Voice content retrieval system applied to insurance sales process

Abstract

The invention discloses a voice content retrieval system applied to an insurance sales process, and the system comprises a recording and text conversion module which is used for recording and converting voice into text content; the text content intelligent analysis module is used for analyzing and extracting the converted text content and generating a structured label for the recording; the retrieval engine module is used for performing retrieval according to the keywords; the character and voice synchronization module is used for synchronously matching the character and voice content of the retrieval result, and can request the character and voice content of the area according to the user requirement; the key content summarizing and recombining module is used for extracting and summarizing key phrases and themes of the insurance industry, recombining the key phrases and themes into text and voice record contents only containing insurance key information, and playing the text and voice record contents; the system has the characteristics of shortening the information processing period, providing data support and realizing rapid backtracking of the sales process.

Description

Technical Field

[0001] This invention relates to the fields of information technology and insurance services, and in particular to a voice content retrieval system applied in the insurance sales process. Background Technology

[0002] In the insurance sales process, customer consultation and communication are crucial, directly impacting the accuracy of product understanding, the grasp of customer needs, and the quality of follow-up services. In actual sales, communication between salespersons and customers often lasts a considerable amount of time and frequently includes content with low relevance to insurance information. Relying on existing note-taking methods or current recording and listening technologies, content confirmation and manual review based on manuscripts or original recordings are inefficient and hinder effective information organization and utilization. With the development of artificial intelligence technology, converting voice data into analyzable text information and performing content retrieval and extraction based on this has become key to improving insurance sales efficiency and customer satisfaction. Summary of the Invention

[0003] To achieve the above objectives, the inventors provide a voice content retrieval system applied in the insurance sales process, comprising:

[0004] The recording and text conversion module is used to record audio and convert speech into text content;

[0005] The intelligent text content analysis module is used to analyze the converted text content, extract key phrases and themes, and generate structured tags for the recording.

[0006] The search engine module is used for searching based on key phrases and topics;

[0007] The text and voice synchronization module is used to synchronize and match the text and voice content of the search results, and can play the text and voice content of the area according to the user's needs.

[0008] The key content summary and reorganization module is used to extract and summarize conversation content that is strongly related to insurance business, targeting key phrases and themes in the insurance industry, and reorganize it into text and voice recordings containing only key insurance information. It can also play the text and voice recordings synchronously as needed by authorized personnel.

[0009] As a preferred embodiment of the present invention, it further includes: a keyword extraction module and a key expression extraction module, used to train materials from the insurance industry to obtain key phrases, themes and semantic databases of the insurance industry, and to serve as keywords and key expressions in the text content intelligent analysis module and the key content summarization and reorganization module.

[0010] In a preferred embodiment of the present invention, the key expressions refer to the relationships between words and between sentences.

[0011] As a preferred embodiment of the present invention

[0012] The output of the keyword extraction module is all the keywords in the insurance industry, as well as the weight of each keyword;

[0013] The output of the key expression extraction module includes key phrases and their weights in the insurance industry, key sentences and their weights, relationships between phrases and their weights, and relationships between sentences and their weights.

[0014] As a preferred embodiment of the present invention, it further includes: a speech rate adjustment module, used to play the text and speech recordings at different speech rates according to the user's needs.

[0015] As a preferred embodiment of the present invention, it further includes: a pitch correction module for correcting the pitch of the played text and speech content.

[0016] As a preferred embodiment of the present invention, it further includes: a data encryption and privacy management module, used to encrypt the recording and text content, and to control and manage authorized personnel who view and use the relevant data.

[0017] As a preferred embodiment of the present invention, it further includes: a sales strategy optimization module, used to provide sales strategies and optimization suggestions based on data analysis and mining of historical information.

[0018] Unlike existing technologies, the above technical solution achieves the following beneficial effects:

[0019] (1) This system can effectively improve sales efficiency. Through real-time voice-to-text and intelligent search functions, it greatly shortens the information processing cycle, enabling sales personnel to respond to customer needs more quickly.

[0020] (2) This system can effectively improve service quality by accurately grasping the customer's concerns and providing more personalized and targeted services, thereby enhancing customer satisfaction;

[0021] (3) This system can effectively enhance decision support, and through rich data analysis reports and sales strategy suggestions, it provides effective data support for the market positioning of insurance products and the training of sales teams;

[0022] (4) This system can effectively realize the rapid retrospective of the sales process. Through records of different speaking speeds, original and condensed versions, it can provide salespersons and users with a rapid means of process retrospective;

[0023] (5) This system can effectively protect information security. Through a powerful data encryption and access control mechanism, it effectively ensures the security and privacy protection of customer data. Attached Figure Description

[0024] Figure 1 The system flowchart is for a specific implementation method. Detailed Implementation

[0025] To explain in detail the technical content, structural features, objectives, and effects of the technical solution, the following description is provided in conjunction with specific embodiments and accompanying drawings.

[0026] like Figure 1 As shown, this embodiment provides a voice content retrieval system applied in the insurance sales process, including:

[0027] The recording and text conversion module is used to record audio and convert speech into text. Specifically, it uses a speech recognition algorithm to automatically record audio during interactions between customers and sales personnel and converts the speech into text in real time. This process ensures high accuracy and low latency and can cover multiple languages ​​and dialects.

[0028] The intelligent text content analysis module is used to analyze the converted text content, extract key phrases and themes, and generate structured tags for the recordings. Specifically, the transcribed text content undergoes in-depth analysis using Natural Language Processing (NLP) technology to extract key phrases and themes, such as "insurance," "personal," "incident," "claims," ​​"medical," and "protection," and generates structured tags for the recordings, such as "personal insurance," "medical protection," and "incident claims," ​​thus facilitating classification and archiving.

[0029] The search engine module is used for searching based on key phrases and topics. Specifically, by building a machine learning-based search engine, it supports multi-dimensional searches, such as rapid retrieval by key phrases and topics like insurance product type, customer needs, and frequently asked questions. This system understands search intent and provides relevance-ranked results to help sales personnel quickly find similar cases or answer templates.

[0030] The text and voice synchronization module is used to automatically synchronize and match the text and voice content of the search results during the search process, and can play the text and voice content of the area according to the user's needs.

[0031] The key content summarization and reorganization module is used to extract and summarize conversation content strongly related to insurance business from key phrases and topics within the insurance industry. This content is then reorganized into text and audio recordings containing only key insurance information, and can be played synchronously on the client side as needed by authorized personnel. Specifically, based on machine learning technology, it automatically extracts, abstracts, and summarizes conversation content strongly related to insurance business from key phrases and topics within the insurance industry, reorganizing it into a high-density, condensed version of text and audio recordings containing only key insurance information. This version can then be played synchronously on the client side as needed by authorized personnel.

[0032] In some embodiments, the system further includes a keyword extraction module and a key expression extraction module, used to train materials from the insurance industry to obtain a database of key phrases, themes, and semantics for the insurance industry, which serve as keywords and key expressions in the text content intelligent analysis module and the key content summarization and reorganization module. Specifically: by designing a corpus trainer, targeting various materials from the insurance industry (hereinafter referred to as a corpus or corpus training library), including insurance college textbooks, insurance industry laws and regulations, various insurance product manuals, insurance standard contract texts, insurance sales training materials, and insurance-related litigation texts, etc., a specialized database of key phrases, themes, and semantics for the insurance industry is obtained after training using artificial intelligence methods, which serves as keywords and key expressions in the above-mentioned text content intelligent analysis module and key content summarization and reorganization module. In the above embodiments, "key expressions" refer to key words and the relationships between words. In layman's terms, words and words further constitute key phrases, as well as the relationships between sentences. In the above embodiments, the output of the keyword extraction module is all keywords in the insurance industry and the weight of each keyword; the output of the key expression extraction module is key phrases and their weights, key sentences and their weights, the relationships between phrases and their weights, and the relationships between sentences and their weights. Taking the automatic extraction algorithm of keywords and key phrases as an example, suppose the text used in the training database of an insurance lawsuit is as follows: "Zhang, an insurance company agent, and Wang, the policyholder, are classmates. When Zhang was selling insurance products to Wang, Wang was on a business trip. So Wang asked Zhang to come to his home to collect the insurance premium from his wife. Zhang went to Wang's home, found Wang's wife, collected the insurance premium, and signed the insurance application on Wang's behalf. The policyholder and insured on the application were both Wang. The type of insurance was critical illness insurance, the insurance period was lifelong, the premium payment period was 20 years, and the annual premium was 2,000 yuan. After Wang returned to Beijing from his business trip, Zhang gave Wang the insurance contract and the premium invoice. Thereafter, Wang paid the insurance premium normally every year, accumulating 12,000 yuan. Until 2006, the relationship between Wang and Zhang deteriorated, and Wang sued the insurance company, demanding a refund of all insurance premiums on the grounds that the insurance application was not signed by him." The corpus text used, originating from a non-insurance industry, included, for example: "My name is Li Haoran, and I am the technical director of a software development company. At the age of 28, I went to Shanghai for a three-month project collaboration. During this time, I participated in an industry exchange conference, where I met many industry experts and engaged in in-depth academic and technical discussions. This experience greatly enriched my professional perspective." The non-insurance industry corpus consisted of broad materials from outside the insurance industry, without industry or scenario restrictions. When each text was input into the corpus trainer, it was manually pre-labeled to indicate whether it belonged to the insurance industry.The corpus trainer calculates insurance industry keywords and their weights based on a comparison of the probability of each word appearing in all insurance-related corpora and the probability of each word appearing in all non-insurance industry corpora, according to a specific algorithm. The output is in the form of "insurance (0.95), salesperson (0.3), insurance application (0.9), sales promotion (0.3), type of insurance (0.8), fee (0.7), invoice (0.5), contract (0.5), critical illness (0.8), etc." These words and the weights expressed by the numbers in parentheses will be automatically included in the insurance industry keyword database. Conversely, words such as "business trip" and "Beijing" have a low probability of appearing in all insurance-related corpora but a high probability of appearing in all non-insurance industry corpora, indicating that these words are common words and not insurance industry keywords, and therefore will not be included in the insurance keyword database. The training of insurance sentences, word relationships, sentence relationships, and other information follows a similar process, and this information is also included in the key expression database of this system. During actual recording, the above-mentioned text content intelligent analysis module and key content summarization and reorganization module will determine whether the identified text, such as words, sentences, word-word combinations, and sentence-sentence combinations, has a high correlation with the training results of the system's keyword and key expression library, i.e., equal or similar. Then, the recording and text will be automatically marked by the system as the key content of the recording.

[0033] In different embodiments, the system further includes a speech rate adjustment module, used to play the text and speech recordings at different speeds according to user needs. Specifically, it can play the original or condensed versions of the text and speech recordings at different speeds according to user needs, and uses a pitch correction module to ensure pitch fidelity at different playback speeds. Simply increasing the playback speed without pitch correction will result in an overall higher sound frequency, causing phenomena such as "male voice turning into female voice." Pitch correction ensures that the person's pitch characteristics remain unchanged when the speech rate is increased. The same principle applies to slowing down the playback speed.

[0034] In addition, this system includes a data encryption and privacy management module, used to encrypt recorded and text content and control the authorized personnel who can view and use related data. Specifically, it employs encryption technologies such as Advanced Encryption Standard (AES) to ensure the security of all recorded and transcribed text data. Simultaneously, strict data management policies are implemented, allowing only authorized personnel to view and use related data, fully protecting customer privacy. This system also includes a sales strategy optimization module, used for data analysis and mining based on historical information to provide sales strategies and optimization suggestions. Specifically, based on data analysis and mining of historical information, this system can further provide sales strategy optimization suggestions, such as recommendations for popular insurance products, customer preference analysis, and communication skills improvement, assisting sales personnel in developing more effective sales plans to achieve higher conversion rates.

[0035] In the specific implementation of the above embodiments, this system can be embedded in online sales platforms, offline sales scenarios, and customer service hotline systems, automatically activating recording and transcription functions from the moment a customer inquires. The sales team can search relevant conversation records at any time through the backend retrieval system by entering keywords, quickly obtaining sales references and customer feedback information, thereby adjusting sales strategies and improving overall performance.

[0036] This system provides an integrated solution that automatically records, transcribes, and summarizes conversations during the insurance sales process. Through intelligent analysis and retrieval technologies, it helps insurance sales personnel quickly identify customer concerns, optimize sales strategies, reduce the difficulty and cost of retrospective analysis, and simultaneously ensure the security and privacy of customer information. It achieves two retrieval functions: first, it automatically converts speech into text and establishes an efficient content indexing mechanism to facilitate later keyword searches for related topics and simultaneous retrospective analysis of text and speech; second, through artificial intelligence technology, it automatically extracts, summarizes, and reorganizes the entire communication content according to insurance industry keywords and key expressions, creating a condensed version containing high-density key information. This condensed version of text and speech is then automatically played back, facilitating quick confirmation during contract signing and rapid retrospective analysis. In summary, by integrating advanced artificial intelligence technologies, this system effectively solves the challenges of processing and efficiently utilizing voice information in the insurance sales process, providing strong technical support for the digital transformation of the insurance industry, and possesses significant innovative value and broad application prospects.

[0037] It should be noted that although the above embodiments have been described herein, this does not limit the scope of patent protection of the present invention. Therefore, any changes and modifications made to the embodiments described herein based on the innovative concept of the present invention, or equivalent structural or procedural transformations made using the content of the present invention's specification and drawings, directly or indirectly applying the above technical solutions to other related technical fields, are all included within the scope of patent protection of the present invention.

Claims

1. A voice content retrieval system applied in the insurance sales process, characterized in that, include: The recording and text conversion module is used to record audio and convert speech into text content; The intelligent text content analysis module is used to analyze the converted text content, extract key phrases and themes, and generate structured tags for the recording. The search engine module is used for searching based on key phrases and topics; The text and voice synchronization module is used to synchronize and match the text and voice content of the search results, and can play the text and voice content of the area according to the user's needs. The key content summarization and reorganization module is used to extract and summarize conversation content that is strongly related to insurance business, targeting key phrases and themes in the insurance industry, and reorganize it into text and voice recordings containing only key insurance information. It can also play this version of text and voice on the client side as needed by authorized personnel.

2. The voice content retrieval system applied to the insurance sales process according to claim 1, characterized in that, Also includes: The keyword extraction module and the key expression extraction module are used to train materials from the insurance industry to obtain a database of key phrases, themes, and semantics in the insurance industry. These serve as keywords and key expressions in the text content intelligent analysis module and the key content summarization and reorganization module.

3. The voice content retrieval system applied to the insurance sales process according to claim 2, characterized in that: The key expressions mentioned are the relationships between words and the relationships between sentences.

4. The voice content retrieval system applied to the insurance sales process according to claim 2, characterized in that: The output of the keyword extraction module is all the keywords in the insurance industry, as well as the weight of each keyword; The output of the key expression extraction module includes key phrases and their weights in the insurance industry, key sentences and their weights, relationships between phrases and their weights, and relationships between sentences and their weights.

5. The voice content retrieval system applied to the insurance sales process according to any one of claims 1 to 4, characterized in that, Also includes: The speech rate adjustment module is used to play text and voice recordings at different speeds according to the user's needs.

6. The voice content retrieval system applied to the insurance sales process according to claim 5, characterized in that, Also includes: The pitch correction module is used to correct the pitch of the played text and speech content.

7. The voice content retrieval system applied to the insurance sales process according to any one of claims 1 to 4, characterized in that, Also includes: The data encryption and privacy management module is used to encrypt audio and text content and control the authorized personnel who can view and use related data.

8. The voice content retrieval system applied to the insurance sales process according to any one of claims 1 to 4, characterized in that, Also includes: The sales strategy optimization module is used to provide sales strategies and optimization suggestions based on data analysis and mining of historical information.

Images