, products, and content, thereby reducing the time and cost of manual operations, standardizing and effectively expressing sales scripts, increasing the success rate of customer marketing, achieving the transformation and upgrading of traditional manual communication methods, reducing enterprise operation costs, and improving customer service efficiency; Intelligent robots streamline the entire service process. Complex voice outbound tasks and voice inbound customer service and text online customer service configurations can be completed with simple interface operations, thus solving the problems of a large number of difficult-to-find customers in the outbound scenario, difficult achievement of connection rate indicators in the inbound scenario, and high cost and easy loss of employees. During the communication with customers, the intelligent robot performs speech recognition on the real-time voice stream, mines customer intentions, and communicates with customers in the form of real-person voice recording or TTS broadcast or a mixture of TTS / recording broadcast according to the preset sales script template. In the text scenario, interactions are carried out through text. By identifying and screening the content of each call, potential customers can be efficiently and accurately locked, thus achieving the purpose of improving efficiency and reducing costs; Among them, ASR automatic speech recognition technology is a technology that converts human speech signals into text. It has a wide range of applications in many fields, such as intelligent voice assistants, customer service robots, voice input, etc. Through ASR technology, devices can quickly and accurately identify users' voice commands or speech content, providing users with a more convenient and efficient interaction experience. During AI intelligent outbound calls, ASR technology can help accurately identify customers' voice responses, thus realizing a more intelligent outbound call process and better customer interaction effects; The outbound platform adopts a containerized deployment method. The overall function consists of three parts: one is the robot management platform, which is the application side of the business capabilities of the entire system; the second is the intelligent middle platform, which is the core of the system's AI capabilities and business capabilities; the third is the communication module of the system; Robot management platform: Includes data template configuration, campaign configuration management, etc.; Intelligent middle platform: Includes ASR speech recognition, multi-round conversation, intention recognition, TTS speech synthesis, etc., which are the core AI components that make up the robot; Communication module: Includes basic inbound and outbound capabilities, and needs to initiate calls, collect and carry voice streams, generate recordings, store recording files, and provide listening services; Integrating with the call center platform IVR / SBC through the SIP protocol can achieve signaling control and voice streaming media services for telephone voice calls, and can view and listen to historical outbound tasks and recordings; It also provides a backend service for the intelligent outbound call console function, including functions such as task distribution, task scheduling, calling the IVR to issue calls, dialogue interfaces, dialogue log recording, etc.; Among them, it supports identification according to the relay feedback instruction and automatic judgment according to the ringback tone. The recognizable results at least include multiple results such as invalid number, user busy, user rejection, out of service, power off, no answer, etc.; For the case of an invalid number, the throughput of intelligent outbound calls can be improved by setting that the number is no longer called; Recording management supports timely saving of outbound call recordings with a saving duration of 1 month, supports online playback and downloading of recording files, and the recording files need to be synchronized to a third-party recording system.

[0021] TTS pre-synthesis, TTS pre-synthesis supports TTS pre-synthesis of prompt tones, including TTS and recording splicing, TTS sound library selection, etc., to reduce the pressure on the real-time TTS engine and the response time of long-text TTS synthesis; It should be noted that different from the prior art solutions where the historical outbound call data of different intelligent outbound call platforms are isolated from each other, resulting in data islands, which leads to the inability to implement diversified supervision and data processing and utilization of the historical outbound call data of different intelligent outbound call platforms, and further leads to poor outbound call service effects for different outbound call objects; In the embodiments of the present invention, by performing matching analysis on the outbound call numbers of different outbound call objects in the early stage of outbound calls by the outbound call platform, reliable historical outbound call matching processing data support can be provided for implementing targeted outbound call service solutions for different outbound call objects during subsequent outbound calls; The outbound call shared database can be obtained by performing privacy processing and shared combination on the historical outbound call big data of the outbound call subjects to which different intelligent outbound call platforms belong. Here, performing privacy processing on the historical outbound call big data refers to encrypting the outbound call numbers so that the outbound call numbers cannot be visually seen in all links of outbound call data matching processing; If the input outbound call number does not exist in the outbound call shared database, the outbound call object to which the outbound call number belongs is marked as the first outbound call object; If the input outbound call number exists in the outbound call shared database, the outbound call object to which the outbound call number belongs is marked as the second outbound call object, and data analysis is performed on the historical outbound call effect coefficient associated with the second outbound call object, and the subsequent outbound calls of the second outbound call object are dynamically controlled according to the analysis results; Among them, the steps for obtaining the historical outbound call effect coefficients corresponding to different second outbound call objects include: Obtain the total number of historical outbound calls and the total number of historical outbound call answers corresponding to the second outbound call object, as well as the corresponding reply data for each answer, and through the formula Calculate and obtain the historical outbound call effect coefficient HJ corresponding to the second outbound call object; where N0 and N1 are the total number of historical outbound calls and the total number of answered historical outbound calls corresponding to the second outbound call object respectively; α is the answer effect factor; It should be noted that the historical outbound call effect coefficient is used to process and calculate the historical outbound call data of the second outbound call object corresponding to different intelligent outbound call platforms, so as to digitally represent the corresponding historical outbound call effect; Among them, the steps for obtaining the answer effect factor α include: Obtain the reply data corresponding to each answer of the second outbound call object, and classify according to the reply behavior in the reply data; If the second outbound call object does not reply, mark the corresponding historical outbound call as the first outbound call; the first outbound call includes the cases where the outbound call object has not actively answered the outbound call historically, and has actively answered the outbound call but has not communicated, such as not replying and not asking; If the second outbound call object replies, mark the corresponding historical outbound call as the second outbound call, and count the outbound call communication duration corresponding to the second outbound call; the unit of the outbound call communication duration is seconds; Count and analyze the total number of all first outbound calls corresponding to the second outbound call object. If the total number of first outbound calls is the same as its total number of historical inbound calls, update the label of the corresponding second outbound call object to the third outbound call object, and set the corresponding answer effect factor to 0; the third outbound call object indicates that its interest in the intelligent outbound call platform is low; If the total number of first outbound calls is different from its total number of historical inbound calls, then pass the supervision data corresponding to all its first outbound calls and second outbound calls through the formula Calculate and obtain the answer effect factor of the second outbound call object; where N2 is the total number of first outbound calls; NZ is the total number of first outbound calls and second outbound calls; i is different second outbound calls, i = 1, 2, 3,..., n; n is a positive integer, which is the total number of second outbound calls; Ti is the outbound call communication duration corresponding to different second outbound calls; T0 is the standard outbound call communication duration, which can be determined according to the preliminary test data of the intelligent outbound call platform; A is the first reply standard value; B is the second reply standard value; both the first reply standard value and the second reply standard value can be determined according to the operation design requirement data of the intelligent outbound call platform, or can be determined according to the historical outbound call data of the intelligent outbound call platform; Analyze the data of the historical outbound call effect coefficient associated with the second outbound call object through the outbound call supervision recognition model, and output the corresponding outbound call supervision evaluation value WP; Among them, the expression of the outbound call supervision recognition model is ; where HJ0 is the historical outbound call supervision standard value, which can be determined according to the median of the outbound call supervision evaluation values of all second outbound call objects corresponding to the intelligent outbound call platform; An outbound supervision evaluation value of 0 indicates that the historical outbound call effect of the second outbound call object is slightly effective, and the corresponding second outbound call object label is updated to the fourth outbound call object; An outbound supervision evaluation value of 1 indicates that the historical outbound call effect of the third outbound call object is severely effective, and the corresponding second outbound call object label is updated to the fifth outbound call object; In the embodiment of the present invention, by processing and analyzing the outbound supervision data corresponding to different outbound call objects, and dynamically labeling different outbound call objects according to the analysis results, reliable outbound supervision data support can be provided for the targeted outbound calls corresponding to different outbound call objects in the future; According to the labels, the same first outbound call objects, third outbound call objects, fourth outbound call objects, and fifth outbound call objects are sorted and combined respectively to obtain the corresponding first outbound call object set, third outbound call object set, fourth outbound call object set, and fifth outbound call object set, and the first outbound supervision plan is implemented for all the first outbound call objects in the first outbound call object set, and the second outbound supervision plan is implemented for all the fourth outbound call objects in the fourth outbound call object set and all the fifth outbound call objects in the fifth outbound call object set; Among them, the first outbound supervision plan can specifically be to conduct outbound calls and supervise the outbound call process based on the existing conventional outbound call implementation plan; The second external supervision plan can specifically be to conduct outbound calls and supervise the outbound call process based on the existing customized outbound call implementation plan, and the customized outbound call implementation plan is obtained by processing the conventional outbound call implementation plan in combination with the historical call data of the corresponding call object, including but not limited to combining the historical outbound call time period and the specific outbound call conversation content of the call object; the supervision items of the second outbound supervision plan are more than those of the first outbound supervision plan; In the embodiment of the present invention, by monitoring and statistically analyzing the outbound call sharing data associated with different outbound call objects by the outbound call platform in the early stage of outbound call implementation, and dynamically managing the subsequent outbound call implementation of different outbound call objects according to the analysis results, targeted processing and analysis of the historical outbound call status of outbound call objects are realized in the early stage of outbound call, so that targeted outbound supervision plans can be implemented for different outbound call objects in the future, improving the autonomous supervision analysis and classification effect of different outbound call objects in the early stage of outbound call, and the targeted service effect in the outbound call implementation process.

[0022] The outbound call implementation late-stage supervision module is used to monitor and statistically analyze the outbound call process implemented by the outbound call platform for the first outbound call object set, the fourth outbound call object set, and the fifth outbound call object set, and perform data analysis on the monitoring and statistical data corresponding to different outbound call object sets, and dynamically label different outbound call objects according to the analysis results, and upload and share the outbound call sharing data of the labeled different outbound call objects; including: When conducting outbound supervision on all the first outbound objects in the first outbound object set according to the first outbound supervision plan, update the marked first outbound objects to the third outbound objects, the fourth outbound objects, or the fifth outbound objects according to the answering behaviors of the first outbound objects corresponding to unanswered calls, answered but not replied calls, or answered and replied calls. In addition, when conducting outbound supervision on all the fourth outbound objects in the fourth outbound object set and all the fifth outbound objects in the fifth outbound object set according to the second outbound supervision plan, calculate and update the corresponding historical outbound effect coefficients according to the answering behaviors of the fourth outbound objects and the fifth outbound objects corresponding to unanswered calls, answered but not replied calls, or answered and replied calls, and conduct data analysis on the calculated and updated historical outbound effect coefficients, and dynamically mark the corresponding fourth outbound objects and fifth outbound objects according to the analysis results. Upload and share-update the corresponding outbound shared data based on the dynamic marking results of all the first outbound objects, the fourth outbound objects, and the fifth outbound objects.

[0023] It should be noted that by supervising the outbound processes corresponding to the outbound objects with different outbound markings and conducting data processing management, and by performing real-time data update and upload sharing on the outbound shared data corresponding to the outbound objects with different outbound markings, subsequent other intelligent outbound platforms can conduct targeted outbound processing management in a timely manner.

[0024] In the embodiments of the present invention, by supervising the outbound processes implemented for different outbound object sets after preliminary screening and processing, conducting data analysis on the monitored and statistical data corresponding to different outbound object sets, dynamically marking different outbound objects according to the analysis results, and uploading and sharing the outbound shared data for the marked different outbound objects, it realizes targeted traceability and expansion management for the outbound objects with different classification markings in the early stage of calls, and dynamically updates and shares the call markings corresponding to different call objects, improving the management effect of historical call data processing and shared utilization among different intelligent outbound platforms.

[0025] In addition, the formulas involved above are all calculated by removing the dimension and taking their numerical values, and are obtained by a software through simulating a large amount of collected data to get a formula that is closest to the real situation.

[0026] In several embodiments provided by the present invention, it should be understood that the disclosed system can be implemented in other ways. For example, the described embodiments of the invention are only illustrative. For example, the division of modules is only a logical function division, and there can be other division methods in actual implementation.

[0027] The modules described as separate components may or may not be physically separated. The components shown as modules may or may not be physical modules. They may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0028] In addition, in each embodiment of the present invention, the functional modules can be integrated into one processing module, or each module can exist physically alone, or two or more modules can be integrated into one module. The above integrated modules can be implemented in the form of hardware, or in the form of a combination of hardware and software functional modules.

[0029] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and without departing from the basic characteristics of the present invention, the present invention can be implemented in other specific forms.

[0030] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An intelligent outbound calling platform based on AI technology, characterized in that: It includes an outbound call pre-implementation supervision module, which is used to monitor and count the outbound call sharing data associated with different outbound call objects in the early stage of outbound call implementation, process and analyze them, and dynamically manage the subsequent outbound call implementation of different outbound call objects based on the analysis results; The analysis result includes a first outbound call object set, a third outbound call object set, a fourth outbound call object set and a fifth outbound call object set obtained by processing the combination; The outbound call implementation post-supervision module is used to monitor and count the outbound call processes implemented by the outbound call platform for the first outbound call object set, the fourth outbound call object set and the fifth outbound call object set, and to perform data analysis on the monitoring and statistical data corresponding to different outbound call object sets, and to dynamically mark different outbound call objects based on the analysis results, as well as to upload and share outbound call shared data for the marked different outbound call objects.

2. According to claim 1, the intelligent outbound calling platform based on AI technology is characterized in that: Obtain the outbound call numbers corresponding to different outbound call objects in the outbound call platform in the early stage of outbound call implementation, and input the outbound call numbers corresponding to different outbound call objects into the outbound call shared database for traversal retrieval; If the input outbound call number does not exist in the outbound call sharing database, the outbound call object to which the outbound call number belongs is marked as the first outbound call object; If the input outbound call number exists in the outbound call sharing database, the outbound call object to which the outbound call number belongs is marked as the second outbound call object, and data analysis is performed on the historical outbound call effect coefficient associated with the second outbound call object. Based on the analysis results, dynamic control is performed on the implementation of subsequent outbound calls corresponding to the second outbound call object.

3. According to claim 2, the intelligent outbound calling platform based on AI technology is characterized in that: The steps of obtaining the historical outbound call effect coefficients corresponding to different second outbound call objects include: Get the total number of historical outbound calls and the total number of historical outbound calls answered corresponding to the second outbound call object, as well as the corresponding reply data for each answer, and use the formula Calculate and obtain the historical outbound call effect coefficient HJ corresponding to the second outbound call object; where N0 and N1 are the total number of historical outbound calls and the total number of historical outbound call answers corresponding to the second outbound call object, respectively; α is the answering effect factor.

4. According to claim 3, the intelligent outbound calling platform based on AI technology is characterized in that: The steps for obtaining the answering effect factor α include: Obtaining the reply data corresponding to each time the second outbound call recipient answers the call, and classifying the reply data according to the reply behavior in the reply data; If the second outbound call recipient does not respond, the historical outbound call will be marked as the first outbound call; If the second outbound call recipient responds, the historical outbound call will be marked as the second outbound call, and the outbound call communication duration corresponding to the second outbound call will be counted; The total number of all first outbound calls corresponding to the second outbound call object is counted and analyzed. If the total number of first outbound calls is the same as the total number of historical outbound calls, the mark of the second outbound call object is updated to the third outbound call object, and the corresponding answering effect factor is set to 0.

5. According to claim 4, the intelligent outbound calling platform based on AI technology is characterized in that: If the total number of first outbound calls is different from the total number of historical outbound calls, the supervision data corresponding to all first outbound calls and second outbound calls are calculated by the formula Calculate and obtain the answering effect factor of the second outbound call object; where N2 is the total number of first outbound calls; NZ is the total number of first outbound calls and second outbound calls; i is different second outbound calls, i=1, 2, 3, ..., n; n is a positive integer, which is the total number of second outbound calls; Ti is the outbound call communication duration corresponding to different second outbound calls; T0 is the standard outbound call communication duration; A is the first reply standard value; B is the second reply standard value; The historical outbound call effect coefficient associated with the second outbound call object is analyzed through the outbound call supervision identification model, and the corresponding outbound call supervision evaluation value WP is output.

6. The AI-based intelligent outbound calling platform according to claim 5, characterized in that: The expression of the outbound call supervision identification model is: ; In the formula, HJ0 is the historical outbound call supervision standard value; The outbound call supervision evaluation value of 0 indicates that the historical outbound call effect of the second outbound call object is slightly effective, and the corresponding second outbound call object mark is updated to the fourth outbound call object; The outbound call supervision evaluation value of 1 indicates that the historical outbound call effect of the third outbound call object is highly effective, and the corresponding second outbound call object is marked as updated to the fifth outbound call object.

7. The intelligent outbound calling platform based on AI technology according to claim 6 is characterized in that: According to the labels, the same first outbound call object, third outbound call object, fourth outbound call object and fifth outbound call object are sorted and combined respectively to obtain the corresponding first outbound call object set, third outbound call object set, fourth outbound call object set and fifth outbound call object set, and the first outbound call supervision plan is implemented for all first outbound call objects in the first outbound call object set, and the second outbound call supervision plan is implemented for all fourth outbound call objects in the fourth outbound call object set and all fifth outbound call objects in the fifth outbound call object set.

8. The intelligent outbound calling platform based on AI technology according to claim 7 is characterized in that: When performing outbound call supervision of the first outbound call supervision scheme on all first outbound call objects in the first outbound call object set, the marked first outbound call object is updated to the third outbound call object, the fourth outbound call object or the fifth outbound call object according to the answering behavior of the first outbound call object corresponding to the missed call, the answered but not replied, or the answered and replied.

9. The intelligent outbound calling platform based on AI technology according to claim 8, characterized in that: When performing outbound call supervision of the second outbound call supervision scheme on all fourth outbound call objects in the fourth outbound call object set and all fifth outbound call objects in the fifth outbound call object set, the corresponding historical outbound call effect coefficients are calculated and updated according to the answering behaviors of the fourth outbound call objects and the fifth outbound call objects, such as missed calls, answered calls but not replied, or answered calls and replied. Data analysis is performed on the calculated and updated historical outbound call effect coefficients, and the corresponding marked fourth outbound call objects and fifth outbound call objects are dynamically marked according to the analysis results.

10. The intelligent outbound calling platform based on AI technology according to claim 9, characterized in that: The dynamic marking results corresponding to all the first outbound call objects, the fourth outbound call objects and the fifth outbound call objects are used to upload and share and update the corresponding outbound call shared data.

Images