WO2021031985A1 - 智能拨号方法、装置、设备与计算机可读存储介质 - Google Patents

智能拨号方法、装置、设备与计算机可读存储介质 Download PDF

Info

Publication number
WO2021031985A1
WO2021031985A1 PCT/CN2020/109045 CN2020109045W WO2021031985A1 WO 2021031985 A1 WO2021031985 A1 WO 2021031985A1 CN 2020109045 W CN2020109045 W CN 2020109045W WO 2021031985 A1 WO2021031985 A1 WO 2021031985A1
Authority
WO
WIPO (PCT)
Prior art keywords
dialing
case
urging
interval
connection rate
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.)
Ceased
Application number
PCT/CN2020/109045
Other languages
English (en)
French (fr)
Inventor
蔡远航
郑少杰
易剑韬
杨波
范增虎
江旻
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
WeBank Co Ltd
Original Assignee
WeBank Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by WeBank Co Ltd filed Critical WeBank Co Ltd
Publication of WO2021031985A1 publication Critical patent/WO2021031985A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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
    • G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/02—Banking, e.g. interest calculation or account maintenance
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04M—TELEPHONIC COMMUNICATION
    • H04M3/00—Automatic or semi-automatic exchanges
    • H04M3/22—Arrangements for supervision, monitoring or testing
    • H04M3/2218—Call detail recording
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04M—TELEPHONIC COMMUNICATION
    • H04M3/00—Automatic or semi-automatic exchanges
    • H04M3/22—Arrangements for supervision, monitoring or testing
    • H04M3/36—Statistical metering, e.g. recording occasions when traffic exceeds capacity of trunks
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04M—TELEPHONIC COMMUNICATION
    • H04M3/00—Automatic or semi-automatic exchanges
    • H04M3/42—Systems providing special services or facilities to subscribers
    • H04M3/50—Centralised arrangements for answering calls; Centralised arrangements for recording messages for absent or busy subscribers ; Centralised arrangements for recording messages
    • H04M3/51—Centralised call answering arrangements requiring operator intervention, e.g. call or contact centers for telemarketing
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04M—TELEPHONIC COMMUNICATION
    • H04M3/00—Automatic or semi-automatic exchanges
    • H04M3/42—Systems providing special services or facilities to subscribers
    • H04M3/50—Centralised arrangements for answering calls; Centralised arrangements for recording messages for absent or busy subscribers ; Centralised arrangements for recording messages
    • H04M3/527—Centralised call answering arrangements not requiring operator intervention

Definitions

  • This application relates to the field of financial technology (Fintech) technology, in particular to smart dialing methods, devices, equipment and computer-readable storage media.
  • Financial technology Fetech
  • the traditional urging method is to manually make calls to untrustworthy users for urging.
  • urging process is relatively simple, most financial institutions use AI technology (Artificial Intelligence)-based voice dialogue robots to make urging calls.
  • AI technology Artificial Intelligence
  • the robot can achieve no mood swings, user information is strictly confidential and easy to manage.
  • the main purpose of this application is to propose an intelligent dialing method, device, equipment, and computer-readable storage medium, aiming to improve the intelligence of the collection robot, reduce dialing costs, and realize intelligent dialing.
  • an intelligent dialing method which includes the following steps:
  • the dialing instruction is detected, obtaining the urging cases in the current urging queue, and determining the connection rate of each urging case in the dialing interval corresponding to the dialing instruction;
  • dialing is performed according to the dialing plan.
  • the step of detecting the dialing instruction, acquiring the prompting cases in the current prompting queue, and determining the connection rate of each prompting case in the dialing interval corresponding to the dialing instruction includes:
  • Detect the dialing instruction obtain the pending cases and the re- reminders in the current reminder queue, and determine whether the re- reminders meet the preset re- reminder rules;
  • the step of determining the connection rate of each of the urging cases in the dialing interval corresponding to the dialing instruction includes:
  • the feature vector is input into a pre-trained connection rate model to obtain the connection rate of each urging case in the dialing interval corresponding to the dialing instruction.
  • the smart dialing method before the step of inputting the feature vector into the pre-trained connection rate model, the smart dialing method further includes:
  • connection rate model is obtained through training.
  • the step of determining the maximum dial amount of the dialing interval, and formulating a dial plan for the urging case in the dialing interval based on the connection rate and the maximum dial amount includes:
  • a target case whose number is the maximum dial amount is determined, and the target case is used as a dial plan for the dialing interval.
  • the step of determining a target case whose number is the maximum dial amount in the urging case based on the connection rate, and using the target case as the dialing plan for the dialing interval includes :
  • the urging queue includes a re-urging queue
  • the step of dialing according to the dialing plan based on the dialing instruction includes:
  • the target case flow corresponding to the target number is transferred to the manual processing channel.
  • the present application also provides an intelligent dialing device, the intelligent dialing device includes:
  • the determining module is configured to detect the dialing instruction, obtain the urging cases in the current urging queue, and determine the connection rate of each urging case in the dialing interval corresponding to the dialing instruction;
  • a formulation module configured to determine the maximum dial amount of the dialing interval, and formulate a dial plan for the urging case in the dialing interval based on the connection rate and the maximum dial amount;
  • the dialing module is configured to dial according to the dialing plan based on the dialing instruction.
  • the determining module is further configured to:
  • Detect the dialing instruction obtain the pending cases and the re- reminders in the current reminder queue, and determine whether the re- reminders meet the preset re- reminder rules;
  • the determining module is further configured to:
  • the feature vector is input into a pre-trained connection rate model to obtain the connection rate of each urging case in the dialing interval corresponding to the dialing instruction.
  • the smart dialing method before the step of inputting the feature vector into the pre-trained connection rate model, the smart dialing method further includes:
  • connection rate model is obtained through training.
  • the formulation module is further configured to:
  • a target case whose number is the maximum dial amount is determined, and the target case is used as a dial plan for the dialing interval.
  • the formulation module is further configured to:
  • the dialing module is further configured to:
  • the target case flow corresponding to the target number is transferred to the manual processing channel.
  • the present application also provides an intelligent dialing device, the intelligent dialing device comprising: a memory, a processor, and an intelligent dialing program stored in the memory and running on the processor, so When the smart dial program is executed by the processor, the steps of the smart dial method described above are implemented.
  • the present application also provides a computer-readable storage medium having an intelligent dialing program stored on the computer-readable storage medium, and when the intelligent dialing program is executed by a processor, the intelligent dialing as described above Method steps.
  • the smart dialing method proposed in this application detects a dialing instruction, obtains the urging cases in the current urging queue, and determines the connection rate of each urging case in the dialing interval corresponding to the dialing instruction; determines the maximum dialing interval Dialing amount, and formulating a dialing plan for the urging case in the dialing interval based on the connection rate and the maximum dialing amount; and dialing according to the dialing plan based on the dialing instruction.
  • this application formulates a dialing plan for the current dialing range through the connection rate of the urging case and the maximum dialing volume of the current dialing range, and dials according to the dialing plan, so as to flexibly urge the urging case and reduce the dialing cost , Realize smart dialing.
  • FIG. 1 is a schematic diagram of a device structure of a hardware operating environment involved in a solution of an embodiment of the present application
  • Fig. 2 is a schematic flowchart of the first embodiment of the smart dialing method of this application.
  • FIG. 1 is a schematic diagram of the device structure of the hardware operating environment involved in the solution of the embodiment of the present application.
  • the device in the embodiment of this application may be a PC or a server device.
  • the device may include a processor 1001, such as a CPU, a network interface 1004, a user interface 1003, a memory 1005, and a communication bus 1002.
  • the communication bus 1002 is used to implement connection and communication between these components.
  • the user interface 1003 may include a display screen (Display) and an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface.
  • the network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface).
  • the memory 1005 may be a high-speed RAM memory, or a non-volatile memory (non-volatile memory), such as a magnetic disk memory.
  • the memory 1005 may also be a storage device independent of the foregoing processor 1001.
  • FIG. 1 does not constitute a limitation on the device, and may include more or fewer components than those shown in the figure, or a combination of certain components, or different component arrangements.
  • the memory 1005 as a computer storage medium may include an operating system, a network communication module, a user interface module, and a smart dial program.
  • the operating system is a program that manages and controls smart dialing equipment and software resources, and supports the operation of network communication modules, user interface modules, smart dialing programs, and other programs or software;
  • the network communication module is used to manage and control the network interface 1002; users
  • the interface module is used to manage and control the user interface 1003.
  • the smart dialing device calls the smart dialing program stored in the memory 1005 through the processor 1001, and executes the operations in each embodiment of the following smart dialing method.
  • Fig. 2 is a schematic flowchart of a first embodiment of an intelligent dialing method according to this application, and the method includes:
  • Step S10 detecting the dialing instruction, obtaining the prompting cases in the current prompting queue, and determining the connection rate of each prompting case in the dialing interval corresponding to the dialing instruction;
  • Step S20 Determine the maximum dial amount of the dialing interval, and formulate a dial plan for the urging case in the dialing interval based on the connection rate and the maximum dial amount;
  • Step S30 based on the dialing instruction, dialing is performed according to the dialing plan.
  • the smart dialing method of this embodiment is applied to smart dialing devices of financial institutions such as financial institutions or banking systems.
  • the smart dialing device may be a terminal, a robot or a PC device.
  • the smart dialing device is described by taking a dialing robot as an example. For overdue and untrustworthy users, the dialing robot urges them to dial the number.
  • the specific urging process is to dial the phone of the overdue user who has not repaid to determine whether the call is connected, whether the contact person is the overdue user who has not repaid, and whether he promises to repay in the near future Etc., when it is confirmed that the person connected to is the user who has not repaid the payment, and the user promises to repay in the near future, the dialing robot will record the relevant information to complete the urge; if the user cannot promise to repay, the dialing robot forwards the case to The channels are handled manually and urged by the corresponding personnel.
  • the dialing robot detects that the call is not connected, or the person connected is not the user who has not repaid overdue, it will abandon this reminder and generate a corresponding re reminder case so that the dialing robot can re remind it.
  • overdue and non-keeping of promises takes the overdue payment as an example.
  • the smart dialing device of this embodiment that is, the dialing robot, detects the dialing instruction, determines the connection rate of the urging case and the maximum dialing volume of the current dialing interval, formulates the corresponding dialing plan, and performs dialing prompts according to the dialing plan,
  • the dialing robot can flexibly urge and realize intelligent dialing.
  • Step S10 detecting the dialing instruction, obtaining the prompting cases in the current prompting queue, and determining the connection rate of each prompting case in the dialing interval corresponding to the dialing instruction;
  • the dial-up robot monitors the overdue status of the user in real time, and when it detects that a user has overdue and has not kept the promise, obtains the user information of the user, generates the user information into a pending case, and adds the pending case to the pending reminder
  • the case take the overdue payment as an example. As long as there is a new case of overdue payment, the case will be added to the waiting queue.
  • the user information includes name, gender, age, education, occupation, contact number, Information such as the mobile phone attribution, the amount owed, and the number of days in arrears.
  • the dialing robot When the dialing robot detects a dialing instruction, it will go to the prompting queue to obtain a reminder case.
  • the dialing instruction can be an instruction issued at any time by the user or a preset dialing time.
  • the dialing robot detects that the current time is At the time of dialing, the dialing instruction is triggered, and the urging case in the current urging queue is obtained based on the dialing instruction.
  • the dialing instruction is preferably triggered when it is detected that the current time is the dialing time
  • the preset dialing time is preferably ten, corresponding to ten dialing intervals, respectively [08:00, 09:00) , [09:00, 10:00), [10:00, 11:00), [11:00, 12:00), [14:00, 15:00), [15:00, 16:00) , [16:00, 17:00), [17:00, 18:00), [18:00, 19:00), [19:00, 20:00), that is, at 8 o'clock, 9 o'clock, At 10, 11, 12, 14, 15, 16, 17, 18, 19, and 20, the dialing instruction is triggered. This is to take into account the impact of other call times that affect overdue users who have not repaid Possibility of working life.
  • the urging queue includes the waiting queue and the re-urging queue, that is, the urging cases include the waiting cases in the waiting queue and the re-urging cases in the re-urging queue.
  • the dialing robot After obtaining the urging case, the dialing robot further confirms the connection rate of each urging case in the dialing interval corresponding to the dialing instruction. If the current dialing instruction is triggered at 9 o'clock, it will urge the urging case in the queue when it gets 9 o’clock. , And determine the connection rate of each urging case in the dialing interval [09:00, 10:00).
  • the steps of determining the connection rate of each urging case in the current dialing interval include:
  • Step a obtaining user information of the urging case, and determining the feature vector corresponding to each urging case based on the user information;
  • the dialing robot further obtains the user information of the urging case.
  • the user information is preferably gender, age, mobile phone attribution, number of days in arrears and current dialing interval, etc.
  • the feature vector corresponding to each urging case is constructed, specifically a 5-dimensional feature vector.
  • Step b Input the feature vector into a pre-trained connection rate model to obtain the connection rate of each urging case in the dialing interval corresponding to the dialing instruction.
  • the dialing robot inputs the feature vector corresponding to each reminder case into the pre-trained connection rate model, thereby outputting the connection rate of each reminder case in the current dialing interval through the connection rate model.
  • the training process of the connection rate model includes:
  • the dialing robot extracts historical cases with overdue payments in history, and extracts historical user information of historical cases, including gender, age, mobile phone attribution, number of days in arrears, and dialing interval.
  • desensitization extraction can be performed when extracting historical cases.
  • a corresponding historical feature vector is constructed based on historical user information and a preset dialing interval, where the preset dialing interval includes [08:00, 09:00), [09:00, 10:00), [ 10:00, 11:00), [11:00, 12:00), [14:00, 15:00), [15:00, 16:00), [16:00, 17:00), [ 17:00, 18:00), [18:00, 19:00), [19:00, 20:00) ten. It is understandable that in the specific implementation, it can be adjusted according to the actual situation.
  • Each historical case has 10 5-dimensional feature vectors, and the label value of each feature vector is the answering status of the historical case in the corresponding dialing interval. If it is connected, it is represented by 1; if it is not connected, it is represented by 0 means.
  • the dialing robot determines the connection status of each historical case in each dialing interval, and merges the historical cases with the same historical feature vector together to obtain the historical connection rate of each historical feature vector in each dialing interval, such as
  • the historical feature vectors of historical case A and historical case B are the same, but the dialing interval connected by A is [08:00, 09:00), and the dialing interval connected by B is [16:00, 17:00), so ,
  • the historical feature vector of A or B in the interval [08:00, 09:00) has a historical call-through rate of 50%, and the historical call-in rate of the dialing interval in [16:00, 17:00) The pass rate is 50%.
  • a call-through rate model is trained.
  • connection rate model Based on historical feature vectors and historical connection rates, train the connection rate model. Specifically, a logistic regression algorithm can be used to train a logistic regression model as the connection rate model. Since the logistic regression algorithm is a mature algorithm, I won’t repeat it here. In addition, in addition to logistic regression algorithms, mature classification algorithms such as SVM (Support Vector Machine), Naive Bayes, and decision trees can also be used.
  • SVM Small Vector Machine
  • Naive Bayes Naive Bayes
  • decision trees can also be used.
  • Step S20 Determine the maximum dialing amount of the dialing interval, and formulate a dialing plan for the urging case in the dialing interval based on the connection rate and the maximum dialing amount.
  • the dialing robot after determining the connection rate of each urging case in the current dialing interval, the dialing robot further determines the maximum dialing amount of the current dialing interval. It is understandable that in a dialing interval, the number of dialing robots is limited. Therefore, you only need to know the call duration of each urging case to determine the maximum dialing volume of the current dialing interval. Among them, the call time of each urging case can be inferred from historical cases, and the specific can be based on the average of historical cases The call duration is used as the call duration of each urging case. If each call takes 15 minutes, the dialing robot can only make 4 calls in a dialing interval, that is, the maximum dialing amount of the dialing robot in the dialing interval is 4.
  • a corresponding dialing plan is formulated, that is, among all urging cases, the target case with the largest dialing amount is selected Make a call plan. If there are 10 urging cases in the current urging queue and the maximum dialing volume of the current dialing interval is 4, the designated dialing plan will select which 4 urging cases to be urged. In the specific implementation, the one with the higher connection rate will be selected. Urging the case as a dial plan.
  • Step S30 based on the dialing instruction, dialing is performed according to the dialing plan.
  • the dialing robot based on the dialing instruction, performs dialing according to the established dialing plan, that is, when the dialing robot detects the dialing instruction, it specifies the dialing plan and urges the overdue user who has not repaid according to the dialing plan.
  • the urging queue includes a re-urging queue
  • step S30 includes:
  • the dialing robot When the dialing robot detects the dialing instruction, it makes a corresponding dialing plan, reads the target number in the dialing plan, dials the target number, and judges whether the target number is connected.
  • the judging method can confirm whether the human voice is collected within the preset time; or whether the unreachable prompt sound is heard within the preset time, etc. It should be noted that the user who has not repaid the overdue payment is also judged as not connected if he connects to the phone but does not speak; the answering person corresponding to the target number is not overdue and the person is judged as not connected.
  • the specific judgment method can be to ask the other party whether Be the host, and semantically analyze whether the other party’s reply voice contains confirmation sentences or negative sentences.
  • the dialing robot After confirming that the dialing robot determines that the target number is not connected, it marks the target case corresponding to the target number as a re-alerting case and transferring the re-alerted case to the re-alerting queue, so that subsequent dialing robots can re-aler it.
  • the re-alerting queue Included in the urging queue, the specific urging queue includes the waiting queue and the re-urging queue. So far, this dialing urges the end.
  • the dialing robot After confirming that the dialing robot determines that the target number is connected, it further determines whether the user of the target number is performing credit, that is, confirming whether the overdue user has promised to repay, and after confirming that the user has promised to repay in the near future, confirm that the user performs credit; confirm the After the user is unable to promise the repayment in the near future, it is determined that the user cannot fulfill the credit.
  • the dialing robot If the dialing robot confirms that the user of the target number performs credit, it records the corresponding trustworthy date, that is, the date of repayment promised by the user, and marks the target case corresponding to the target number as urged. At this point, the dialing urge ends.
  • the target case flow corresponding to the target number is transferred to the manual processing channel.
  • the dialing robot confirms that the user of the target number cannot fulfill the credit, that is, the user cannot promise to repay in the near future, the target case corresponding to the target number will be transferred to the manual processing channel, and the relevant personnel of the manual processing channel will be manually urged. So far, this dialing urges the end.
  • the urging cases in the current urging queue are acquired, and the connection rate of each urging case in the dialing interval corresponding to the dialing instruction is determined; the maximum dialing amount of the dialing interval is determined, And based on the connection rate and the maximum dialing amount, formulate a dial plan for the urging case in the dialing interval; based on the dialing instruction, dial according to the dialing plan.
  • this application formulates a dialing plan for the current dialing range through the connection rate of the urging case and the maximum dialing volume of the current dialing range, and dials according to the dialing plan, so as to flexibly urge the urging case and reduce the dialing cost , Realize smart dialing.
  • step S10 includes:
  • Step c Detect the dialing instruction, obtain the pending cases and the re-due cases in the current urging queue, and determine whether the re-due cases meet the preset re-due rules;
  • Step d Set the reminder cases and the cases to be reminded that satisfy the preset reminder rules as reminder cases, and determine the connection rate of each reminder case in the dialing interval corresponding to the dial instruction.
  • the reminder cases in this embodiment include pending reminders and reminder cases. Therefore, when determining the connection rate of each reminder case, it is also necessary to consider the reminder rules of reminder cases to avoid frequent reminders of the same reminder case, increase the number of calls, and increase Drive costs.
  • Step c Detect the dialing instruction, obtain the pending cases and the re-due cases in the current urging queue, and determine whether the re-due cases meet the preset re-due rules;
  • the urging queue includes the waiting queue and the re-urging team, that is, the urging case includes the waiting urging case and the re-urging case.
  • the waiting urging case is specifically a case of overdue payment for the first time
  • the re-urging case is specifically a dial urging Two or more cases of overdue payments.
  • the rules for re-urging are specifically that the interval of re-urging of the same urging case must be greater than a preset time.
  • the interval of two urgings of the same urging case must be greater than 1 hour. Understandably, this is to prevent frequent notifications to overdue.
  • the repayment user makes a urging call, which increases the number of calls and increases the cost of dialing.
  • Step d Set the reminder cases and the cases to be reminded that satisfy the preset reminder rules as reminder cases, and determine the connection rate of each reminder case in the dialing interval corresponding to the dial instruction.
  • the dialing robot sets the cases to be reminded and the cases that meet the preset re- reminder rules as reminded cases, and then determines the connection rate of each reminded case in the current dialing interval, that is, it will not satisfy Pre-determined re-due case filtering, and the rest are urging cases.
  • the reminder cases in this embodiment include pending reminders and re reminders.
  • the connection rate of each reminder case only the connection rates of pending reminders and reminder cases that meet the re reminder rules are calculated.
  • the regular re-alerting cases do not need to calculate the call-through rate, and do not need to dial-up urging them, realize smart dialing and reduce dialing costs.
  • step S20 includes:
  • Step e determining the overdue type of each urging case, and determining the call duration of each urging case based on the overdue type;
  • Step f Determine the maximum dial amount of the dialing interval based on the call duration
  • Step g Based on the connection rate, in the urging case, determine a target case whose number is the maximum dialing amount, and use the target case as the dialing interval.
  • the call duration of each urging case is determined by the overdue type category of the urging case, and the maximum dialing amount of the current dialing interval is determined according to the call duration, that is, the maximum dialing amount of the dialing interval is dynamically changed, and the dialing robot is based on each Different urging cases for dialing zones can obtain different maximum dialing amounts for different dialing zones, which reduces dialing costs and improves dialing efficiency.
  • Step e Determine the overdue type of each urging case, and determine the call duration of each urging case based on the overdue type.
  • the dialing robot determines the overdue type of each urging case, and determines the call duration of each urging case according to the overdue type.
  • the overdue type includes: 1 day overdue, overdue
  • overdue types can also specifically include the first time arrears not paid, the second and above outstanding debts, etc., in specific implementation, according to specific circumstances, such as can also be divided according to business types.
  • Step f Determine the maximum dial amount of the dialing interval based on the call duration.
  • the maximum dial amount of the current dialing interval when calculating the maximum dial amount of the current dialing interval, obtain the urging cases of the current urging queue, and determine the call duration of each urging case, and then divide the duration of the dialing interval by the duration of each call to determine that the same can be processed at most.
  • the number of overdue urging cases is determined, and the maximum number of dialing numbers is determined based on the number of overdue urging cases.
  • the dialing robot It can only handle 1 urging case with a call duration of 10 minutes, 2 urging cases with a call duration of 15 minutes and 1 urging case with a 20-minute call duration, and the maximum number of dialing numbers for the current dialing interval is 4.
  • Step g Based on the connection rate, in the urging case, determine a target case whose number is the maximum dialing amount, and use the target case as a dialing plan for the dialing interval.
  • the dialing robot determines the target data with the maximum dialing amount in all the prompting cases according to the connection rate of each prompting case, that is, how much is the maximum dialing amount, that is, selecting the corresponding one from the prompting cases The number of target cases, with the target case as the dial plan.
  • the maximum dialing volume of the current dialing interval is 4, then 1 urging case with a call duration of 10 minutes, 2 urging cases with a call duration of 15 minutes, and a call duration of 20 minutes are selected. There is 1 urging case, and the urging case with a call duration of 20 minutes is the one with the highest connection rate among the 4 urging cases with a call duration of 20 minutes. These four urging cases are the dial plans for the current dialing zone.
  • step g includes:
  • each urging case when formulating the dial plan for the current dialing interval, it is also necessary to obtain the average call duration of the urging case, and the maximum and minimum dialing times of each urging case. Among them, to avoid frequently giving the same overdue Unpaid users dial numbers, increasing unnecessary dialing costs.
  • Each urging case has a maximum number of calls.
  • each urging case has a minimum number of dialing.
  • the dialing robot will determine the connection rate of each reminder case, the maximum dial amount of the current dialing interval, the call duration of each reminder case, the average call duration of all reminder cases, the maximum number of calls and the minimum number of calls for each reminder case The number of times is input and constructed in the objective function and the constraint conditions of the objective function to obtain the dial plan of the current dialing interval.
  • the objective function maximizes the expected connection rate of all overdue users who have not repaid.
  • the specific formula is:
  • the dial duration of the dialing interval should be less than the upper limit of the maximum call duration
  • the number of calls for each urging case should be between the maximum number of calls and the minimum number of calls.
  • 1 ⁇ i ⁇ the number of dialing intervals (as described in the above embodiment, the number of dialing intervals is 10), M means urging a case of heatstroke, 1 ⁇ p ⁇ M, and i and p are integers;
  • api The answering willingness of the overdue payment user corresponding to the p-th case in the i-th time interval, determined by the connection rate model;
  • tp the call duration of the overdue and unpaid user corresponding to the p-th case
  • mp the maximum number of calls made by the debtor corresponding to the p-th case
  • m’p the minimum number of calls made by the debtor corresponding to the p-th case
  • xpi Whether to make a call to the debtor corresponding to the p-th case in the i-th time interval, x ⁇ 0,1, 0 means no call, 1 means call, obtained by the above objective function and constraint conditions;
  • ci the maximum dial amount of the i-th dialing interval
  • the dialing robot needs to call the objective function and constraint conditions to update the optimal dialing of the current dialing zone when dialing in each dialing zone. plan. That is, the objective function and constraint conditions can be used to determine which numbers to dial in which dialing interval.
  • This embodiment integrates multiple parameters and the natural laws that each parameter follows under actual conditions to ensure optimal dialing.
  • the optimal dialing plan for the current dialing interval is formulated through objective functions and constraint conditions to realize intelligent dialing.
  • the application also provides an intelligent dialing device.
  • the intelligent dialing device of this application includes:
  • the determining module is configured to detect the dialing instruction, obtain the urging cases in the current urging queue, and determine the connection rate of each urging case in the dialing interval corresponding to the dialing instruction;
  • a formulation module configured to determine the maximum dial amount of the dialing interval, and formulate a dial plan for the urging case in the dialing interval based on the connection rate and the maximum dial amount;
  • the dialing module is configured to dial according to the dialing plan based on the dialing instruction.
  • the determining module is further configured to:
  • Detect the dialing instruction obtain the pending cases and the re- reminders in the current reminder queue, and determine whether the re- reminders meet the preset re- reminder rules;
  • the determining module is further configured to:
  • the feature vector is input into a pre-trained connection rate model to obtain the connection rate of each urging case in the dialing interval corresponding to the dialing instruction.
  • the smart dialing method further includes:
  • connection rate model is obtained through training.
  • formulation module is also configured as:
  • a target case whose number is the maximum dial amount is determined, and the target case is used as a dial plan for the dialing interval.
  • formulation module is also configured as:
  • dialing module is also configured to:
  • the target case flow corresponding to the target number is transferred to the manual processing channel.
  • the application also provides a computer-readable storage medium.
  • the computer-readable storage medium of the present application stores an intelligent dialing program, and when the intelligent dialing program is executed by a processor, the steps of the intelligent dialing method described above are realized.

Landscapes

  • Engineering & Computer Science (AREA)
  • Signal Processing (AREA)
  • Business, Economics & Management (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Marketing (AREA)
  • Economics (AREA)
  • Development Economics (AREA)
  • Strategic Management (AREA)
  • Technology Law (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Telephonic Communication Services (AREA)

Abstract

本申请公开了一种智能拨号方法，包括：在检测到拨打指令时，获取当前催促队列中的催促案件，并确定各所述催促案件在所述拨打指令对应的拨打区间的接通率；确定所述拨打区间的最大拨号量，并基于所述接通率和所述最大拨号量制定所述催促案件在所述拨打区间的拨打计划；基于所述拨打指令，按照所述拨打计划进行拨号。本申请还公开了一种智能拨号装置、设备和计算机可读存储介质。

Description

智能拨号方法、装置、设备与计算机可读存储介质
本申请要求于2019年08月21日提交中国专利局、申请号为201910776534.6、发明名称为“智能拨号方法、装置、设备与计算机可读存储介质”的中国专利申请的优先权，其全部内容通过引用结合在申请中。
技术领域
本申请涉及金融科技(Fintech)技术领域，尤其涉及智能拨号方法、装置、设备与计算机可读存储介质。
背景技术
近年来，随着金融科技(Fintech)，尤其是互联网金融的不断发展，人工智能技术被引入银行等金融机构的日常服务中。在银行等金融机构的日常服务过程中，常常会遇到用户逾期未守信的情况，如逾期未还款等情况，此时的金融机构需要催促用户尽快履行信用职责，那么如何催促逾期未守信用户履行信用职责？
传统催促方式是由人工拨打电话给未守信用户进行催促，考虑到催促过程较为简单，因此大多金融机构采用基于AI技术(Artificial Intelligence，人工智能)的语音对话机器人来拨打催促电话。相比于人工拨打电话的方式，机器人可以做到无情绪波动，用户信息严格保密，易于管理。
但是，现有的机器人，在拨打催促电话时，遵循的是先到先拨打的逻辑，且只要存在未完成催促的待催案件，机器人就会拨打，这样的拨打逻辑没有考虑未守信用户在当前时刻的接听可能性，出现每个目标用户的平均拨打次数较多的问题，拨打成本较高，显然，现有的机器人在催促逾期未守信用户履行信用职责时，智能性还不够高，无法灵活催促。
技术解决方案
本申请的主要目的在于提出一种智能拨号方法、装置、设备与计算机可读存储介质，旨在提高催收机器人的智能性，减少拨号成本，实现智能拨号。
为实现上述目的，本申请提供一种智能拨号方法，所述智能拨号方法包括如下步骤：
检测到拨打指令，获取当前催促队列中的催促案件，并确定各所述催促案件在所述拨打指令对应的拨打区间的接通率；
确定所述拨打区间的最大拨号量，并基于所述接通率和所述最大拨号量制定所述催促案件在所述拨打区间的拨打计划；
基于所述拨打指令，按照所述拨打计划进行拨号。
在一实施例中，所述检测到拨打指令，获取当前催促队列中的催促案件，并确定各所述催促案件在所述拨打指令对应的拨打区间的接通率的步骤包括：
检测到拨打指令，获取当前催促队列中的待催案件和重催案件，并确定所述重催案件是否满足预设重催规则；
将满足所述预设重催规则的重催案件和所述待催案件设置为催促案件，并确定各所述催促案件在所述拨打指令对应的拨打区间的接通率。
在一实施例中，所述确定各所述催促案件在所述拨打指令对应的拨打区间的接通率的步骤包括：
获取所述催促案件的用户信息，并基于所述用户信息，确定各所述催促案件对应的特征向量；
将所述特征向量输入预先训练好的接通率模型中，以得到各所述催促案件在所述拨打指令对应的拨打区间的接通率。
在一实施例中，所述将所述特征向量输入预先训练好的接通率模型中的步骤之前，所述智能拨号方法还包括：
采集历史案件，并提取所述历史案件的历史用户信息；
基于所述历史用户信息和预设拨打区间，构建对应的历史特征向量，并确定所述历史特征向量在所述预设拨打区间的历史接通率；
基于所述历史特征向量和所述历史接通率，训练得到所述接通率模型。
在一实施例中，所述确定所述拨打区间的最大拨号量，并基于所述接通率和所述最大拨号量制定所述催促案件在所述拨打区间的拨打计划的步骤包括：
确定各所述催促案件的逾期类型，并基于所述逾期类型确定各所述催促案件的通话时长；
基于所述通话时长，确定所述拨打区间的最大拨号量；
基于所述接通率，在所述催促案件中，确定数量为所述最大拨号量的目标案件，以所述目标案件作为所述拨打区间的拨打计划。
在一实施例中，所述基于所述接通率，在所述催促案件中，确定数量为所述最大拨号量的目标案件，以所述目标案件作为所述拨打区间的拨打计划的步骤包括：
确定所述催促案件的平均通话时长，以及各所述催促案件的最大拨打次数和最小拨打次数；
将所述接通率、所述最大拨号量、所述通话时长、所述平均通话时长、所述最大拨打次数和所述最小拨打次数，输入预构建的目标函数和所述目标函数的约束条件中，以得到所述拨打区间对应的目标案件，以所述目标案件作为所述拨打区间的拨打计划。
在一实施例中，所述催促队列包括重催队列，所述基于所述拨打指令，按照所述拨打计划进行拨号的步骤包括：
基于所述拨打指令，读取所述拨打计划中的目标号码，并拨打所述目标号码，确定所述目标号码是否接通；
确定所述目标号码未接通后，将所述目标号码对应的目标案件标记为重催案件，并将所述重催案件流转至所述重催队列；
确定所述目标号码接通后，确定所述目标号码的用户是否履行信用；
确定所述目标号码的用户履行信用后，记录对应的守信日期，并将所述目标号码对应的目标案件标记为已催促；
确定所述目标号码的用户未履行信用后，将所述目标号码对应的目标案件流转至人工处理通道。
此外，为实现上述目的，本申请还提供一种智能拨号装置，所述智能拨号装置包括：
确定模块，配置为检测到拨打指令，获取当前催促队列中的催促案件，并确定各所述催促案件在所述拨打指令对应的拨打区间的接通率；
制定模块，配置为确定所述拨打区间的最大拨号量，并基于所述接通率和所述最大拨号量制定所述催促案件在所述拨打区间的拨打计划；
拨号模块，配置为基于所述拨打指令，按照所述拨打计划进行拨号。
在一实施例中，所述确定模块还配置为：
检测到拨打指令，获取当前催促队列中的待催案件和重催案件，并确定所述重催案件是否满足预设重催规则；
将满足所述预设重催规则的重催案件和所述待催案件设置为催促案件，并确定各所述催促案件在所述拨打指令对应的拨打区间的接通率。
在一实施例中，所述确定模块还配置为：
获取所述催促案件的用户信息，并基于所述用户信息，确定各所述催促案件对应的特征向量；
将所述特征向量输入预先训练好的接通率模型中，以得到各所述催促案件在所述拨打指令对应的拨打区间的接通率。
在一实施例中，所述将所述特征向量输入预先训练好的接通率模型中的步骤之前，所述智能拨号方法还包括：
采集历史案件，并提取所述历史案件的历史用户信息；
基于所述历史用户信息和预设拨打区间，构建对应的历史特征向量，并确定所述历史特征向量在所述预设拨打区间的历史接通率；
基于所述历史特征向量和所述历史接通率，训练得到所述接通率模型。
在一实施例中，所述制定模块还配置为：
确定各所述催促案件的逾期类型，并基于所述逾期类型确定各所述催促案件的通话时长；
基于所述通话时长，确定所述拨打区间的最大拨号量；
基于所述接通率，在所述催促案件中，确定数量为所述最大拨号量的目标案件，以所述目标案件作为所述拨打区间的拨打计划。
在一实施例中，所述制定模块还配置为：
确定所述催促案件的平均通话时长，以及各所述催促案件的最大拨打次数和最小拨打次数；
将所述接通率、所述最大拨号量、所述通话时长、所述平均通话时长、所述最大拨打次数和所述最小拨打次数，输入预构建的目标函数和所述目标函数的约束条件中，以得到所述拨打区间对应的目标案件，以所述目标案件作为所述拨打区间的拨打计划。
在一实施例中，所述拨号模块还配置为：
基于所述拨打指令，读取所述拨打计划中的目标号码，并拨打所述目标号码，确定所述目标号码是否接通；
确定所述目标号码未接通后，将所述目标号码对应的目标案件标记为重催案件，并将所述重催案件流转至所述重催队列；
确定所述目标号码接通后，确定所述目标号码的用户是否履行信用；
确定所述目标号码的用户履行信用后，记录对应的守信日期，并将所述目标号码对应的目标案件标记为已催促；
确定所述目标号码的用户未履行信用后，将所述目标号码对应的目标案件流转至人工处理通道。
此外，为实现上述目的，本申请还提供一种智能拨号设备，所述智能拨号设备包括：存储器、处理器及存储在所述存储器上并可在所述处理器上运行的智能拨号程序，所述智能拨号程序被所述处理器执行时实现如上所述的智能拨号方法的步骤。
此外，为实现上述目的，本申请还提供一种计算机可读存储介质，所述计算机可读存储介质上存储有智能拨号程序，所述智能拨号程序被处理器执行时实现如上所述的智能拨号方法的步骤。
本申请提出的智能拨号方法，检测到拨打指令，获取当前催促队列中的催促案件，并确定各所述催促案件在所述拨打指令对应的拨打区间的接通率；确定所述拨打区间的最大拨号量，并基于所述接通率和所述最大拨号量制定所述催促案件在所述拨打区间的拨打计划；基于所述拨打指令，按照所述拨打计划进行拨号。本申请在检测到拨打指令时，通过催促案件的接通率以及当前拨打区间的最大拨号量，制定当前拨打区间的拨打计划，并按照拨打计划进行拨号，对催促案件进行灵活催促，减少拨号成本，实现智能拨号。
附图说明
图1是本申请实施例方案涉及的硬件运行环境的设备结构示意图；
图2为本申请智能拨号方法第一实施例的流程示意图。
本申请目的的实现、功能特点及优点将结合实施例，参照附图做进一步说明。
本申请的实施方式
应当理解，此处所描述的具体实施例仅仅用以解释本申请，并不用于限定本申请。
如图1所示，图1是本申请实施例方案涉及的硬件运行环境的设备结构示意图。
本申请实施例设备可以是PC机或服务器设备。
如图1所示，该设备可以包括：处理器1001，例如CPU，网络接口1004，用户接口1003，存储器1005，通信总线1002。其中，通信总线1002用于实现这些组件之间的连接通信。用户接口1003可以包括显示屏(Display)、输入单元比如键盘(Keyboard)，可选用户接口1003还可以包括标准的有线接口、无线接口。网络接口1004可选的可以包括标准的有线接口、无线接口(如WI-FI接口)。存储器1005可以是高速RAM存储器，也可以是稳定的存储器(non-volatile memory)，例如磁盘存储器。存储器1005可选的还可以是独立于前述处理器1001的存储装置。
本领域技术人员可以理解，图1中示出的设备结构并不构成对设备的限定，可以包括比图示更多或更少的部件，或者组合某些部件，或者不同的部件布置。
如图1所示，作为一种计算机存储介质的存储器1005中可以包括操作系统、网络通信模块、用户接口模块以及智能拨号程序。
其中，操作系统是管理和控制智能拨号设备与软件资源的程序，支持网络通信模块、用户接口模块、智能拨号程序以及其他程序或软件的运行；网络通信模块用于管理和控制网络接口1002；用户接口模块用于管理和控制用户接口1003。
在图1所示的智能拨号设备中，所述智能拨号设备通过处理器1001调用存储器1005中存储的智能拨号程序，并执行下述智能拨号方法各个实施例中的操作。
基于上述硬件结构，提出本申请智能拨号方法实施例。
参照图2，图2为本申请智能拨号方法第一实施例的流程示意图，所述方法包括：
步骤S10，检测到拨打指令，获取当前催促队列中的催促案件，并确定各所述催促案件在所述拨打指令对应的拨打区间的接通率；
步骤S20，确定所述拨打区间的最大拨号量，并基于所述接通率和所述最大拨号量制定所述催促案件在所述拨打区间的拨打计划；
步骤S30，基于所述拨打指令，按照所述拨打计划进行拨号。
本实施例智能拨号方法运用于理财机构或者银行系统等金融机构的智能拨号设备中，智能拨号设备可以是终端、机器人或者PC设备，为描述方便，智能拨号设备以拨号机器人为例进行描述。对于逾期未守信用户，拨号机器人对其进行拨号催促，具体催促过程为拨打逾期未还款用户的电话，判断电话是否接通，接通人是否为逾期未还款用户本人，是否承诺近期还款等，在确认接通人为逾期未还款用户本人，并且该用户承诺近期还款时，拨号机器人会记录相关信息完成本次催促；若该用户无法承诺还款，则该案件有拨号机器人转发到人工处理通道，并由相应人员进行人工催促。在这过程中，若拨号机器人检测到电话未接通，或者接通人不是逾期未还款用户本人，则放弃本次催促，并生成对应的重催案件，以便拨号机器人对其进行重催。为描述方便，以下逾期未守信以逾期未还款为例进行描述。
本实施例的智能拨号设备，也即拨号机器人，检测到拨打指令，通过确定催促案件的接通率，以及当前拨打区间的最大拨号量，制定对应的拨打计划，并按照拨打计划进行拨号催促，使得拨号机器人能灵活催促，实现智能拨号。
以下将对各个步骤进行详细说明：
步骤S10，检测到拨打指令，获取当前催促队列中的催促案件，并确定各所述催促案件在所述拨打指令对应的拨打区间的接通率；
在本实施例中，拨号机器人实时监测用户的逾期情况，在监测到有用户逾期未守信时，获取该用户的用户信息，将该用户信息生成待催案件，并将该待催案件加入待催案件中，以逾期未还款为例，只要有新的逾期未还款案件发生，该案件即被加入待催队列中，其中，用户信息包括姓名、性别、年龄、学历、职业、联系电话、手机归属地、欠款金额、欠款天数等信息。
需要说明的是，用户信息中的欠款天数随时间变动。
拨号机器人在检测到拨打指令时，会去催促队列中获取催促案件，其中，拨打指令可 以是用户下达的任意时刻下达的指令，也可以是预设的拨打时刻，拨号机器人在检测到当前时刻为拨打时刻时，触发拨打指令，并基于拨打指令获取当前催促队列中的催促案件。在具体实施时，拨打指令优选为在检测到当前时刻为拨打时刻时触发，并且，预设的拨打时刻优为十个，分别对应十个拨打区间，分别为[08:00，09:00)，[09:00，10:00)，[10:00，11:00)，[11:00，12:00)，[14:00，15:00)，[15:00，16:00)，[16:00，17:00)，[17:00，18:00)，[18:00，19:00)，[19:00，20:00)，也即在8点，9点，10点，11点，12点，14点，15点，16点，17点，18点，19点和20点时，触发拨打指令，这是考虑到其他的拨打时刻存在影响逾期未还款用户工作生活的可能性。
需要说明的是，催促队列包括待催队列和重催队列，也即催促案件包括待催队列中的待催案件，以及重催队列中的重催案件。
拨号机器人在获取到催促案件后，进一步确认各催促案件在拨打指令对应的拨打区间的接通率，如当前的拨打指令为9点时所触发，则获取9点时，催促队列中的催促案件，并确定各催促案件在[09:00，10:00)这一拨打区间的接通率。
具体的，确定各催促案件在当前拨打区间的接通率的步骤包括：
步骤a，获取所述催促案件的用户信息，并基于所述用户信息，确定各所述催促案件对应的特征向量；
在该步骤中，拨号机器人在获取到催促案件后，进一步获取催促案件的用户信息，其中，用户信息在具体实施时，优选为性别，年龄，手机归属地，欠款天数和当前拨打区间等，并基于用户信息，构建各催促案件对应的特征向量，具体为5维特征向量。
步骤b，将所述特征向量输入预先训练好的接通率模型中，以得到各所述催促案件在所述拨打指令对应的拨打区间的接通率。
在该步骤中，拨号机器人将各催促案件对应的特征向量输入预先训练好的接通率模型中，从而通过接通率模型，输出各催促案件在当前拨打区间的接通率。
其中，接通率模型的训练过程包括：
采集历史案件，并提取所述历史案件的历史用户信息；
在该步骤中，拨号机器人抽取历史上存在逾期未还款的历史案件，并提取历史案件的历史用户信息，具体包括性别，年龄，手机归属地，欠款天数和拨打区间。其中，在抽取历史案件时，为保证历史用户信息安全，在抽取历史案件时，可进行脱敏抽取。
基于所述历史用户信息和预设拨打区间，构建对应的历史特征向量，并确定所述历史特征向量在所述预设拨打区间的历史接通率；
在该步骤中，基于历史用户信息和预设拨打区间，构建对应的历史特征向量，其中，预设拨打区间包括[08:00，09:00)，[09:00，10:00)，[10:00，11:00)，[11:00，12:00)，[14:00，15:00)，[15:00，16:00)，[16:00，17:00)，[17:00，18:00)，[18:00，19:00)，[19:00，20:00)十个，可以理解的，在具体实施时，可根据实际情况进行调整，此时，每个历史案件都有10个5维特征向量，每个特征向量的标签值即为该历史案件在对应的拨打区间的接听情况，若接通，则以1表示；若未接通，则以0表示。
需要说明的是，若某个历史案件在某一拨打区间接通，如在[14:00，15:00)这一拨打区间接通，则删掉15点之后的拨打区间对应的特征向量。
由于历史可知，因此，拨号机器人确定各历史案件在各拨打区间的接通情况，并将相同历史特征向量的历史案件合并在一起，得到各历史特征向量在各拨打区间的历史接通率，如历史案件A和历史案件B的历史特征向量一致，但A接通的拨打区间为[08:00，09:00)，而B接通的拨打区间为[16:00，17:00)，因此，A或者B的历史特征向量在[08:00，09:00)这一区间的历史接通率为百分之五十，在[16:00，17:00)这一拨打区间的历史接通率为百分之五十。
基于所述历史特征向量和所述历史接通率，训练接通率模型。
基于历史特征向量和历史接通率，训练接通率模型，具体可采用逻辑回归算法训练一 个逻辑回归模型作为接通率模型，由于逻辑回归算法是一个成熟的算法，在此不做过多赘述，此外，除了逻辑回归算法，还可以采用SVM(Support Vector Machine，支持向量机)，朴素贝叶斯，决策树等成熟的分类算法。
步骤S20，确定所述拨打区间的最大拨号量，并基于所述接通率和所述最大拨号量制定所述催促案件在所述拨打区间的拨打计划。
在本实施例中，拨号机器人在确定各催促案件在当前拨打区间的接通率之后，进一步确定当前拨打区间的最大拨号量，可以理解的，在一个拨打区间内，拨号机器人的拨打数量是有限的，因此，只需知道每个催促案件的通话时长，即可确定当前拨打区间的最大拨号量，其中，每个催促案件的通话时间可根据历史案件推测得出，具体可以以历史案件的平均通话时长作为每个催促案件的通话时长，如每一通电话都需要15分钟，那么拨号机器人在一个拨打区间内仅能拨打4通电话，也即拨号机器人在该拨打区间的最大拨号量为4。接着，基于各催促案件在当前拨打区间的接通率，以及当前拨打区间的最大拨号量，制定相应的拨打计划，也即，在所有的催促案件中，挑选出数量为最大拨号量的目标案件制定拨打计划。如当前催促队列中有10个催促案件，当前拨打区间的最大拨号量为4，则指定的拨打计划则为具体挑选哪4个催促案件进行催促，在具体实施时，将接通率较高的催促案件作为拨打计划。
步骤S30，基于所述拨打指令，按照所述拨打计划进行拨号。
在本实施例中，基于拨打指令，拨号机器人按照制定的拨打计划进行拨号，也即拨号机器人在检测到拨打指令时，指定拨打计划，并按照拨打计划对逾期未还款用户进行催促。
具体的，催促队列包括重催队列，步骤S30包括：
基于所述拨打指令，读取所述拨打计划中的目标号码，并拨打所述目标号码，确定所述目标号码是否接通；
拨号机器人在检测到拨打指令时，制定对应的拨打计划，并读取拨打计划中的目标号码，并拨打目标号码，判断目标号码是否接通。判断方式可在预设时间内确认是否采集到人声；或者，在预设时间内是否接听到无法接通的提示音等。需要说明的是，逾期未还款用户接通电话但未发声也判定为未接通；目标号码对应的接听人不是逾期未还款本人也判定为未接通，具体判断方式可以为询问对方是否为机主，并语义分析对方的回复语音是否包含确认语句或者否定语句等。
确定所述目标号码未接通后，将所述目标号码对应的目标案件标记为重催案件，并将所述重催案件流转至所述重催队列；
确定拨号机器人判定目标号码未接通后，将目标号码对应的目标案件标记为重催案件，并将重催案件流转到重催队列，以便后续拨号机器人对其进行重催，其中，重催队列包含在催促队列中，具体催促队列包括待催队列以及重催队列。至此，本次拨号催促结束。
确定所述目标号码接通后，确定所述目标号码的用户是否履行信用；
确定拨号机器人判定目标号码接通后，进一步确定目标号码的用户是否履行信用，也即确认逾期未还款用户是否承诺还款，确定该用户承诺近期还款后，确定该用户履行信用；确定该用户无法承诺近期还款后，确定该用户无法履行信用。
确定所述目标号码的用户履行信用后，记录对应的守信日期，并将所述目标号码对应的目标案件标记为已催促；
拨号机器人若确认目标号码的用户履行信用，则记录对应的守信日期，也即该用户承诺的还款日期，并将目标号码对应的目标案件标记为已催促，至此，本次拨号催促结束。
确定所述目标号码的用户未履行信用后，将所述目标号码对应的目标案件流转至人工处理通道。
拨号机器人若确认目标号码的用户无法履行信用，也即该用户无法承诺近期还款，则将该目标号码对应的目标案件流转至人工处理通道，由人工处理通道的相关人员进行人工催促。至此，本次拨号催促结束。
本实施例在检测到拨打指令时，获取当前催促队列中的催促案件，并确定各所述催促案件在所述拨打指令对应的拨打区间的接通率；确定所述拨打区间的最大拨号量，并基于所述接通率和所述最大拨号量制定所述催促案件在所述拨打区间的拨打计划；基于所述拨打指令，按照所述拨打计划进行拨号。本申请在检测到拨打指令时，通过催促案件的接通率以及当前拨打区间的最大拨号量，制定当前拨打区间的拨打计划，并按照拨打计划进行拨号，对催促案件进行灵活催促，减少拨号成本，实现智能拨号。
进一步地，基于本申请智能拨号方法第一实施例，提出本申请智能拨号方法第二实施例。
智能拨号方法的第二实施例与智能拨号方法的第一实施例的区别在于，步骤S10包括：
步骤c，检测到拨打指令，获取当前催促队列中的待催案件和重催案件，并确定所述重催案件是否满足预设重催规则；
步骤d，将满足所述预设重催规则的重催案件和所述待催案件设置为催促案件，并确定各所述催促案件在所述拨打指令对应的拨打区间的接通率。
本实施例的催促案件包括待催案件和重催案件，因此在确定各催促案件的接通率时，还需要考虑重催案件的重催规则，避免同一催促案件频繁催促，增加拨打次数，增加催促成本。
以下将对各个步骤进行详细说明：
步骤c，检测到拨打指令，获取当前催促队列中的待催案件和重催案件，并确定所述重催案件是否满足预设重催规则；
在本实施例中，催促队列包括待催队列和重催队里，也即催促案件包括待催案件和重催案件，待催案件具体为首次逾期未还款案件，重催案件具体为拨号催促两次及两次以上的逾期未还款案件。拨号机器人在检测到拨打指令时，获取当前催促队列中的待催案件和重催案件，并确定催促队列中的重催案件是否满足预设重催规则。
其中，重催规则具体为同一催促案件再一次催促的间隔时间必须大于一个预设时间，如同一催促案件两次催促的间隔时间必须大于1小时，可以理解的，这是为了防止频繁向逾期未还款用户拨打催促电话，导致拨打次数增多，增加拨号成本。
因此，在获取到催促队列中的催促案件后，先区分哪些是待催案件，哪些是重催案件，并确定重催案件中哪些是刚刚催促过的，也即在前一个小时催促过的，从而确定催促案件中的重催案件哪些满足预设重催规则，哪些不满足预设重催规则。
步骤d，将满足所述预设重催规则的重催案件和所述待催案件设置为催促案件，并确定各所述催促案件在所述拨打指令对应的拨打区间的接通率。
在本实施例中，拨号机器人将待催案件，和满足预设重催规则的重催案件设置为催促案件，再去确定各催促案件在当前拨打区间的接通率，也即，将不满足预设重催规则的重催案件过滤，剩下的即为催促案件。
本实施例的催促案件包括待催案件和重催案件，在确定各催促案件的接通率时，只计算待催案件和满足重催规则的重催案件的接通率，对于不满足重催规则的重催案件不需要计算其接通率，且也不需要对其进行拨号催促，实现智能拨号，降低拨号成本。
进一步地，基于本申请智能拨号方法第一、第二实施例，提出本申请智能拨号方法第三实施例。
智能拨号方法的第三实施例与智能拨号方法的第一、第二实施例的区别在于，步骤S20包括：
步骤e，确定各所述催促案件的逾期类型，并基于所述逾期类型确定各所述催促案件的通话时长；
步骤f，基于所述通话时长，确定所述拨打区间的最大拨号量；
步骤g，基于所述接通率，在所述催促案件中，确定数量为所述最大拨号量的目标案件，以所述目标案件作为所述拨打区间的。
本实施例通过催促案件的逾期类型类确定每个催促案件的通话时长，并根据通话时长确定当前拨打区间的最大拨号量，也即拨打区间的最大拨号量是动态变化的，拨号机器人根据每个拨打区间的催促案件的不同，得到不同拨打区间的不同最大拨号量，在降低拨号成本的同时，提高拨号效率。
以下将对各个步骤进行详细说明：
步骤e，确定各所述催促案件的逾期类型，并基于所述逾期类型确定各所述催促案件的通话时长。
在本实施例中，拨号机器人在获取到催促队列中催促案件之后，确定各催促案件的逾期类型，并根据逾期类型，确定各催促案件的通话时长，其中，逾期类型包括：逾期1天，逾期1-4天，逾期4天以上三种，可以理解的，不同的逾期天数对应的催促电话话术也会存在差异，因此，不同逾期天数的催促案件的通话时长也是不同的，具体的，根据逾期1天的历史案件的平均通话时长，确定逾期1天的催促案件的通话时长，同理，确定逾期1-4天的催促案件的通话时长和逾期4天以上的催促案件的通话时长。
进一步地，逾期类型还可具体包括首次欠款未还，第二次及以上欠款未还等，在具体实施时，根据具体情况进行限定，如还可根据业务类型进行划分。
步骤f，基于所述通话时长，确定所述拨打区间的最大拨号量。
在本实施例中，计算当前拨打区间的最大拨号量时，获取当前催促队列的催促案件，并确定各催促案件的通话时长，再用拨打区间的时长除以各通话时长，确定最多能处理相同逾期类型的催促案件有多少个，再根据各逾期类型的催促案件的数量，确定最大拨号量。如拨打区间为1小时，通话时长为15分钟的催促案件有两个，通话时长为20分钟的有4个，通话时长为10分钟的为1个，则通话时长15分钟的催促案件最多可以处理4个，通话时长20分钟的催促案件最多可以处理3个，通话时长10分钟的催促案件最多可以处理6个，由于通话时长为15分钟和通话时长为10分钟的催促案件没有那么多，拨号机器人仅能处理1个通话时长为10分钟的催促案件，2个通话时长为15分钟的催促案件和1个20分钟的催促案件，得到当前拨打区间的最大拨号量为4个。
步骤g，基于所述接通率，在所述催促案件中，确定数量为所述最大拨号量的目标案件，以所述目标案件作为所述拨打区间的拨打计划。
在本实施例中，拨号机器人根据各催促案件的接通率，在所有的催促案件中，确定数量为最大拨号量的目标数据，也即最大拨号量是多少，即从催促案件中挑选出对应的数量的目标案件，以目标案件作为拨打计划。
具体的，以上述例子为例，当前拨打区间的最大拨号量为4个，则选取通话时长为10分钟的催促案件1个，通话时长为15分钟的催促案件2个，通话时长为20分钟的催促案件1个，且该通话时长为20分钟的催促案件是4个通话时长为20分钟的催促案件中接通率最高的那个，这四个催促案件即为当前拨打区间的拨打计划。
进一步地，步骤g包括：
确定所述催促案件的平均通话时长，以及各所述催促案件的最大拨打次数和最小拨打次数；
在另一实施例中，在制定当前拨打区间的拨打计划时，还需获取所述催促案件的平均通话时长，以及各催促案件的最大拨打次数和最小拨打次数，其中，为避免频繁给同一逾期未还款用户拨号，增加不必要的拨号成本，每个催促案件都有最大拨打次数，同时，为保证每个催促案件都有拨号催促，每个催促案件都有最小拨打次数。
将所述接通率、所述最大拨号量、所述通话时长、所述平均通话时长、所述最大拨打次数和所述最小拨打次数，输入预构建的目标函数和所述目标函数的约束条件中，以得到所述拨打区间对应的目标案件，以所述目标案件作为所述拨打区间的拨打计划。
接着，拨号机器人将确定给的各催促案件的接通率、当前拨打区间的最大拨号量、各催促案件的通话时长、所有的催促案件的平均通话时长、各催促案件的最大拨打次数和最 小拨打次数，输入与构建的目标函数和目标函数的约束条件中，从而得到当前拨打区间的拨打计划。
具体的，目标函数使得所有逾期未还款用户的期望接通率最大化，具体公式为：
Figure PCTCN2020109045-appb-000001
该目标函数的约束条件为：
1、拨打区间的拨号时长应当小于最大拨号量通话时长上限；
2、每个催促案件的拨打次数应当介于最大拨打次数和最小拨打次数之间。
具体公式为：
Figure PCTCN2020109045-appb-000002
最后求解上述具有约束条件的目标函数，得到当前拨打区间的拨打计划。
其中，1≤i≤拨打区间数(如上述实施例所述，拨打区间数为10)，M表示催促案件中暑，1≤p≤M，且i和p属于整数；
api：第p个案件对应的逾期未还款用户在第i个时间区间内的接听意愿，通过接通率模型确定；
tp：第p个案件对应的逾期未还款用户的通话时长；
mp：第p个案件对应的欠款人的最大拨打次数；
m’p：第p个案件对应的欠款人的最小拨打次数；
xpi：在第i个时间区间内是否对第p个案件对应的欠款人拨打电话，x∈0,1，为0表示不拨打，为1表示拨打，由上述目标函数和约束条件求得；
ci：第i个拨打区间的最大拨号量；
Figure PCTCN2020109045-appb-000003
所有催促案件的平均通话时长。
由于在拨号过程中，存在新的待催案件，以及重催案件产生，因此，拨号机器人在每个拨打区间进行拨号时，都需要调用该目标函数和约束条件来更新当前拨打区间的最优拨打计划。也即通过该目标函数和约束条件可得出在哪一拨打区间该拨打哪些号码。
本实施例综合多种参数，以及各参数在实际情况下遵循的自然规律，确保拨号最优，通过目标函数和约束条件制定当前拨打区间的最优拨打计划，实现智能拨号。
本申请还提供一种智能拨号装置。本申请智能拨号装置包括：
确定模块，配置为检测到拨打指令，获取当前催促队列中的催促案件，并确定各所述催促案件在所述拨打指令对应的拨打区间的接通率；
制定模块，配置为确定所述拨打区间的最大拨号量，并基于所述接通率和所述最大拨号量制定所述催促案件在所述拨打区间的拨打计划；
拨号模块，配置为基于所述拨打指令，按照所述拨打计划进行拨号。
进一步地，所述确定模块还配置为：
检测到拨打指令，获取当前催促队列中的待催案件和重催案件，并确定所述重催案件是否满足预设重催规则；
将满足所述预设重催规则的重催案件和所述待催案件设置为催促案件，并确定各所述 催促案件在所述拨打指令对应的拨打区间的接通率。
进一步地，所述确定模块还配置为：
获取所述催促案件的用户信息，并基于所述用户信息，确定各所述催促案件对应的特征向量；
将所述特征向量输入预先训练好的接通率模型中，以得到各所述催促案件在所述拨打指令对应的拨打区间的接通率。
进一步地，所述将所述特征向量输入预先训练好的接通率模型中的步骤之前，所述智能拨号方法还包括：
采集历史案件，并提取所述历史案件的历史用户信息；
基于所述历史用户信息和预设拨打区间，构建对应的历史特征向量，并确定所述历史特征向量在所述预设拨打区间的历史接通率；
基于所述历史特征向量和所述历史接通率，训练得到所述接通率模型。
进一步地，所述制定模块还配置为：
确定各所述催促案件的逾期类型，并基于所述逾期类型确定各所述催促案件的通话时长；
基于所述通话时长，确定所述拨打区间的最大拨号量；
基于所述接通率，在所述催促案件中，确定数量为所述最大拨号量的目标案件，以所述目标案件作为所述拨打区间的拨打计划。
进一步地，所述制定模块还配置为：
确定所述催促案件的平均通话时长，以及各所述催促案件的最大拨打次数和最小拨打次数；
将所述接通率、所述最大拨号量、所述通话时长、所述平均通话时长、所述最大拨打次数和所述最小拨打次数，输入预构建的目标函数和所述目标函数的约束条件中，以得到所述拨打区间对应的目标案件，以所述目标案件作为所述拨打区间的拨打计划。
进一步地，所述拨号模块还配置为：
基于所述拨打指令，读取所述拨打计划中的目标号码，并拨打所述目标号码，确定所述目标号码是否接通；
确定所述目标号码未接通后，将所述目标号码对应的目标案件标记为重催案件，并将所述重催案件流转至所述重催队列；
确定所述目标号码接通后，确定所述目标号码的用户是否履行信用；
确定所述目标号码的用户履行信用后，记录对应的守信日期，并将所述目标号码对应的目标案件标记为已催促；
确定所述目标号码的用户未履行信用后，将所述目标号码对应的目标案件流转至人工处理通道。
本申请还提供一种计算机可读存储介质。
本申请计算机可读存储介质上存储有智能拨号程序，所述智能拨号程序被处理器执行时实现如上所述的智能拨号方法的步骤。
其中，在所述处理器上运行的智能拨号程序被执行时所实现的方法可参照本申请智能拨号方法各个实施例，此处不再赘述。
需要说明的是，在本文中，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者系统不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者系统所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括该要素的过程、方法、物品或者系统中还存在另外的相同要素。
上述本申请实施例序号仅仅为了描述，不代表实施例的优劣。
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到上述实施例方法可 借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件，但很多情况下前者是更佳的实施方式。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品存储在如上所述的一个存储介质(如ROM/RAM、磁碟、光盘)中，包括若干指令用以使得一台终端设备(可以是手机，计算机，服务器，空调器，或者网络设备等)执行本申请各个实施例所述的方法。
以上仅为本申请的优选实施例，并非因此限制本申请的专利范围，凡是利用本申请说明书与附图内容所作的等效结构或等效流程变换，或直接或间接运用在其他相关的技术领域，均同理包括在本申请的专利保护范围内。

Claims (10)

  1. 一种智能拨号方法，所述智能拨号方法包括如下步骤：
    检测到拨打指令，获取当前催促队列中的催促案件，并确定各所述催促案件在所述拨打指令对应的拨打区间的接通率；
    确定所述拨打区间的最大拨号量，并基于所述接通率和所述最大拨号量制定所述催促案件在所述拨打区间的拨打计划；
    基于所述拨打指令，按照所述拨打计划进行拨号。
  2. 如权利要求1所述的智能拨号方法，其中，所述检测到拨打指令，获取当前催促队列中的催促案件，并确定各所述催促案件在所述拨打指令对应的拨打区间的接通率的步骤包括：
    检测到拨打指令，获取当前催促队列中的待催案件和重催案件，并确定所述重催案件是否满足预设重催规则；
    将满足所述预设重催规则的重催案件和所述待催案件设置为催促案件，并确定各所述催促案件在所述拨打指令对应的拨打区间的接通率。
  3. 如权利要求2所述的智能拨号方法，其中，所述确定各所述催促案件在所述拨打指令对应的拨打区间的接通率的步骤包括：
    获取所述催促案件的用户信息，并基于所述用户信息，确定各所述催促案件对应的特征向量；
    将所述特征向量输入预先训练好的接通率模型中，以得到各所述催促案件在所述拨打指令对应的拨打区间的接通率。
  4. 如权利要求3所述的智能拨号方法，其中，所述将所述特征向量输入预先训练好的接通率模型中的步骤之前，所述智能拨号方法还包括：
    采集历史案件，并提取所述历史案件的历史用户信息；
    基于所述历史用户信息和预设拨打区间，构建对应的历史特征向量，并确定所述历史特征向量在所述预设拨打区间的历史接通率；
    基于所述历史特征向量和所述历史接通率，训练得到所述接通率模型。
  5. 如权利要求1所述的智能拨号方法，其中，所述确定所述拨打区间的最大拨号量，并基于所述接通率和所述最大拨号量制定所述催促案件在所述拨打区间的拨打计划的步骤包括：
    确定各所述催促案件的逾期类型，并基于所述逾期类型确定各所述催促案件的通话时长；
    基于所述通话时长，确定所述拨打区间的最大拨号量；
    基于所述接通率，在所述催促案件中，确定数量为所述最大拨号量的目标案件，以所述目标案件作为所述拨打区间的拨打计划。
  6. 如权利要求5所述的智能拨号方法，其中，所述基于所述接通率，在所述催促案件中，确定数量为所述最大拨号量的目标案件，以所述目标案件作为所述拨打区间的拨打计划的步骤包括：
    确定所述催促案件的平均通话时长，以及各所述催促案件的最大拨打次数和最小拨打次数；
    将所述接通率、所述最大拨号量、所述通话时长、所述平均通话时长、所述最大拨打 次数和所述最小拨打次数，输入预构建的目标函数和所述目标函数的约束条件中，以得到所述拨打区间对应的目标案件，以所述目标案件作为所述拨打区间的拨打计划。
  7. 如权利要求1-6任意一项所述的智能拨号方法，其中，所述催促队列包括重催队列，所述基于所述拨打指令，按照所述拨打计划进行拨号的步骤包括：
    基于所述拨打指令，读取所述拨打计划中的目标号码，并拨打所述目标号码，确定所述目标号码是否接通；
    确定所述目标号码未接通后，所述目标号码对应的目标案件标记为重催案件，并将所述重催案件流转至所述重催队列；
    确定所述目标号码接通后，确定所述目标号码的用户是否履行信用；
    确定所述目标号码的用户履行信用后，记录对应的守信日期，并将所述目标号码对应的目标案件标记为已催促；
    确定所述目标号码的用户未履行信用后，将所述目标号码对应的目标案件流转至人工处理通道。
  8. 一种智能拨号装置，其中，所述智能拨号装置包括：
    确定模块，配置为确定检测到拨打指令后，获取当前催促队列中的催促案件，并确定各所述催促案件在所述拨打指令对应的拨打区间的接通率；
    制定模块，配置为确定所述拨打区间的最大拨号量，并基于所述接通率和所述最大拨号量制定所述催促案件在所述拨打区间的拨打计划；
    拨号模块，配置为基于所述拨打指令，按照所述拨打计划进行拨号。
  9. 一种智能拨号设备，其中，所述智能拨号设备包括：存储器、处理器及存储在所述存储器上并可在所述处理器上运行的智能拨号程序，所述智能拨号程序被所述处理器执行时实现如权利要求1至7中任一项所述的智能拨号方法的步骤。
  10. 一种计算机可读存储介质，其中，所述计算机可读存储介质上存储有智能拨号程序，所述智能拨号程序被处理器执行时实现如权利要求1至7中任一项所述的智能拨号方法的步骤。
PCT/CN2020/109045 2019-08-21 2020-08-14 智能拨号方法、装置、设备与计算机可读存储介质 Ceased WO2021031985A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910776534.6A CN110475033B (zh) 2019-08-21 2019-08-21 智能拨号方法、装置、设备与计算机可读存储介质
CN201910776534.6 2019-08-21

Publications (1)

Publication Number Publication Date
WO2021031985A1 true WO2021031985A1 (zh) 2021-02-25

Family

ID=68513487

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/109045 Ceased WO2021031985A1 (zh) 2019-08-21 2020-08-14 智能拨号方法、装置、设备与计算机可读存储介质

Country Status (2)

Country Link
CN (1) CN110475033B (zh)
WO (1) WO2021031985A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110475033B (zh) * 2019-08-21 2021-08-03 深圳前海微众银行股份有限公司 智能拨号方法、装置、设备与计算机可读存储介质
CN111091460B (zh) * 2019-11-22 2024-07-02 深圳前海微众银行股份有限公司 一种数据处理方法及装置
CN112734243A (zh) * 2021-01-13 2021-04-30 零犀(北京)科技有限公司 一种确定还款案件的贡献度的方法和智能终端

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016036144A1 (ko) * 2014-09-04 2016-03-10 (주)네오메카 그룹화된 상담단말기와 콜데이타를 이용한 아웃바운드 시스템
CN109859032A (zh) * 2019-01-22 2019-06-07 深圳壹账通智能科技有限公司 基于智能语音的账款催收方法、装置、设备和存储介质
CN110139288A (zh) * 2019-04-08 2019-08-16 深圳众赢维融科技有限公司 一种网络通话方法、装置、系统和记录介质
CN110378575A (zh) * 2019-06-28 2019-10-25 上海上湖信息技术有限公司 逾期事件回款催收方法及装置、计算机可读存储介质
CN110475033A (zh) * 2019-08-21 2019-11-19 深圳前海微众银行股份有限公司 智能拨号方法、装置、设备与计算机可读存储介质
CN111435482A (zh) * 2019-01-14 2020-07-21 顺丰科技有限公司 一种外呼模型的构建方法、外呼方法、装置和存储介质

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5521965A (en) * 1992-08-14 1996-05-28 Bell Atlantic Network Services, Inc. Apparatus and method for handling busy calls in telephone network
US8483383B2 (en) * 2007-03-02 2013-07-09 Aspect Software, Inc. Method of scheduling calls
CN102938802B (zh) * 2012-11-15 2015-06-17 上海过河兵电子商务有限公司 智能拨号控制系统及其控制方法
US9332118B2 (en) * 2014-03-25 2016-05-03 CallerReady LLC System and method for call distribution
CN105898086A (zh) * 2016-03-31 2016-08-24 广东亿迅科技有限公司 一种预测外呼的方法
CN108650426B (zh) * 2018-02-08 2021-01-26 中移在线服务有限公司 一种呼叫中心预测外呼方法及系统
CN109600522A (zh) * 2018-12-10 2019-04-09 中国银行股份有限公司 一种外呼控制方法及装置

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016036144A1 (ko) * 2014-09-04 2016-03-10 (주)네오메카 그룹화된 상담단말기와 콜데이타를 이용한 아웃바운드 시스템
CN111435482A (zh) * 2019-01-14 2020-07-21 顺丰科技有限公司 一种外呼模型的构建方法、外呼方法、装置和存储介质
CN109859032A (zh) * 2019-01-22 2019-06-07 深圳壹账通智能科技有限公司 基于智能语音的账款催收方法、装置、设备和存储介质
CN110139288A (zh) * 2019-04-08 2019-08-16 深圳众赢维融科技有限公司 一种网络通话方法、装置、系统和记录介质
CN110378575A (zh) * 2019-06-28 2019-10-25 上海上湖信息技术有限公司 逾期事件回款催收方法及装置、计算机可读存储介质
CN110475033A (zh) * 2019-08-21 2019-11-19 深圳前海微众银行股份有限公司 智能拨号方法、装置、设备与计算机可读存储介质

Also Published As

Publication number Publication date
CN110475033A (zh) 2019-11-19
CN110475033B (zh) 2021-08-03

Similar Documents

Publication Publication Date Title
US12236936B2 (en) Systems and methods for providing automated natural language dialogue with customers
US11032421B2 (en) Systems for transitioning telephony-based and in-person servicing interactions to and from an artificial intelligence (AI) chat session
US12452197B2 (en) System and method for proactive intervention to reduce high cost channel usage
EP3511887A1 (en) Automated chat assistant for providing interactive data using npl - natural language processing - system and method
US8527413B2 (en) Method and system for mobile bill presentment and payment messaging and marketing
US12081699B2 (en) Systems for transitioning telephony-based and in-person servicing interactions to and from an artificial intelligence (AI) chat session
US11586618B2 (en) Query parsing from natural language questions supported by captured subject matter knowledge
US12067138B2 (en) Systems and methods for linking a screen capture to a user support session
CN110310020B (zh) 基于数据分析的项目方案管理方法、相关装置及存储介质
CN108711085A (zh) 一种交易请求的响应方法及其设备
WO2021031985A1 (zh) 智能拨号方法、装置、设备与计算机可读存储介质
CN107944851B (zh) 群组资金账户的管理方法、装置及系统
KR20180042763A (ko) 채팅형 금융 로봇 및 이를 이용한 금융 서비스 제공방법
US10171671B1 (en) Customer service call routing and/or scheduling system and method
CN107862525A (zh) 动态调整移动支付额度的方法、装置和系统
TW202117584A (zh) 基於自然語言處理的智能對話管理方法及系統
US10930391B2 (en) Device for reducing fraud, waste, and abuse in the ordering and performance of medical testing and methods for using the same
CN114723455B (zh) 业务处理方法、装置、电子设备和存储介质
CN107563588A (zh) 一种个人信用的获取方法和获取系统
KR102132936B1 (ko) 고객 맞춤형 금융거래 서비스 운영 방법
CN120634540A (zh) 一种基于医疗信息的退费处理方法、装置及自动退费机
US9832316B1 (en) Customer service model-based call routing and/or scheduling system and method
CN109408558A (zh) 一种自助设备数据智能分析与应用系统
US12619756B2 (en) Systems and methods for linking a screen capture to a user support session
CN112164450A (zh) 一种医疗服务的签约方法和系统

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20854000

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20854000

Country of ref document: EP

Kind code of ref document: A1