Online insurance buying and claim settlement integrated processing method and system
By establishing an initial risk state vector during the insurance application stage and dynamically updating the trust state vector during the claims stage, combined with state-space equations and risk pricing models, the problem of information fragmentation between online insurance application and claims processes is solved, achieving continuity and dynamic updates in risk assessment, and improving risk control and customer experience.
Patent Information
- Application Number
- CN202511338896.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-18
- Publication Date
- 2025-12-09
AI Technical Summary
The existing online insurance application and claims processes are fragmented, resulting in a lack of continuity in risk assessment. Claims risk assessments are static and difficult to update dynamically, affecting the accuracy of risk control and customer experience.
By establishing an initial risk state vector during the insurance application stage and dynamically updating it in conjunction with the trust state vector during the claims stage, the prepayment amount is calculated using state space equations and risk pricing models, thus achieving seamless integration between insurance application and claims and dynamic risk assessment.
It has achieved consistency and accuracy in risk assessment from insurance application to claims settlement, improved the quality of automated processing decisions, enhanced the ability to identify and control potential claims risks, and optimized operational efficiency and customer satisfaction.
Smart Images
Figure CN121094751A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of financial data processing technology, specifically to an integrated online insurance application and claims processing method and system. Background Technology
[0002] With the development of information technology, the insurance industry is undergoing a profound digital transformation. Online insurance application and online claims processing, as core business processes, have become key competitive advantages in terms of automation and intelligence. Therefore, integrating the insurance application and claims processes to achieve intelligent processing throughout the entire process from risk underwriting to claims service is an important direction for current technological development in the industry.
[0003] While online insurance application and claims processing are now widespread, they are typically operated as independent processes. The application process focuses on facilitating policy issuance, while the claims process involves separate document verification. This results in the ineffective utilization of customer risk information accumulated during the application process, requiring separate risk assessment for claims and hindering its integration with the risk profile established during the application phase. This creates an "information gap" between the application and claims processes.
[0004] The aforementioned shortcomings result in a lack of consistency in risk assessment within existing solutions, making it difficult to accurately identify risks. Furthermore, the claims processing model is relatively static, making it difficult to dynamically update risk judgments based on new evidence within the claims cycle, leading to rigid assessment results. Therefore, when making automated decisions such as advance payments, the processing flow struggles to achieve risk-based, refined pricing, often relying on uniform standards or complete manual intervention. This not only impacts customer experience and processing efficiency but also limits the accuracy of risk control.
[0005] Therefore, this invention proposes an integrated online insurance application and claims processing method and system to address the shortcomings of existing technologies. Summary of the Invention
[0006] To address the shortcomings of existing technologies, this invention provides an integrated online insurance application and claims processing method and system, which solves the problem of information separation between the application and claims processes in existing technologies, leading to a lack of accurate risk pricing basis in the static and automated decision-making process of claims risk assessment.
[0007] To achieve the above objectives, the present invention provides the following technical solution: an integrated online insurance application and claims processing method, comprising the following steps:
[0008] S1. Receive the insurance policy generated from the online insurance application, and establish an initial risk state vector for the insurance application;
[0009] S2. When an online claim application is received based on the policy generated from the insurance application, the trust state vector corresponding to the claim application is initialized based on the initial risk state vector.
[0010] S3. During the processing cycle of the online claim application, when updated information related to the online claim application is received, the trust state vector is updated based on the state space equation.
[0011] S4. Based on the updated trust state vector, calculate the trust scalar corresponding to the claim application;
[0012] S5. Based on the trust scalar and the estimated total compensation amount, calculate the prepayment amount corresponding to the claim application through a risk pricing model;
[0013] S6. In response to the policyholder's confirmation instruction, transfer the prepayment amount to the financial account designated by the policyholder.
[0014] Preferably, in step S1, the steps of receiving the insurance policy generated from the online insurance application and establishing an initial risk state vector for the insurance application include:
[0015] Receive and process online insurance application data, including insurance information, to generate policies and store them in the policy database;
[0016] Retrieves the structured policyholder information, risk information, and underwriting terms corresponding to the policy from the policy database;
[0017] Using the unique identifier in the policyholder's information as the query index, historical risk data associated with the policyholder is retrieved from the historical database. The historical database is a risk information database, which stores the policyholder's historical insurance records, historical claims records, and external credit data.
[0018] The information of the insured party, the information of the risk object, the underwriting terms and historical risk data are mapped in a preset numerical manner to generate a set of multi-dimensional risk feature values;
[0019] A set of multi-dimensional risk feature values are combined into a vector to construct the initial risk state vector.
[0020] Preferably, the insured information includes: insured identification information, historical claims records, and third-party credit scores;
[0021] The information regarding the risk object includes: vehicle model, the insured's health declaration, and the location of the property;
[0022] The terms and conditions of coverage include: scope of insurance liability, exclusions, deductibles, and payout limits.
[0023] Preferably, in step S2, when an online claim application initiated by a policy generated based on the insurance application is received, the step of initializing the trust state vector corresponding to the claim application based on the initial risk state vector includes:
[0024] When an online claim application is received based on an insurance policy generated from an insurance application, the policy identifier and claim event information are parsed from the online claim application.
[0025] Based on the policy identifier, retrieve the initial risk state vector corresponding to the policy from the storage medium;
[0026] The claims event information is subjected to feature extraction and numerical processing to generate a set of claims event feature values;
[0027] The elements within the initial risk state vector, combined with the set of claim event feature values, are used to construct the trust state vector.
[0028] Preferably, the step of using elements within the initial risk state vector and combining them with the set of claim event feature values to construct the trust state vector includes:
[0029] The elements in the initial risk state vector are defined as components of the insured historical risk dimension.
[0030] Define the set of claim event feature values as the current claim event dimension components;
[0031] The risk dimension component of the insurance history is combined with the current claim event dimension component to form the trust state vector.
[0032] Preferably, in step S3, during the processing cycle of the online claim application, when updated information related to the online claim application is received, the step of updating the trust state vector based on the state space equation includes:
[0033] During the processing period of an online claim application, receive updated information associated with the online claim application;
[0034] The updated information is quantified to generate an update event vector U. k ;
[0035] Using a preset state transition matrix A and control input matrix B, the updated trust state vector T is calculated through state-space equations. k ;
[0036] The formula for calculating the state-space equation is as follows:
[0037] Tk =A·T k-1 +B·U k ;
[0038] In the formula, T k-1 This is the trust state vector before the update.
[0039] Preferably, in step S4, the step of calculating the trust scalar corresponding to the claim application based on the updated trust state vector includes:
[0040] Retrieve the preset weight vector corresponding to the updated trust state vector dimension;
[0041] Each element in the updated trust state vector is multiplied by the corresponding element in the weight vector to generate multiple product values.
[0042] The multiple product values are summed, and the summation result is used as the trust scalar corresponding to the claim application.
[0043] Preferably, in step S5, the step of calculating the prepayment amount corresponding to the claim application based on the trust scalar and the estimated total compensation amount through a risk pricing model includes:
[0044] The risk pricing model converts the trust scalar into a prepayment ratio. The risk pricing model is a preset function used to define the mapping relationship between the trust scalar and the prepayment ratio, and the prepayment ratio increases as the trust scalar increases.
[0045] Multiply the estimated total compensation amount by the prepayment ratio to calculate the prepayment amount corresponding to the claim application;
[0046] The risk pricing model is as follows:
[0047] In the formula, P k This represents the prepayment ratio calculated by the model at time k, where 0 ≤ P k ≤P max S k P represents the trust scalar calculated in the previous stage at time k; max For a preset maximum prepayment ratio, 0 <P max ≤1; α is a preset gain or scaling parameter, α>0; β is a preset offset or threshold parameter, P max The three model parameters, α, β, and β, can all be statistically calibrated based on historical claims data or set by risk control strategy experts to provide differentiated risk pricing for different product lines and customer groups.
[0048] Preferably, in step S6, the step of transferring the prepayment amount to the financial account designated by the insured in response to the policyholder's confirmation instruction includes:
[0049] Present the insured party with a payment confirmation interface that includes the prepayment amount and financial account information;
[0050] Receive the confirmation instruction triggered by the insured party on the payment confirmation interface;
[0051] In response to the confirmation instruction, a payment instruction including the prepayment amount and the financial account information is generated;
[0052] The payment instruction is sent to the payment processing system to transfer the prepayment amount to the financial account, thereby processing the prepayment for the online claim application.
[0053] This invention also provides an integrated online insurance application and claims processing system, the system comprising:
[0054] The risk vector establishment module is used to receive the insurance policy generated from the online insurance application and establish an initial risk state vector for the insurance application.
[0055] The trust vector initialization module is used to initialize the trust state vector corresponding to the claim application based on the initial risk state vector when an online claim application initiated by the policy generated based on the insurance application is received.
[0056] The trust vector update module is used to update the trust state vector based on the state space equation when it receives update information related to the online claim during the processing cycle of the online claim application.
[0057] The trust scalar calculation module is used to calculate the trust scalar corresponding to the claim application based on the updated trust state vector.
[0058] The advance payment calculation module is used to calculate the advance payment amount corresponding to the claim application based on the trust scalar and the estimated total compensation amount through a risk pricing model.
[0059] The prepayment processing module is used to transfer the prepayment amount to the financial account designated by the insured in response to the policyholder's confirmation instruction.
[0060] This invention provides a method and system for integrated online insurance application and claims processing. It offers the following advantages:
[0061] 1. This invention establishes an initial risk state vector during the insurance application stage, uniformly quantifying information such as the insured's historical data, the risk object, and the underwriting terms. This lays an objective and consistent risk assessment benchmark for subsequent claims processing. This design breaks down the information barriers in the traditional insurance application and claims process, achieving seamless risk assessment from source to end. This results in higher consistency and accuracy in risk judgment throughout the integrated online insurance application and claims process, improving the decision-making quality of automated processing.
[0062] 2. This invention introduces a trust state vector and applies state-space equations to dynamically model updated information within the claims processing cycle, achieving continuous and iterative updates to the trust level of claims cases. This method transforms claims trust assessment from a static, one-time judgment into a dynamic, evolving process, enabling the real-time incorporation of new evidence into the assessment system. This allows the trust assessment results to more accurately approximate the true situation of the case, thereby effectively improving the ability to identify and control potential claims risks.
[0063] 3. This invention automates the processing of claims prepayments by transforming dynamically updated trust state vectors into a single trust scalar and automatically calculating the prepayment amount matching the risk level based on a risk pricing model. This method provides differentiated and precise prepayment strategies for claims applications with different trust levels, significantly accelerating the processing speed of high-trust cases and greatly improving customer satisfaction. It also optimizes the allocation of claims resources, allowing manual review to focus more on high-risk cases, achieving an effective balance between operational efficiency and risk control. Attached Figure Description
[0064] Figure 1 This is a flowchart of the method of the present invention;
[0065] Figure 2 This is a schematic diagram of the dynamic trust assessment and quantification process of the present invention;
[0066] Figure 3 This is a schematic diagram of the automated prepayment decision-making and execution process of the present invention;
[0067] Figure 4 This is a system architecture diagram of the present invention. Detailed Implementation
[0068] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0069] Please see the appendix Figure 1 -Appendix Figure 3 This invention provides an integrated online insurance application and claims processing method, comprising the following steps:
[0070] S1. Receive the insurance policy generated from the online insurance application, and establish an initial risk state vector for the insurance application;
[0071] In this embodiment, an initial risk state vector is established for the online insurance application, aiming to lay an objective and quantitative risk benchmark for the subsequent claims trust assessment process. This initial risk state vector is a mathematical characterization of the inherent risk characteristics of a policy at the time of underwriting, and its establishment process provides a crucial initial state for the subsequent dynamic trust evolution.
[0072] Specifically, this processing method first receives online insurance application data submitted by the policyholder through a pre-defined online service interface. The online service interface can be a portal website, a mobile device application, or a dedicated application programming interface (API) to adapt to different business access scenarios. After receiving and processing the online insurance application data, including the insurance information, an electronic policy is generated, and the policy and all associated information are persistently stored in a policy database. This step ensures the secure storage of policy data and convenient access to subsequent processes.
[0073] To establish a comprehensive risk assessment benchmark, the process needs to retrieve the structured policyholder information, risk information, and underwriting terms associated with the new policy from the policy database.
[0074] Meanwhile, to incorporate the insured's historical behavioral risk dimension, the processing flow uses a unique identifier (such as an individual's ID number or a company's unified credit code) from the insured's information as a query index to retrieve historical risk data associated with that insured from a separate, pre-built historical database. In one specific implementation, this historical database is a risk information database that stores the insured's past insurance records, claims records, and legally obtainable external credit data. Introducing historical risk data aims to ensure that risk assessment is not limited to the current insurance purchase itself, but is based on behavioral patterns over a longer period, thereby obtaining a more stable and reliable risk profile.
[0075] In a preferred embodiment, the data used to establish the initial risk state vector may specifically include:
[0076] The insured information, used to assess the credit and historical risk of the insured entity, may specifically include: insured entity identification information to uniquely identify the entity; the entity's past claims records as a direct reference for its risk behavior; or third-party credit scores obtained from external authoritative institutions as an objective evaluation of its overall credit.
[0077] The information regarding the risk object is used to assess the specific risk characteristics of the insured object, and its content varies depending on the type of insurance product. For example, in the case of vehicle insurance, the risk object information may be the vehicle model; in the case of health insurance, it may be the insured's health declaration, which details the insured's current health status, past medical history, and other key information; and in the case of property insurance, it may be the location information of the insured property to assess the environmental risks of its location.
[0078] The aforementioned underwriting terms clarify the specific scope of rights and responsibilities of the insurance contract, which may include: the scope of insurance liability, defining which losses are eligible for compensation; exclusion clauses, clarifying the specific circumstances under which compensation will not be paid; and the deductible and compensation limits agreed upon by both parties, all of which directly affect the degree of risk exposure.
[0079] Next, we proceed to the crucial numerical mapping step. The purpose of this step is to perform a pre-defined numerical mapping on the insured's information, the information on the risky object, the underwriting terms, and historical risk data. This transforms the previously collected, diverse data (including text descriptions, classification options, qualitative declarations, etc.) into unified, standardized numerical values suitable for mathematical modeling, generating a set of multi-dimensional risk characteristic values. This mapping process is accomplished through a series of pre-defined rules or transformation functions. For example, different vehicle models can be mapped to a specific risk score based on their market share, parts-to-vehicle ratio, and historical accident rate; the insured's health status declaration, such as "good," "fair," or "pre-existing medical history," can be mapped to different numerical levels, such as 1, 2, and 3; and third-party credit scores can be normalized to fall within a fixed numerical range, such as 0 to 1, to facilitate unified dimensional calculations with other characteristic values.
[0080] Finally, the set of multi-dimensional risk feature values generated through numerical mapping will be combined into a mathematical vector according to a pre-set dimensional order to construct the initial risk state vector.
[0081] This vector, in a compact mathematical form, comprehensively encapsulates the overall risk level of the policy at the outset. It is not only the final output of this insurance risk assessment, but also the cornerstone for initializing the trust state vector in the subsequent claims process, ensuring the continuity and consistency of risk assessment from the insurance application stage to the claims process.
[0082] S2. When an online claim application is received based on the policy generated from the insurance application, the trust state vector corresponding to the claim application is initialized based on the initial risk state vector.
[0083] In this embodiment, this step is a key bridge connecting the risk assessment in the insurance application stage and the trust assessment in the claims stage. Its purpose is to establish an assessment starting point that can reflect the initial full picture of a claims case and can be used for subsequent dynamic evolution.
[0084] Specifically, when an online claim application is received based on an insurance policy generated from the aforementioned insurance application, the first step in the processing flow is to parse the policy identifier and claim event information from the online claim application. The policy identifier, such as the policy number, is a credential associated with a unique and valid policy. The claim event information provides a preliminary description of the incident, such as the time and location of the incident, a brief description of what happened, and a preliminary estimate of the loss.
[0085] Subsequently, the processing flow retrieves the initial risk status vector corresponding to the policy from the storage medium based on the policy identifier. The storage medium can be the aforementioned policy database or a dedicated risk vector library. This step ensures that the quantified static risk profile presented to the customer at the time of insurance application can be seamlessly transferred to the claims process, providing a historical benchmark for subsequent trust assessment.
[0086] Simultaneously, the processing flow requires feature extraction and numerical processing of the claims event information to generate a set of claims event feature values. In one specific implementation, this process may include: comparing the reporting time with the accident occurrence time, calculating the timeliness of the reporting and mapping it to a score; and classifying and encoding the accident type (such as single-vehicle accident, two-vehicle accident, personal injury accident, etc.). These quantified claims event feature values collectively constitute a preliminary measure of the immediate risk of this claims event.
[0087] In order to construct a comprehensive assessment vector that can simultaneously reflect both the customer's historical inherent risks and the immediate risks of current claims, the processing flow will use the elements in the initial risk state vector and combine them with a set of claim event feature values to jointly form the trust state vector.
[0088] In a preferred embodiment, this step of constructing the trust state vector is specifically implemented as follows:
[0089] First, the elements within the initial risk state vector are defined as components of the insurance history risk dimension. This component fully inherits all historical risk characteristics of the customer from the start of insurance purchase, serving as a static basis for trust assessment.
[0090] Secondly, a set of feature values of the claims event is defined as the current claims event dimension component. This component represents the immediate risk characteristics of the current claims event itself, serving as a dynamic input for trust assessment.
[0091] Finally, the risk dimension component of the insurance history and the current claim event dimension component are combined to form the trust state vector.
[0092] In mathematical implementation, this combination process can be represented as a vector concatenation operation. Assuming the retrieved initial risk state vector is R, which is an n-dimensional vector, it can be expressed as R = [r1, r2, ..., r...]. n ] T This represents the risk dimension component of the insurance history; the newly generated claim event feature value vector is C, which is an m-dimensional vector, and can be represented as C = [c1, c2, ..., c...]. m ] T This represents the current claim event dimension component. The initialized trust state vector T0 can then be constructed using the following formula:
[0093]
[0094] In this way, the dimension of the initialized trust state vector T0 is expanded to n+m, and the information it carries becomes more comprehensive, laying a solid mathematical foundation for subsequent dynamic updates and evolution based on state space equations within the claims period.
[0095] S3. During the processing cycle of the online claim application, when updated information related to the online claim application is received, the trust state vector is updated based on the state space equation.
[0096] In this embodiment, the core purpose of this step is to give the trust assessment model a dynamic evolution capability, so that it can adjust its assessment results in real time and iteratively as the information on claims cases becomes more complete.
[0097] Specifically, throughout the entire processing cycle of a claim case—from the initial submission of the claim application to its eventual closure—the processing workflow continuously receives updated information associated with the online claim application. This updated information can take various forms, including but not limited to: accident scene photos, detailed vehicle damage diagrams, medical diagnostic reports, expense receipts, damage assessment reports or repair quotations issued by repair shops, or liability determination letters or assessment conclusions issued by authoritative third-party institutions (such as traffic management departments or medical appraisal institutions).
[0098] Whenever a new update is received, the processing flow initiates an iteration of trust state vector update. First, the update information needs to be quantified to generate an update event vector U. k This step aims to transform newly received, multimodal, and often unstructured information into a standardized numerical vector usable by a mathematical model. For example, an official liability determination clearly stating that the insured party is not at fault can be vectorized into a vector with a strong positive impact on trust levels; conversely, a vague repair list with a price significantly higher than the fair market value can be vectorized into a vector with a negative impact. This vectorization process ensures that all external input information can be uniformly incorporated into subsequent calculations.
[0099] Subsequently, the processing flow applies a preset state transition matrix A and control input matrix B to calculate the updated trust state vector T through state-space equations. k The state-space equation is the core mathematical model for dynamic evaluation in this invention, and it is an effective tool for describing the evolution of dynamic processes.
[0100] The calculation formula for the state-space equation is as follows:
[0101] T k =A·T k-1 +B·U k ;
[0102] In the formula, T k-1 Let T be the trust state vector before the update. During the first iteration update, T... k-1 This refers to the trust state vector T0 generated during the initialization in the previous stage (S2). In subsequent iterations, it represents the evaluation result of the previous time step (k-1).
[0103] A is a predefined state transition matrix. This matrix describes the inherent evolutionary relationships between the components of the trust state vector, independent of external new information input. In one specific implementation, the predefined state transition matrix A can be a square matrix close to the identity matrix, indicating that the trust state has high persistence and its own changes are minimal in the absence of new information shocks. In another implementation, A can also contain off-diagonal elements to represent the intrinsic correlations and mutual influences between the dimensions of the trust state, or simulate the natural decay effect of trust by setting diagonal elements less than 1. This matrix is predefined based on extensive historical data analysis or expert experience.
[0104] B is a preset control input matrix. The function of this matrix is to convert the external input update event vector U... k The effect is mapped to the trust state vector T.k On the corresponding dimensions, it defines how each type of updated information affects each dimension of the trust status and the weight of that impact. For example, a column in the pre-defined control input matrix B might specifically define the impact coefficient of information such as "official liability determination letter" on trust dimensions such as "fraud risk component" and "liability ratio component." Therefore, the pre-defined control input matrix B essentially encapsulates the business rules and knowledge about how external evidence affects trust judgments.
[0105] U k This refers to the update event vector generated after the aforementioned numerical processing of the new update information.
[0106] T k This is the updated trust state vector calculated in this step. It incorporates the state from the previous time step and the influence of the latest information, representing the most accurate trust assessment for the claim at the current time step (k).
[0107] Through this iterative update mechanism, the trust state vector is no longer a static, one-time evaluation result, but a dynamic variable that constantly approaches the real situation as the chain of evidence becomes richer, providing a solid and dynamically updated foundation for the subsequent accurate calculation of the trust scalar and the prepayment amount.
[0108] S4. Based on the updated trust state vector, calculate the trust scalar corresponding to the claim application;
[0109] In this embodiment, the purpose of this step is to transform the multi-dimensional, mathematical trust state vector generated and dynamically maintained in the previous stage into a single, intuitive, and directly applicable quantitative indicator for subsequent automated decision support. This transformation process is essentially an information dimensionality reduction and weighted aggregation process, aiming to derive a final score that comprehensively reflects the overall credibility of the case.
[0110] To achieve this transformation, the processing flow first needs to retrieve the preset weight vector corresponding to the updated trust state vector dimension. The preset weight vector is a coefficient vector with the same dimension as the trust state vector, where each element represents the contribution or importance of the corresponding dimension component in the trust state vector to the final trust assessment result. The preset weight vector is pre-defined and embedded in the processing flow based on statistical analysis and machine learning modeling of massive amounts of historical claims data (e.g., obtaining the coefficients of each feature through training a logistic regression model), or in conjunction with a prior knowledge base of insurance claims experts. This ensures that the calculation of the trust scalar has an objective basis and business interpretability.
[0111] After obtaining the preset weight vector, the processing flow begins to execute the core calculation of mapping the multi-dimensional trust state vector to a single trust scalar. This calculation process specifically includes:
[0112] First, each element in the updated trust state vector is multiplied by its corresponding element in the weight vector to generate multiple product values. This element-wise multiplication aims to weight each dimension component in the trust state vector according to its pre-defined importance. Each weight value in the weight vector quantifies the contribution of its corresponding component (such as historical claim frequency, timeliness of reporting, etc.) to the final trust assessment. The product value obtained after multiplication can be considered as the contribution score of that single dimension component to the overall trust in the current state. For example, multiplying a feature value representing high risk by a negative weight will result in a negative contribution score, and vice versa.
[0113] Next, the processing flow sums the multiple product values and uses the sum as the trust scalar corresponding to the claim application. Essentially, this summation operation linearly combines the independent contribution scores of all dimensions to form a single indicator that represents the overall trust level of the current claim case.
[0114] Mathematically, the aforementioned calculation process is fully represented as the dot product of two vectors. The trust scalar S calculated at time k... k It can be given by the following formula:
[0115]
[0116] In the formula, T k This represents the trust state vector updated by the state-space equations at time k. It is a p-dimensional column vector of the form T. k = [t1, t2, ..., t p ] T In the formula, k represents the time steps or iterations within the claims processing cycle, k = 0 indicates the initial state, and k ≥ 1 indicates the state after at least one update; p is the total dimension of the trust state vector, which is equal to the sum of the dimension n of the insured historical risk dimension component and the dimension m of the current claims event dimension component (i.e., p = n + m); t i (where i = 1, 2, ..., p) is the i-th element of the vector, which is a specific numerical value representing the value of an independent trust or risk dimension component, such as a quantified credit score, a timely reporting score, or a vehicle risk coefficient.
[0117] W represents a predefined weight vector that matches the dimension of the trust state vector. It is also a p-dimensional column vector, in the form W = [w1, w2, ..., w...]. p ]T w i (where i = 1, 2, ..., p) is the i-th weight coefficient of the vector, and t i One-to-one correspondence. This coefficient is a preset real number, the magnitude of which represents the degree of influence of the i-th dimension component on the final trust scalar, and its positive or negative sign indicates the direction of influence (positive or negative).
[0118] S k This represents the trust scalar calculated at time k.
[0119] W T ·T k Represents the transpose of vector W and vector T k The dot product operation.
[0120] It is the expanded form of the dot product operation, which explicitly indicates that corresponding elements of the two vectors are multiplied one by one, and then all products (i.e., w1t1, w2t2, ..., w...) are calculated. p t p The process of summing up.
[0121] The trust scalar S obtained through this calculation k It is a dimensionless value that incorporates the impact of a customer's historical risks, current claim event characteristics, and all updated information within the claim period, providing a quantitative snapshot of the "credibility" of the entire claim at the current point in time.
[0122] S5. Based on the trust scalar and the estimated total compensation amount, calculate the prepayment amount corresponding to the claim application through a risk pricing model;
[0123] In this embodiment, this step is the final output of the processing method. It aims to transform the abstract trust level dynamically evaluated in the previous steps into a specific and automated financial decision, namely, to determine a prepayment amount that can both improve customer experience and ensure fund security under the premise of controllable risk.
[0124] This calculation process specifically includes the following steps:
[0125] First, the processing flow converts the trust scalar into an upfront payment ratio using the risk pricing model. The core of this process is the risk pricing model, which is essentially a pre-defined function that technically defines the mapping relationship between the trust scalar and the upfront payment ratio. The model's design follows a core principle: the upfront payment ratio increases as the trust scalar increases. This ensures that the more credible the claim, the faster and at a higher percentage of the customer receives the upfront payment.
[0126] In a preferred embodiment, to make the mapping relationship more closely resemble the nonlinear risk-return relationship in actual business scenarios, the risk pricing model can be constructed using a Siqmoid-like function. This function can smoothly map a potentially unbounded trust scalar input to a bounded scalar located in [0, P]. max The prepayment ratio for the specified interval is output. Its mathematical expression can be defined as:
[0127]
[0128] In the formula, P k This represents the prepayment ratio calculated by the model at time k, which is between 0 and P. max The values between; f(·) represents the function itself of the risk pricing model; S k P represents the trust scalar calculated in the previous stage at time k; max α is a preset maximum prepayment ratio, with a value ranging from (0,1], for example, it can be set to 0.9 or 1.0. This parameter is the upper limit of the business, defining the maximum proportion of the estimated total compensation that the prepayment can account for under the most ideal trust conditions; α is a preset gain or scaling parameter (α>0), which controls the steepness of the S-curve. A larger α value means that the prepayment ratio is more sensitive to changes in the trust scalar, that is, the prepayment ratio will change drastically within a small range of changes in the trust scalar. Conversely, a smaller α value represents a smoother transition; β is a preset offset or threshold parameter.
[0129] The above P max The three model parameters, α, β, can all be statistically calibrated based on historical claims data, or set by risk control strategy experts, in order to achieve differentiated risk pricing for different product lines and different customer groups.
[0130] The prepayment ratio P was calculated using a risk pricing model. k Then, the processing flow performs the final calculation step: multiplying the estimated total compensation amount by the prepayment ratio to calculate the prepayment amount corresponding to the claim application.
[0131] The calculation formula is as follows:
[0132] Adv k =L est ×P k ;
[0133] In the formula, Adv k L represents the prepayment amount calculated at time k. estThis represents the estimated total compensation for the current case. This figure can be derived from the claims adjuster's initial damage assessment, the damage assessment report from the cooperating repair shop, or other valid loss assessment methods.
[0134] In this way, the processing method transforms a dynamic, multi-dimensional trust assessment result into a specific prepayment amount in a precise and automated manner, achieving a closed-loop process from risk quantification to financial decision-making.
[0135] S6. In response to the policyholder's confirmation instruction, transfer the prepayment amount to the financial account designated by the policyholder;
[0136] In this embodiment, this step constitutes the final execution stage of the entire automated claims process. Its core purpose is to safely, accurately, and efficiently transform the financial decisions calculated in the preceding steps into actual fund transfer actions, thereby fulfilling the prepayment commitment to the customer.
[0137] Specifically, this execution process is accomplished through a series of well-defined interaction and processing steps:
[0138] First, after the processing flow calculates the exact prepayment amount using a risk pricing model, payment is not executed immediately. Instead, a crucial user confirmation step is initiated. During this step, the processing flow presents the policyholder with a payment confirmation interface that includes the prepayment amount and their financial account information.
[0139] In a preferred embodiment, the interface is pushed to the online service terminal used by the policyholder, such as an application on their mobile device or their personal account page on a portal website. The interface clearly lists the specific amount to be prepaid and displays pre-set financial account information (such as bank name, account holder name, and account number), which can be retrieved in advance from the policy or user profile. This step is designed to provide a clear human-computer interaction verification point, giving the policyholder the right to final review and confirmation before fund transfer, ensuring the authenticity of the payment intention and the accuracy of the receiving information.
[0140] Next, the processing flow waits for and receives the confirmation instruction triggered by the insured party on the payment confirmation interface. After verifying that the information is correct, the insured party can issue the confirmation instruction by clicking the confirmation button, entering the payment password, or using biometric identification (such as fingerprint or facial recognition). From a technical perspective, this instruction is a digital signal generated by the client with the user's explicit authorization intent. Its receipt is the only and necessary prerequisite for initiating all subsequent automated payment processes, thereby ensuring the compliance and non-repudiation of the operation.
[0141] Upon successful receipt of the confirmation instruction, the backend of the processing flow will immediately respond to the confirmation instruction and generate a payment instruction including the prepayment amount and the financial account information. This payment instruction is an internal, structured data object that encapsulates the amount to be paid and core payment elements such as the target account. Preferably, the instruction may also include a unique identifier for this claim, a timestamp, and a transaction serial number for verification and tracking, forming a complete, self-contained payment request.
[0142] Finally, the processing flow sends the generated payment instruction to the payment processing system to transfer the prepayment amount to the financial account, thereby processing the prepayment for the online claim. The payment processing system can be an internal financial clearing center or an external third-party payment gateway or direct bank connection channel connected via a secure application programming interface (API). The sending of the payment instruction is completed through an encrypted channel to ensure the secure transmission of financial information. After receiving and verifying the validity of the instruction, the payment processing system executes the actual fund transfer operation.
[0143] Through the aforementioned series of interconnected steps, this processing method not only achieves a high degree of automation in prepayment payments, but also ensures the security, accuracy, and user's right to know throughout the entire fund transfer process through confirmation interactions at key nodes, ultimately completing a closed-loop operation for integrated online insurance application and claims processing.
[0144] The online insurance application and claims processing system described below can be referred to in conjunction with the online insurance application and claims processing method described above.
[0145] Please see the appendix Figure 4 The present invention also provides an integrated online insurance application and claims processing system, the system comprising:
[0146] The risk vector establishment module is used to receive the insurance policy generated from the online insurance application and establish an initial risk state vector for the insurance application.
[0147] The trust vector initialization module is used to initialize the trust state vector corresponding to the claim application based on the initial risk state vector when an online claim application initiated by the policy generated based on the insurance application is received.
[0148] The trust vector update module is used to update the trust state vector based on the state space equation when it receives update information related to the online claim during the processing cycle of the online claim application.
[0149] The trust scalar calculation module is used to calculate the trust scalar corresponding to the claim application based on the updated trust state vector.
[0150] The advance payment calculation module is used to calculate the advance payment amount corresponding to the claim application based on the trust scalar and the estimated total compensation amount through a risk pricing model.
[0151] The prepayment processing module is used to transfer the prepayment amount to the financial account designated by the insured in response to the policyholder's confirmation instruction.
[0152] The system in this embodiment can be used to execute the above method embodiments, and its principle and technical effect are similar, so they will not be described again here.
[0153] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.
Claims
1. A method for integrated online insurance application and claims processing, characterized in that, Includes the following steps: S1. Receive the insurance policy generated from the online insurance application, and establish an initial risk state vector for the insurance application; S2. When an online claim application is received based on the policy generated from the insurance application, the trust state vector corresponding to the claim application is initialized based on the initial risk state vector. S3. During the processing cycle of the online claim application, when updated information related to the online claim application is received, the trust state vector is updated based on the state space equation. S4. Based on the updated trust state vector, calculate the trust scalar corresponding to the claim application; S5. Based on the trust scalar and the estimated total compensation amount, calculate the prepayment amount corresponding to the claim application through a risk pricing model; S6. In response to the policyholder's confirmation instruction, transfer the prepayment amount to the financial account designated by the policyholder.
2. The integrated online insurance application and claims processing method according to claim 1, characterized in that, Step S1, which involves receiving the insurance policy generated from the online insurance application and establishing an initial risk state vector for the application, includes: Receive and process online insurance application data, including insurance information, to generate policies and store them in the policy database; Retrieves the structured policyholder information, risk information, and underwriting terms corresponding to the policy from the policy database; Using the unique identifier in the policyholder's information as the query index, historical risk data associated with the policyholder is retrieved from the historical database. The historical database is a risk information database, which stores the policyholder's historical insurance records, historical claims records, and external credit data. The information of the insured party, the information of the risk object, the underwriting terms and historical risk data are mapped in a preset numerical manner to generate a set of multi-dimensional risk feature values; A set of multi-dimensional risk feature values are combined into a vector to construct the initial risk state vector.
3. The integrated online insurance application and claims processing method according to claim 2, characterized in that, The insured party information includes: insured party identification information, historical claims records, and third-party credit scores; The information regarding the risk object includes: vehicle model, the insured's health declaration, and the location of the property; The terms and conditions of coverage include: scope of insurance liability, exclusions, deductibles, and payout limits.
4. The integrated online insurance application and claims processing method according to claim 1, characterized in that, In step S2, when an online claim application is received based on the policy generated from the insurance application, the step of initializing the trust state vector corresponding to the claim application based on the initial risk state vector includes: When an online claim application is received based on an insurance policy generated from an insurance application, the policy identifier and claim event information are parsed from the online claim application. Based on the policy identifier, retrieve the initial risk state vector corresponding to the policy from the storage medium; The claims event information is subjected to feature extraction and numerical processing to generate a set of claims event feature values; The elements within the initial risk state vector, combined with the set of claim event feature values, are used to construct the trust state vector.
5. The integrated online insurance application and claims processing method according to claim 4, characterized in that, The steps of using the elements within the initial risk state vector and combining them with the set of claim event feature values to construct the trust state vector include: The elements in the initial risk state vector are defined as components of the insured historical risk dimension. Define the set of claim event feature values as the current claim event dimension components; The risk dimension component of the insurance history is combined with the current claim event dimension component to form the trust state vector.
6. The integrated online insurance application and claims processing method according to claim 1, characterized in that, In step S3, during the processing cycle of the online claim application, when updated information related to the online claim application is received, the step of updating the trust state vector based on the state space equation includes: During the processing period of an online claim application, receive updated information associated with the online claim application; The updated information is quantified to generate an update event vector U. k ; Using a preset state transition matrix A and control input matrix B, the updated trust state vector T is calculated through state-space equations. k ; The formula for calculating the state-space equation is as follows: T k =A·T k-1 +B·U k ; In the formula, T k-1 This is the trust state vector before the update.
7. The integrated online insurance application and claims processing method according to claim 1, characterized in that, In step S4, the step of calculating the trust scalar corresponding to the claim application based on the updated trust state vector includes: Retrieve the preset weight vector corresponding to the updated trust state vector dimension; Each element in the updated trust state vector is multiplied by the corresponding element in the weight vector to generate multiple product values. The multiple product values are summed, and the summation result is used as the trust scalar corresponding to the claim application.
8. The integrated online insurance application and claims processing method according to claim 1, characterized in that, Step S5, which involves calculating the prepayment amount corresponding to the claim application based on the trust scalar and the estimated total compensation amount using a risk pricing model, includes: The risk pricing model converts the trust scalar into a prepayment ratio. The risk pricing model is a preset function used to define the mapping relationship between the trust scalar and the prepayment ratio, and the prepayment ratio increases as the trust scalar increases. Multiply the estimated total compensation amount by the prepayment ratio to calculate the prepayment amount corresponding to the claim application; The risk pricing model is as follows: In the formula, P k This represents the prepayment ratio calculated by the model at time k, where 0 ≤ P k ≤P max S k P represents the trust scalar calculated in the previous stage at time k; max For a preset maximum prepayment ratio, 0 <P max ≤1; α is a preset gain or scaling parameter, α>0; β is a preset offset or threshold parameter, P max The three model parameters, α, β, and β, can all be statistically calibrated based on historical claims data or set by risk control strategy experts to provide differentiated risk pricing for different product lines and customer groups.
9. The integrated online insurance application and claims processing method according to claim 1, characterized in that, In step S6, the step of transferring the prepayment amount to the financial account designated by the policyholder in response to the policyholder's confirmation instruction includes: Present the insured party with a payment confirmation interface that includes the prepayment amount and financial account information; Receive the confirmation instruction triggered by the insured party on the payment confirmation interface; In response to the confirmation instruction, a payment instruction including the prepayment amount and the financial account information is generated; The payment instruction is sent to the payment processing system to transfer the prepayment amount to the financial account, thereby processing the prepayment for the online claim application.
10. An integrated online insurance application and claims processing system, applied to the integrated online insurance application and claims processing method as described in any one of claims 1-9, characterized in that, The system includes: The risk vector establishment module is used to receive the insurance policy generated from the online insurance application and establish an initial risk state vector for the insurance application. The trust vector initialization module is used to initialize the trust state vector corresponding to the claim application based on the initial risk state vector when an online claim application initiated by the policy generated based on the insurance application is received. The trust vector update module is used to update the trust state vector based on the state space equation when it receives update information related to the online claim during the processing cycle of the online claim application. The trust scalar calculation module is used to calculate the trust scalar corresponding to the claim application based on the updated trust state vector. The advance payment calculation module is used to calculate the advance payment amount corresponding to the claim application based on the trust scalar and the estimated total compensation amount through a risk pricing model. The prepayment processing module is used to transfer the prepayment amount to the financial account designated by the insured in response to the policyholder's confirmation instruction.