Online data outbound call test processing method and device and electronic equipment

Abstract

This invention provides an online data outbound call testing processing method, apparatus, and electronic device. The method includes: determining a target test scenario for the outbound call system under test; acquiring a target simulated voice stream matching the target test scenario; establishing a test call for the outbound call system under test matching the target test scenario; in response to triggering a preset trigger condition based on the voice stream of the outbound call system under test being monitored during the test call, injecting the target simulated voice stream into the voice stream of the outbound call system under test to obtain an interactive call voice stream; and performing outbound call testing on the outbound call system under test based on the interactive call voice stream. This achieves low-cost and high-fidelity reproduction of complex online call scenarios, thereby enabling low-cost and high-fidelity outbound call testing of the outbound call system under test.

Description

Technical Field

[0001] This invention relates to the field of outbound call testing technology, and in particular to an online data outbound call testing processing method, apparatus, and electronic device. Background Technology

[0002] During the development and testing phase of outbound calling systems, especially intelligent AI outbound calling systems, it is necessary to test the outbound calling performance of the outbound calling system.

[0003] According to relevant technologies, outbound call testing is currently often conducted manually or using scripts. However, while manual outbound call testing offers better flexibility and realism, it suffers from low execution efficiency and struggles to accurately reproduce complex online interaction scenarios. Script-based outbound call testing, on the other hand, improves execution efficiency, but the simulated dialogue flow is rather rigid and cannot realistically replicate the continuous, real-time, and dynamic characteristics of online call audio streams.

[0004] Therefore, finding a low-cost, high-fidelity outbound call testing method that can reproduce complex online call scenarios has become a current research hotspot. Summary of the Invention

[0005] This invention provides an online data outbound call testing processing method, device, and electronic device, which enables low-cost and high-fidelity reproduction of complex online call scenarios, thereby completing outbound call testing of the outbound call system under test in a low-cost and high-fidelity manner.

[0006] This invention provides an online data outbound call testing processing method, the method comprising: determining a target test scenario for an outbound call system to be tested; acquiring a target simulated voice stream matching the target test scenario, wherein the target simulated voice stream is used to simulate a client voice stream interacting with the outbound call system to be tested under the target test scenario; establishing a test call for the outbound call system to be tested that matches the target test scenario; in response to triggering a preset trigger condition based on the voice stream of the outbound call system to be tested being monitored in the test call, injecting the target simulated voice stream into the voice stream of the outbound call system to be tested to obtain an interactive call voice stream; and performing outbound call testing on the outbound call system to be tested based on the interactive call voice stream.

[0007] According to an online data outbound call testing processing method provided by the present invention, before obtaining a target simulated voice stream matching the target test scenario based on the target test scenario, the method further includes: calling a pre-maintained test data management library, wherein the test data management library includes the correspondence between different test scenarios and different simulated voice streams; obtaining a target simulated voice stream matching the target test scenario based on the target test scenario includes: obtaining a target simulated voice stream matching the target test scenario based on the correspondence between the target test scenario and the test data management library.

[0008] According to an online data outbound call testing processing method provided by the present invention, the simulated voice stream is obtained in the following manner: acquiring the historical interactive call voice stream between the outbound call system and the client; pruning the historical interactive call voice stream from the call connection time to obtain the pruned historical interactive call voice stream; extracting the client voice stream under the pruned historical interactive call voice stream based on the pruned historical interactive call voice stream; and using the client voice stream under the pruned historical interactive call voice stream as the simulated voice stream.

[0009] According to an online data outbound call testing processing method provided by the present invention, the simulated voice stream is obtained in the following manner: based on pre-set interactive response rules, a simulated interactive call voice stream of client-outbound system interaction in a test scenario is recorded, wherein the interactive response rules define client response scripts and response time intervals of client response scripts, and the client response scripts are used to characterize the client's response rules under preset trigger conditions triggered by the voice stream on the outbound system side; based on the simulated interactive call voice stream, the client voice stream under the simulated interactive call voice stream is extracted; and the client voice stream under the simulated interactive call voice stream is used as the simulated voice stream.

[0010] According to an online data outbound call testing method provided by the present invention, before performing outbound call testing on the outbound call system under test based on the interactive call voice stream, the method further includes: determining a standard interactive call voice stream under the target test scenario, wherein the standard interactive call voice stream defines a standard response script and a standard response time interval of the standard response script, and the standard response script is a response reply script template of the outbound call system under test in response to the target simulated voice stream; the outbound call testing on the outbound call system under test based on the interactive call voice stream includes: extracting the actual response script of the outbound call system under test in response to the target simulated voice stream based on the interactive call voice stream, and the actual response time interval of the actual response script; and performing outbound call testing on the outbound call system under test based on the standard response script, the actual response script, the standard response time interval, and the actual response time interval.

[0011] According to an online data outbound call testing method provided by the present invention, the outbound call testing of the outbound call system under test based on the standard response script, the actual response script, the standard response time interval, and the actual response time interval includes: determining that the outbound call test result of the outbound call system under test meets the test requirements if the standard response script and the actual response script match and the actual response time interval is less than or equal to the standard response time interval; and determining that the outbound call test result of the outbound call system under test does not meet the test requirements if the standard response script and the actual response script do not match and / or the actual response time interval is greater than the standard response time interval.

[0012] This invention also provides an online data outbound call testing processing device, the device comprising: a determining module, configured to determine a target test scenario for the outbound call system under test; an acquiring module, configured to acquire a target simulated voice stream matching the target test scenario, wherein the target simulated voice stream is used to simulate a client voice stream interacting with the outbound call system under test in the target test scenario; an establishing module, configured to establish a test call for the outbound call system under test matching the target test scenario, and in response to triggering a preset trigger condition based on the voice stream of the outbound call system under test being monitored in the test call, injecting the target simulated voice stream into the voice stream of the outbound call system under test to obtain an interactive call voice stream; and a testing module, configured to perform outbound call testing on the outbound call system under test based on the interactive call voice stream.

[0013] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the online data outbound call test processing method as described above.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the online data outbound call test processing method as described above.

[0015] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the online data outbound call test processing method as described above.

[0016] This invention provides an online data outbound call testing processing method, apparatus, and electronic device. The method includes: determining a target test scenario for the outbound call system under test; acquiring a target simulated voice stream matching the target test scenario; establishing a test call for the outbound call system under test matching the target test scenario; in response to triggering a preset trigger condition based on the voice stream of the outbound call system under test being monitored during the test call, injecting the target simulated voice stream into the voice stream of the outbound call system under test to obtain an interactive call voice stream; and performing outbound call testing on the outbound call system under test based on the interactive call voice stream. This achieves low-cost and high-fidelity reproduction of complex online call scenarios, thereby enabling low-cost and high-fidelity outbound call testing of the outbound call system under test. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating the online data outbound call testing processing method provided by the present invention.

[0019] Figure 2 This is a schematic diagram of the process for obtaining an analog speech stream provided by the present invention.

[0020] Figure 3 This is a schematic diagram of the process for performing outbound call testing on the outbound call system under test based on the interactive call voice stream provided by the present invention.

[0021] Figure 4 This is a schematic diagram illustrating an application scenario of the online data outbound call testing and processing method provided by the present invention.

[0022] Figure 5 This is a schematic diagram of the online data outbound call testing and processing device provided by the present invention.

[0023] Figure 6This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0025] The online data outbound call testing processing method provided by this invention can collect complete call recordings (corresponding to the target simulated voice stream mentioned above) from the production environment after removing the ringing phase as test data material. Through a media server in the test environment, the recordings are used as the voice stream of the simulated user and accurately injected back into the call channel established with the outbound call system under test, thereby restoring a high-fidelity and controllable online interaction scenario and realizing low-cost, high-fidelity outbound call testing of the outbound call system under test.

[0026] Figure 1 This invention provides an online data outbound call testing and processing method, which will be discussed below in conjunction with... Figure 1 The process of the online data outbound call testing method provided by this invention is described.

[0027] In an exemplary embodiment of the present invention, combined with Figure 1 As can be seen, the online data outbound call test processing method can be applied to the automated test platform of the intelligent AI outbound call system, which includes steps 110 to 140. Each step will be described below.

[0028] In step 110, the target test scenario for the outbound calling system to be tested is determined.

[0029] In one embodiment, test engineers or automated testing platforms can determine a target test scenario for the outbound calling system under test based on business needs. This target test scenario could be: "After hearing the robot announce the keyword 'annual interest rate of X percent,' the customer immediately interrupts and expresses 'that's too high.'" This scenario represents a typical, time-sensitive critical interaction behavior in online calls, used to verify whether the internal logic of the system under test (e.g., whether it can interrupt the current script and transition to a reassurance or discount explanation process) is correct when precisely interrupted by a customer.

[0030] In step 120, based on the target test scenario, a target simulated voice stream matching the target test scenario is obtained. The target simulated voice stream is used to simulate the client voice stream that interacts with the outbound call system under the target test scenario.

[0031] In one embodiment, a target simulated voice stream matching a defined target test scenario can be obtained from a pre-built test data management library. This target simulated voice stream simulates the client voice stream interacting with the outbound calling system under test in the target test scenario, and includes all randomness, critical states, and timing information.

[0032] In one example, for the aforementioned target test scenario where "after hearing the robot announce the keyword 'annual interest rate of X percent,' the customer immediately interrupts and says 'That's too high,'" the target simulated voice stream could be an audio file extracted from a real production environment call recording. The audio file could contain a real customer saying "That's too high" in a slightly hurried tone. This recording fully preserves the original call's intonation, background noise, and precise duration characteristics, enabling a high-fidelity simulation of the client's voice stream interacting with the outbound calling system under test in this test scenario.

[0033] In step 130, a test call is established for the outbound calling system under test that matches the target test scenario. In response to the triggering of a preset trigger condition based on the voice stream of the outbound calling system under test being monitored in the test call, the target simulated voice stream is injected into the voice stream of the outbound calling system under test to obtain an interactive call voice stream.

[0034] In one embodiment, a test call command (SIP INVITE) can be sent to the media server FreeSWITCH, initiating a test call to the system under test. After the system under test responds, a real-time communication channel is established between the two. The test call can be monitored in real time.

[0035] When the preset keyword "annual interest rate is X percent" is detected in the voice stream of the outbound calling system under test, the preset trigger condition is determined to be met. A high-priority command can be immediately sent to the media server (FreeSWITCH). This command instructs the media server to inject a pre-loaded target analog voice stream containing the audio "too high" into the current call media stream with extremely low latency (e.g., a target latency of 300 milliseconds). FreeSWITCH executes the command, mixing or directly switching the simulated customer voice stream of "too high" with the voice stream being played by the outbound calling system under test, thereby generating an interactive call voice stream containing "robot script" and "precise customer interruption".

[0036] In step 140, an outbound call test is performed on the outbound call system under test based on the interactive call voice stream.

[0037] In another embodiment, throughout the entire test call, the complete interactive call audio stream generated by the interaction between the two parties (including the robot's previously broadcast script, the injected "too high" interruption script, and the system's subsequent reassuring script) is recorded and saved by a media server to form a test result recording (corresponding to the interactive call audio stream). Furthermore, the outbound call system under test can be tested based on the test result recording.

[0038] In one example, test recordings can be analyzed automatically. Speech recognition technology transcribes the recordings into text, which is then compared to the expected results. For instance, it determines whether the outbound calling system under test accurately played the pre-set "preferential rate" script after the customer interrupted with "too high," instead of continuing with the previous price quote. The comparison results will automatically generate a test report, clearly indicating whether the outbound calling system under test passed the test.

[0039] This embodiment successfully transforms real-world, time-sensitive, and complex interactive scenarios into automated, repeatable, and high-fidelity test cases, effectively solving the core technical problems of low fidelity, high cost, and non-repeatability in existing outbound call system testing technologies.

[0040] This invention provides an online data outbound call testing processing method, the method comprising: determining the target test scenario of the outbound call system under test; acquiring a target simulated voice stream matching the target test scenario; establishing a test call for the outbound call system under test matching the target test scenario; in response to triggering a preset trigger condition based on the voice stream of the outbound call system under test being monitored in the test call, injecting the target simulated voice stream into the voice stream of the outbound call system under test to obtain an interactive call voice stream; and performing outbound call testing on the outbound call system under test based on the interactive call voice stream. This method enables low-cost, high-fidelity reproduction of complex online call scenarios, thereby achieving low-cost, high-fidelity outbound call testing of the outbound call system under test.

[0041] In yet another exemplary embodiment of the present invention, continuing with the description of the aforementioned embodiments, before obtaining the target simulated speech stream matching the target test scenario based on the target test scenario, the method further includes: Call the pre-maintained test data management library, which includes the correspondence between different test scenarios and different simulated voice streams; Among them, obtaining the target simulated speech stream that matches the target test scenario can be achieved in the following ways: Based on the correspondence between the target test scenario and the test data management library, a target simulated speech stream matching the target test scenario is obtained.

[0042] In one embodiment, a pre-maintained test data management library can be invoked first. This test data management library is a structured database or file indexing system that pre-stores the mapping relationships between different test scenarios and different simulated voice streams. Specifically, for each defined test scenario, such as "precise interruption scenario," "silence timeout scenario," "emotional excitement scenario," and "key selection scenario," the library can associate one or more matching simulated voice stream file paths and metadata.

[0043] In another embodiment, the determined target test scenario (or its scenario identifier) ​​is used as the query key to perform a matching retrieval in the invoked test data management database. Based on the pre-established correspondence in the database, the record uniquely associated with the target test scenario can be quickly located, and the corresponding analog speech stream file can be extracted from it as the target analog speech stream used in this test.

[0044] Figure 2 This is a schematic diagram of the process for obtaining an analog speech stream provided by the present invention.

[0045] The following will combine Figure 2 The process of obtaining an analog speech stream provided by the present invention will be described.

[0046] In an exemplary embodiment of the present invention, combined with Figure 2 As can be seen, obtaining the analog speech stream may include steps 210 to 240, and each step will be described below.

[0047] In step 210, the historical voice streams of the outbound call system and the client are obtained.

[0048] In one embodiment, during production environment operations, such as when a live online outbound calling system is running, the system is equipped with a recording function. Once a live outbound call is successfully connected, two-way recording can be automatically initiated, fully recording the entire call process from the customer saying "hello" until either party hangs up. This recording file is a historical interactive call audio stream. This audio stream is a dual-channel or mixed-channel file, which simultaneously contains the voice of the outbound calling system (robot script) and the voice of the client (real customer), as well as all original information such as background noise, pauses, and interruptions during the call.

[0049] In step 220, the historical interactive call voice stream is clipped starting from the moment the call is connected, resulting in a clipped historical interactive call voice stream.

[0050] In one embodiment, since production environment recordings may contain ringing tones or other useless segments before a call is connected, the call connection moment of the historical interactive call voice stream can be located first, i.e., the instant the customer actually answers the call, typically at the moment the ringing stops and a human voice or ambient sound appears. Then, starting from this connection moment, the original recording is trimmed, discarding all data before connection, such as ringing tones and operator prompts, while retaining the complete content from the moment the call is connected to the end of the call, thereby obtaining a trimmed historical interactive call voice stream.

[0051] It should be noted that the starting point of the truncated historical interactive call voice stream should be set at the instant the call is connected, excluding the "ringing wait period" after the outgoing call is dialed, and only the complete two-way interactive voice from the moment the customer answers the phone and says "hello" until the call is disconnected. The truncated historical interactive call voice stream file (such as WAV format) can include metadata such as timestamps and caller / callee numbers, and is stored in a recording repository. These recordings are a complete digital mapping of the real online scenario, containing all randomness, critical states, and timing information.

[0052] In step 230, the client voice stream under the clipped historical interactive call voice stream is extracted based on the clipped historical interactive call voice stream.

[0053] In step 240, the client voice stream under the clipped historical interactive call voice stream is used as the analog voice stream.

[0054] In one embodiment, the truncated historical interactive call audio stream is still a mixed stream (or a two-channel stream) containing the voices of both parties. During application, speech separation techniques can be employed, such as voiceprint separation based on timbre differences, channel extraction based on two channels, or separation algorithms based on VAD and energy ratio, to extract the client-specific audio stream from the mixed audio stream. This client-side audio stream fully preserves every word spoken by the client during the call, every pause, every interruption's tone and timing, but removes the interference of robotic speech.

[0055] Furthermore, the client-side voice stream from the clipped historical interactive call voice stream is used as a simulated voice stream for subsequent test back-injection.

[0056] In this embodiment, client voice is extracted directly from historical interactive call voice streams in the production environment, i.e., simulated voice streams. These simulated voice streams fully preserve the natural tone, speaking speed, accent, emotional fluctuations, background noise, and unexpected pauses of real customers. Compared to TTS-synthesized speech or human-simulated recordings, the simulated voice streams generated by this solution have unparalleled high fidelity, making the test scenarios closer to the real online environment.

[0057] The following embodiments describe an alternative implementation for acquiring analog speech streams. This embodiment is suitable for scenarios where sufficient authentic historical recordings cannot be obtained from the production environment, or where it is necessary to construct analog speech streams for new test scenarios not yet available online.

[0058] In an exemplary embodiment of the present invention, the analog speech stream can be obtained in the following manner: Based on pre-set interactive response rules, simulated interactive call voice streams of client-outbound call system interaction in test scenarios are recorded. The interactive response rules define client response scripts and response time intervals of client response scripts. Client response scripts are used to characterize the client's response rules under the preset trigger conditions of voice stream triggering on the outbound call system. Based on the simulated interactive call voice stream, the client voice stream under the simulated interactive call voice stream is extracted; The client's voice stream under the simulated interactive call voice stream is used as the analog voice stream.

[0059] In one embodiment, the test engineer pre-sets a set of interaction response rules based on the business logic of the target test scenario. These rules at least define the client's response script and the response time interval for that script.

[0060] The client response script is used to characterize the content of the client's response when the voice stream on the outbound calling system triggers a preset trigger condition. For example, for the "precise interruption" test scenario, this rule defines the client response script as: "Too high."

[0061] The response time interval of the client-side reply script is used to characterize how long the client needs to wait after detecting a specific trigger point on the outbound calling system before sending the reply script. For example, the rule defines that when the outbound calling system plays the keyword "annual interest rate is 100%", the client should delay for 300 milliseconds before replying "That's too high".

[0062] Furthermore, a simulation recording process is executed: Test engineers or automated tools set up a controlled simulation environment in which a reference system with the same dialogue logic as the outbound calling system under test, or the system under test directly using the test environment, is run. Then, a human test engineer or TTS system simulates customer behavior and conducts real-time voice interaction with the reference system according to the aforementioned interaction response rules. Simultaneously, a recording device is started to record the entire interaction process, thereby obtaining a complete simulated interactive call voice stream. This voice stream includes the dialogue output from the reference system and the simulated client's response voice according to the rules (such as "too high"), as well as the precise time interval between the two.

[0063] Since the recorded simulated interactive call voice stream is an audio file (or a two-channel file) that mixes the voice from the reference system and the voice from the simulated client, in another embodiment, voice separation technology can be used to extract the voice stream that belongs only to the simulated client. This client voice stream retains the simulated client's responses such as "too high" spoken according to rules, as well as the natural pauses and tone characteristics between the responses.

[0064] Furthermore, the extracted client voice stream is used as a simulated voice stream file that can be called by the test system, and it is associated with the corresponding test scenario and stored in the test data management library for subsequent test back-injection.

[0065] When the production environment has not accumulated enough historical call recordings, or when the scenario to be tested involves a newly launched service or extreme abnormal situations (such as specific emotional expressions), it is impossible to directly obtain simulated voice streams from real recordings. This embodiment, by pre-setting interactive response rules and recording simulated interactive voice streams, can proactively and controllably generate simulated voice streams for the required test scenarios under zero real data conditions, greatly expanding the scope of application of this invention.

[0066] Figure 3 This is a schematic diagram of the process for performing outbound call testing on the outbound call system under test based on the interactive call voice stream provided by the present invention.

[0067] The following will combine Figure 3 The present invention describes the process of conducting outbound call tests on the outbound call system under test based on interactive call voice stream.

[0068] In an exemplary embodiment of the present invention, combined with Figure 3 As can be seen, based on the interactive call voice stream, the outbound call test of the outbound call system under test can include steps 310 to 330, and each step will be described below.

[0069] In step 310, a standard interactive call voice stream is determined in the target test scenario. The standard interactive call voice stream defines a standard response script and a standard response time interval for the standard response script. The standard response script is a response reply script template of the outbound call system under test in response to the target simulated voice stream.

[0070] In one embodiment, the standard interactive call voice stream is a predefined benchmark template used to measure the correctness of the behavior of the outbound calling system under test. The standard interactive call voice stream defines at least a standard response script and a standard response time interval. The standard response script is the reply script template that the outbound calling system under test should play after correctly responding to the target analog voice stream (e.g., receiving a customer saying "too high"). For example, for a "precise interruption" scenario, the standard response script is defined as: "Sir / Madam, please don't worry, we have specially applied for a preferential interest rate for you..." This script is used to verify whether the system under test has made the correct semantic response according to business requirements. The standard response time interval is the maximum allowed (or expected) duration from when the outbound calling system under test receives the target analog voice stream to when it begins playing the standard response script. For example, if business rules require the system under test to start playing the reassuring script within 800 milliseconds after the customer says "too high," otherwise it is considered a response timeout, then the standard response time interval can be set to 800 milliseconds.

[0071] The standard interactive call voice stream can be pre-set by test engineers according to business requirements documents and stored in the test scenario and rule base, and uniquely associated with the target test scenario.

[0072] In step 320, based on the interactive call voice stream, the actual response script of the simulated voice stream of the outbound call system under test and the actual response time interval of the actual response script are extracted.

[0073] In one embodiment, the interactive call audio stream can be transcribed into text using Automatic Speech Recognition (ASR) technology, and the actual content spoken by the system under test after the customer interrupts can be located from the text. For example, the recognized actual response might be: "Sir / Madam, please don't worry, we have specially applied for a preferential interest rate for you..." Alternatively, audio energy detection and voice-activated audio (VAD) technologies can be used to segment the response audio segments of the system under test from the audio stream, and then their content can be confirmed manually or automatically.

[0074] In another embodiment, two key time points can be precisely located in the interactive call voice stream: the first time point is the end of the injected target simulated voice stream (i.e., the instant the customer finishes saying "too high"), and the second time point is the start of the tested system playing its response script. The time difference between these two time points is the actual response time interval. For example, the actual response time interval measured by audio analysis is 720 milliseconds.

[0075] In step 330, outbound call testing is performed on the outbound call system under test based on the standard response script, the actual response script, the standard response time interval, and the actual response time interval.

[0076] In another embodiment, it can be determined whether the actual response text matches the standard response text. The matching method can be exact text consistency, semantic similarity calculation, or keyword inclusion matching. It can also be determined whether the actual response time interval meets the requirements of the standard response time interval. For example, if the standard response time interval is no more than 800 milliseconds, and the actual measured time interval is 720 milliseconds, then it is determined to meet the requirement; if the actual measured time interval is 950 milliseconds, then it is determined to have timed out, and the test fails.

[0077] Furthermore, based on the above comparison results, the testing platform can automatically generate test conclusions. If the script content is correct and the response delay meets the requirements, the outbound call test result of the outbound call system under test is determined to meet the test requirements; otherwise, the outbound call test result of the outbound call system under test is determined to not meet the test requirements, and the specific reasons for not meeting the test requirements are recorded, such as "incorrect response script" or "response timeout".

[0078] In this embodiment, by predefining standard response scripts and standard response time intervals, and automatically comparing the actual interaction results with the standards, the reliance on subjective human judgment is completely eliminated. The test results are deterministic and repeatable, enabling outbound call testing to automatically output pass / fail results like unit testing, providing a technical foundation for automated quality control.

[0079] In another exemplary embodiment of the present invention, the following description continues using the previously described embodiments as examples. Based on the standard response script, the actual response script, the standard response time interval, and the actual response time interval, outbound call testing of the outbound call system under test can be performed in the following manner: If the standard response script matches the actual response script, and the actual response time interval is less than or equal to the standard response time interval, then the outbound call test results of the outbound call system under test meet the test requirements. If the standard response script does not match the actual response script, and / or the actual response time interval is greater than the standard response time interval, it is determined that the outbound call test results of the outbound call system under test do not meet the test requirements.

[0080] In one embodiment, when the actual response script matches the standard response script and the actual response time interval is less than or equal to the standard response time interval, it can be determined that the outbound call test results of the outbound call system under test meet the test requirements.

[0081] In another embodiment, when a mismatch is detected between the actual response script and the standard response script, and / or the actual response time interval is greater than the standard response time interval, it can be determined that the outbound call test result of the outbound call system under test meets the requirement of not being tested. The matching method can be complete text consistency, semantic similarity calculation, or keyword inclusion matching.

[0082] Through the above judgment logic, an automated and quantitative evaluation of the correctness of the behavior of the tested outbound calling system in the scenario of precise interruption is achieved.

[0083] To further introduce the online data outbound call testing processing method provided by this invention Figure 4 This is a schematic diagram illustrating an application scenario of the online data outbound call testing processing method provided by the present invention. The following will combine... Figure 4 Please provide an explanation.

[0084] In an exemplary embodiment of the present invention, combined with Figure 4 It is known that online environment data collection is feasible. In the production environment, configure the outbound calling system (e.g., by modifying the dialing plan of FreeSWITCH) or utilize its built-in functions to immediately initiate recording at the start of each successfully connected real outbound call. The key point is that the recording should begin the instant the call connects, eliminating the "ringing wait period" after the outbound call is dialed, and only recording the complete two-way interactive voice from the moment the customer answers the phone and says "hello" until the call ends. The recording files (e.g., in WAV format) contain metadata such as timestamps and caller / called numbers, and are stored in the recording repository. These recordings are a complete digital mapping of the real-world online scenario, including all randomness, critical states, and timing information. During application, the raw recordings can be stored in the test data management module.

[0085] In another embodiment, scenario rules can be associated. In the test scenario and rule base, test engineers can define test scenarios that need to be verified based on business logic (e.g., "interrupt the system if the customer says 'too high' within 300 milliseconds after hearing 'the annual interest rate is...'"). The system uses speech analysis (such as transcription using ASR technology) or manual annotation to filter recordings that match or approximate the scenario from a massive recording repository, associates and indexes them with the test scenario, and stores them in the test data management module. The recording files can be trimmed, retaining only the key interaction parts.

[0086] In another embodiment, a test task can be issued. At the start of the test, the test control engine can send a command to the FreeSWITCH media server, based on the selected test scenario, to initiate a test call (SIP INVITE) to the outbound calling system under test. The system under test responds and establishes a call channel (corresponding to step 2, establishing a test call channel). Simultaneously, the test control engine loads the pre-processed recording file associated with the scenario from the test data management module (corresponding to step 1, selecting a pre-processed recording), and prepares to send the audio media stream to FreeSWITCH. FreeSWITCH plays a crucial media proxy and injection role in this process.

[0087] In another embodiment, after the call is established, the test control engine does not immediately play the recording. Instead, it simultaneously monitors the voice stream (i.e., the robot script) from the outbound calling system under test. Through real-time speech recognition or preset keyword detection, the test control engine can accurately determine the script playback progress of the system under test (corresponding to 3. Sending control commands and synchronization). When playback reaches the target trigger point (e.g., the end of the phrase "annual interest rate is"), the test control engine immediately sends a command to FreeSWITCH to inject the recorded media stream representing the "simulated customer" into the call media stream with extremely low latency (e.g., a target latency of 300 milliseconds) (corresponding to 4. Interactive response). This process simulates the interruption behavior of a real customer at a precise time. FreeSWITCH is responsible for mixing or switching the injected recorded media stream with the media stream from the system under test and transmitting it to the other party.

[0088] In another embodiment, the test proceeds according to a predetermined interaction flow in the recording: after the system under test receives an interruption voice from the "simulated customer," it should trigger its internal logic (such as playing a reassuring script or a transfer strategy), and its response voice is transmitted back again via FreeSWITCH. The test control engine can continue to control subsequent interactions according to rules (e.g., injecting recordings of customer consent or refusal at specific nodes). The audio of both parties in the entire test call (including all outputs of the system under test and the injected customer recordings) is completely recorded by FreeSWITCH, forming a test result recording, which is stored in the test result library (corresponding to recording the test result in step 5).

[0089] Furthermore, the recorded test results are automatically analyzed. The audio is transcribed into text using Automatic Speech Recognition (ASR) and compared to the expected business logic in the test scenario (e.g., asserting whether the system played the preset "preferential interest rate" script after the customer said "too high"). Simultaneously, key time intervals (such as the time from the script trigger point to the breakpoint) can be precisely measured to verify whether business rules are met. The comparison results automatically generate a test report, providing a pass or fail conclusion.

[0090] As described above, the present invention provides an online data outbound call testing processing method, the method comprising: determining the target test scenario of the outbound call system to be tested; acquiring a target simulated voice stream matching the target test scenario; establishing a test call for the outbound call system to be tested that matches the target test scenario; in response to triggering a preset trigger condition based on the voice stream of the outbound call system to be tested being monitored in the test call, injecting the target simulated voice stream into the voice stream of the outbound call system to be tested to obtain an interactive call voice stream; and performing outbound call testing on the outbound call system to be tested based on the interactive call voice stream. This method enables low-cost, high-fidelity reproduction of complex online call scenarios, thereby achieving low-cost, high-fidelity outbound call testing of the outbound call system to be tested.

[0091] The online data outbound call testing and processing device provided by the present invention is described below. The online data outbound call testing and processing device described below can be referred to in correspondence with the online data outbound call testing and processing method described above.

[0092] Figure 5 This is a schematic diagram of the online data outbound call testing and processing device provided by the present invention.

[0093] The following will combine Figure 5 The structure of the online data outbound call testing and processing device provided by the present invention will be described.

[0094] In an exemplary embodiment of the present invention, combined with Figure 5 As can be seen, the online data outbound call test processing device may include a determination module 510, an acquisition module 520, an establishment module 530, and a test module 540. Each module will be described in detail below.

[0095] The determination module 510 can be configured to determine the target test scenario for the outbound call system under test. The acquisition module 520 can be configured to acquire a target simulated voice stream that matches the target test scenario based on the target test scenario, wherein the target simulated voice stream is used to simulate a client voice stream that interacts with the outbound call system under the target test scenario; The module 530 can be configured to establish a test call for the outbound calling system under test that matches the target test scenario. In response to triggering a preset trigger condition based on the voice stream of the outbound calling system under test being monitored in the test call, the target simulated voice stream is injected into the voice stream of the outbound calling system under test to obtain an interactive call voice stream. The test module 540 can be configured to perform outbound call tests on the outbound call system under test based on the interactive call voice stream.

[0096] In an exemplary embodiment of the present invention, the acquisition module 520 may further be configured to: Call the pre-maintained test data management library, wherein the test data management library includes the correspondence between different test scenarios and different simulated voice streams; The acquisition module 520 can acquire a target simulated speech stream that matches the target test scenario in the following ways: Based on the correspondence between the target test scenario and the test data management library, a target simulated speech stream matching the target test scenario is obtained.

[0097] In an exemplary embodiment of the present invention, the acquisition module 520 may obtain the analog speech stream in the following manner: Obtain historical voice streams of voice interactions between the outbound calling system and the client; Starting from the call connection time of the historical interactive call voice stream, the historical interactive call voice stream is trimmed to obtain the trimmed historical interactive call voice stream; Based on the cropped historical interactive call voice stream, the client voice stream under the cropped historical interactive call voice stream is extracted; The client voice stream from the cropped historical interactive call voice stream is used as the simulated voice stream.

[0098] In an exemplary embodiment of the present invention, the acquisition module 520 may obtain the analog speech stream in the following manner: Based on pre-set interactive response rules, a simulated interactive call voice stream of the client interacting with the outbound calling system in a test scenario is recorded. The interactive response rules define the client's response script and the response time interval of the client's response script. The client's response script is used to characterize the client's response rules under the preset triggering conditions of the voice stream on the outbound calling system. Based on the simulated interactive call voice stream, the client voice stream under the simulated interactive call voice stream is extracted; The client voice stream under the simulated interactive call voice stream is used as the simulated voice stream.

[0099] In an exemplary embodiment of the present invention, the test module 540 may further be configured to: A standard interactive call voice stream is determined under the target test scenario, wherein the standard interactive call voice stream defines a standard response script and a standard response time interval of the standard response script, and the standard response script is a response reply script template of the outbound call system under test in response to the target simulated voice stream; The test module 540 can perform outbound call testing on the outbound call system under test based on the interactive call voice stream in the following manner: Based on the interactive call voice stream, the actual response script of the outbound calling system under test in response to the target simulated voice stream, and the actual response time interval of the actual response script are extracted. The outbound call system under test is tested based on the standard response script, the actual response script, the standard response time interval, and the actual response time interval.

[0100] In an exemplary embodiment of the present invention, the testing module 540 may perform outbound call testing on the outbound call system under test based on the standard response script, the actual response script, the standard response time interval, and the actual response time interval in the following manner: If the standard response script matches the actual response script, and the actual response time interval is less than or equal to the standard response time interval, it is determined that the outbound call test result of the outbound call system under test meets the test requirements. If the standard response script does not match the actual response script, and / or the actual response time interval is greater than the standard response time interval, it is determined that the outbound call test result of the outbound call system under test does not meet the test requirements.

[0101] Figure 6 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 6 As shown, the electronic device may include: a processor 610, a communications interface 620, a memory 630, and a communication bus 640, wherein the processor 610, the communications interface 620, and the memory 630 communicate with each other through the communication bus 640. The processor 610 can call logical instructions in the memory 630 to execute an online data outbound call test processing method, which includes: determining a target test scenario for the outbound call system under test; acquiring a target simulated voice stream matching the target test scenario, wherein the target simulated voice stream is used to simulate a client voice stream interacting with the outbound call system under test in the target test scenario; establishing a test call for the outbound call system under test that matches the target test scenario; in response to triggering a preset trigger condition based on the voice stream of the outbound call system under test being monitored in the test call, injecting the target simulated voice stream into the voice stream of the outbound call system under test to obtain an interactive call voice stream; and performing an outbound call test on the outbound call system under test based on the interactive call voice stream.

[0102] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0103] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the online data outbound call testing processing method provided by the above methods. The method includes: determining a target test scenario for the outbound call system under test; acquiring a target simulated voice stream matching the target test scenario based on the target test scenario, wherein the target simulated voice stream is used to simulate a client voice stream interacting with the outbound call system under test in the target test scenario; establishing a test call for the outbound call system under test that matches the target test scenario; in response to triggering a preset trigger condition based on the voice stream of the outbound call system under test being monitored in the test call, injecting the target simulated voice stream into the voice stream of the outbound call system under test to obtain an interactive call voice stream; and performing an outbound call test on the outbound call system under test based on the interactive call voice stream.

[0104] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the online data outbound call testing processing method provided by the above methods. The method includes: determining a target test scenario for the outbound call system under test; acquiring a target simulated voice stream matching the target test scenario, wherein the target simulated voice stream is used to simulate a client voice stream interacting with the outbound call system under test in the target test scenario; establishing a test call for the outbound call system under test matching the target test scenario; in response to triggering a preset trigger condition based on the voice stream of the outbound call system under test detected in the test call, injecting the target simulated voice stream into the voice stream of the outbound call system under test to obtain an interactive call voice stream; and performing an outbound call test on the outbound call system under test based on the interactive call voice stream.

[0105] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0106] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0107] 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 foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for processing online outbound data calls, characterized in that, The method includes: Determine the target test scenario for the outbound calling system to be tested; Based on the target test scenario, a target simulated voice stream matching the target test scenario is obtained, wherein the target simulated voice stream is used to simulate the client voice stream that interacts with the outbound call system under the target test scenario; Establish a test call for the outbound calling system under test that matches the target test scenario. In response to the triggering of a preset trigger condition based on the voice stream of the outbound calling system under test being monitored in the test call, inject the target simulated voice stream into the voice stream of the outbound calling system under test to obtain an interactive call voice stream. Based on the interactive call voice stream, the outbound call system under test is tested for outbound calls.

2. The online outbound data call testing processing method according to claim 1, characterized in that, Before obtaining the target simulated speech stream matching the target test scenario based on the target test scenario, the method further includes: Call the pre-maintained test data management library, wherein the test data management library includes the correspondence between different test scenarios and different simulated voice streams; The step of obtaining a target simulated speech stream that matches the target test scenario includes: Based on the correspondence between the target test scenario and the test data management library, a target simulated speech stream matching the target test scenario is obtained.

3. The online data outbound call testing processing method according to claim 2, characterized in that, The analog speech stream is obtained in the following way: Obtain historical voice streams of voice interactions between the outbound calling system and the client; Starting from the call connection time of the historical interactive call voice stream, the historical interactive call voice stream is trimmed to obtain the trimmed historical interactive call voice stream; Based on the cropped historical interactive call voice stream, the client voice stream under the cropped historical interactive call voice stream is extracted; The client voice stream from the cropped historical interactive call voice stream is used as the simulated voice stream.

4. The online outbound data call testing processing method according to claim 2, characterized in that, The analog speech stream is obtained in the following way: Based on pre-set interactive response rules, a simulated interactive call voice stream of the client interacting with the outbound calling system in a test scenario is recorded. The interactive response rules define the client's response script and the response time interval of the client's response script. The client's response script is used to characterize the client's response rules under the preset triggering conditions of the voice stream on the outbound calling system. Based on the simulated interactive call voice stream, the client voice stream under the simulated interactive call voice stream is extracted; The client voice stream under the simulated interactive call voice stream is used as the simulated voice stream.

5. The online outbound data call testing processing method according to claim 1, characterized in that, Before performing outbound call testing on the outbound call system under test based on the interactive call voice stream, the method further includes: A standard interactive call voice stream is determined under the target test scenario, wherein the standard interactive call voice stream defines a standard response script and a standard response time interval of the standard response script, and the standard response script is a response reply script template of the outbound call system under test in response to the target simulated voice stream; The outbound call test of the outbound call system under test based on the interactive call voice stream includes: Based on the interactive call voice stream, the actual response script of the outbound calling system under test in response to the target simulated voice stream, and the actual response time interval of the actual response script are extracted. The outbound call system under test is tested based on the standard response script, the actual response script, the standard response time interval, and the actual response time interval.

6. The online data outbound call testing processing method according to claim 5, characterized in that, The outbound call test of the system under test, based on the standard response script, the actual response script, the standard response time interval, and the actual response time interval, includes: If the standard response script matches the actual response script, and the actual response time interval is less than or equal to the standard response time interval, it is determined that the outbound call test result of the outbound call system under test meets the test requirements. If the standard response script does not match the actual response script, and / or the actual response time interval is greater than the standard response time interval, it is determined that the outbound call test result of the outbound call system under test does not meet the test requirements.

7. An online data outbound call testing and processing device, characterized in that, The device includes: The determination module is used to determine the target test scenario for the outbound calling system under test; The acquisition module is used to acquire a target simulated voice stream that matches the target test scenario based on the target test scenario, wherein the target simulated voice stream is used to simulate the client voice stream that interacts with the outbound call system under the target test scenario; The module is used to establish a test call for the outbound calling system under test that matches the target test scenario. In response to triggering a preset trigger condition based on the voice stream of the outbound calling system under test being monitored in the test call, the target simulated voice stream is injected into the voice stream of the outbound calling system under test to obtain an interactive call voice stream. The testing module is used to perform outbound call tests on the outbound call system under test based on the interactive call voice stream.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the online data outbound call test processing method as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the online data outbound call test processing method as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the online data outbound call test processing method as described in any one of claims 1 to 6.

Images