Method, system, electronic device and storage medium for overdue analysis of accounts receivable

Abstract

The application relates to a method, system, electronic device and storage medium for overdue analysis of accounts receivable, the method comprising: parsing overdue data of accounts receivable in a first year and overdue data of accounts receivable in a second year into first structured key-value pair data and second structured key-value pair data in an accounts receivable intelligent agent; automatically aligning the first structured key-value pair data and the second structured key-value pair data by the accounts receivable intelligent agent, and calculating statistical indicators of the first year and the second year; obtaining and parsing a natural language query sentence requested by a user in the accounts receivable intelligent agent, determining a type by keyword matching of the query sentence; and generating an overdue analysis report of accounts receivable by the accounts receivable intelligent agent according to the type and based on the statistical indicators of the first year and the second year. The data of heterogeneous tables in different years are aligned by the intelligent agent, the heterogeneous data are integrated, and the efficiency and accuracy of analysis are effectively improved.

Description

Technical Field

[0001] This application relates to the field of intelligent agent data processing technology, and in particular to methods, systems, electronic devices, and storage media for analyzing overdue accounts receivable. Background Technology

[0002] Overdue accounts receivable analysis refers to the process of statistically analyzing, classifying, and risk-assessing accounts receivable that are due but not yet collected. By analyzing overdue accounts receivable data, companies can promptly identify collection risks, optimize collection strategies, and ensure cash flow security.

[0003] Currently, corporate finance personnel typically rely on existing financial systems for accounts receivable management. This requires them to first export overdue data from the first and second years into two separate tables. However, the table structures and data formats for the first and second years are inconsistent, and the data definitions are not uniform, making direct year-on-year comparisons impossible. Therefore, finance personnel must manually compile analysis reports based on these two tables, requiring significant manual adjustments and resulting in low efficiency and accuracy in accounts receivable overdue analysis.

[0004] There is currently no effective solution to the problem of low efficiency and accuracy in the analysis of overdue accounts receivable in related technologies. Summary of the Invention

[0005] This embodiment provides a method, system, electronic device, and storage medium for analyzing overdue accounts receivable, in order to solve the problems of low efficiency and accuracy in the analysis of overdue accounts receivable in related technologies.

[0006] Firstly, this embodiment provides a method for analyzing overdue accounts receivable, including:

[0007] The overdue accounts receivable data for the first year are parsed into first structured key-value pair data in the accounts receivable intelligent agent, and the overdue accounts receivable data for the second year are parsed into second structured key-value pair data in the same agent; wherein the first year and the second year are two adjacent calendar years; the accounts receivable intelligent agent is used to perform data parsing, automatic alignment, data calculation, intent recognition, and report generation;

[0008] The accounts receivable intelligent agent automatically aligns the first structured key-value pair data and the second structured key-value pair data, and calculates the accounts receivable overdue statistical indicators for the first year and the accounts receivable overdue statistical indicators for the second year, respectively.

[0009] The system acquires and parses the natural language query statements of user requests in the accounts receivable intelligent agent, and determines the type of user request by performing keyword matching on the natural language query statements.

[0010] Based on the aforementioned type and the overdue accounts receivable statistics for the first and second years, an overdue accounts receivable analysis report is generated through the accounts receivable agent.

[0011] In some embodiments, parsing the overdue accounts receivable data for the first year into a first structured key-value pair data in the accounts receivable intelligent agent, and parsing the overdue accounts receivable data for the second year into a second structured key-value pair data in the accounts receivable intelligent agent, includes:

[0012] Obtain a spreadsheet file containing overdue accounts receivable data for the first year and overdue accounts receivable data for the second year;

[0013] The spreadsheet file is parsed using the spreadsheet tool plugin in the accounts receivable agent to obtain the first structured key-value pair data and the second structured key-value pair data.

[0014] In some embodiments, the automatic alignment of the first structured key-value pair data and the second structured key-value pair data by the accounts receivable intelligent agent, and the calculation of accounts receivable overdue statistical indicators for the first year and the second year respectively, includes:

[0015] Perform column name mapping, unit conversion, company name normalization, and aging consolidation operations on the second structured key-value pair data to achieve automatic alignment between the first and second structured key-value pair data.

[0016] Based on the first structured key-value pair data, calculate the overdue accounts receivable statistics for the first year;

[0017] Based on the automatically aligned second structured key-value pair data, calculate the accounts receivable overdue statistics for the second year;

[0018] Based on the overdue accounts receivable statistics for the first year and the overdue accounts receivable statistics for the second year, calculate the year-on-year change rate between the first and second years.

[0019] In some embodiments, performing column name mapping, unit conversion, company name normalization, and aging consolidation operations on the second structured key-value pair data to achieve automatic alignment of the first and second structured key-value pair data includes:

[0020] By matching semantic similarity, the original column names of the second year are mapped to the standard column names of the first year;

[0021] By using a preset conversion factor, the amount in the second year is scaled to the same unit as in the first year;

[0022] By using a fuzzy matching algorithm, the company name for the second year is standardized to be the same as the name for the first year;

[0023] By merging intervals, multiple aging segments of accounts receivable in the second year are merged into the same aging segment as in the first year;

[0024] Using the structure of the first structured key-value pair data as a benchmark, the second structured key-value pair data is automatically aligned.

[0025] In some embodiments, the step of acquiring and parsing the natural language query statement of the user request in the accounts receivable intelligent agent, and determining the type of the user request by performing keyword matching on the natural language query statement, includes:

[0026] Obtain the natural language query statement requested by the user;

[0027] The natural language query statement is segmented to obtain keywords;

[0028] The keywords are semantically analyzed and compared with preset keywords to obtain matching results; the preset keywords have a corresponding relationship with the type.

[0029] The type of the user request is determined based on the matching result, and the request is routed to the corresponding path in the accounts receivable agent according to the type.

[0030] In some embodiments, the user request type is determined based on the matching result, and traffic is routed to the corresponding path in the accounts receivable agent according to the type; the type includes: mixed type, chart type, single company type, or company-wide report type; wherein:

[0031] When the natural language query statement contains image and text keywords, the type is determined to be a mixed type, and the query is routed to a mixed path;

[0032] When the natural language query contains chart-related keywords, the type is determined to be chart type, and the query is routed to the chart path;

[0033] When the natural language query contains company name keywords, the type is determined to be a single company type, and the query is routed to the single company path.

[0034] When the natural language query does not contain image / text keywords, chart / graph keywords, or company name keywords, the type is determined to be the company-wide type, and the query is routed to the company-wide reporting path.

[0035] In some embodiments, generating an accounts receivable overdue analysis report through the accounts receivable agent based on the type and the accounts receivable overdue statistical indicators for the first year and the second year includes:

[0036] When the type is either company-wide or single-company, a structured markup-formatted analysis report text is generated using a large language model, and the analysis report text is then converted into a document format for output.

[0037] When the type is a chart type, the chart engine is called to generate a visual chart based on the overdue accounts receivable statistics;

[0038] When the type is a mixed type, the analysis report text and chart configuration information in a structured markup format are generated synchronously in the same call process through the large language model, and the chart configuration information is embedded in the analysis report text for output.

[0039] Secondly, this embodiment provides an accounts receivable overdue analysis system, including: a terminal device, a transmission device, and a server device; wherein the terminal device is connected to the server device through the transmission device;

[0040] The terminal device is used to acquire files uploaded by users that contain overdue accounts receivable data for the first year and overdue accounts receivable data for the second year, as well as to acquire natural language query statements requested by users.

[0041] The transmission device is used to transmit the file and the natural language query statement from the terminal device to the server device, and to transmit the overdue accounts receivable analysis report generated by the server device back from the server device to the terminal device;

[0042] The server device is used to execute the accounts receivable overdue analysis method described in any of the first aspects.

[0043] Thirdly, this embodiment provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the accounts receivable overdue analysis method described in the first aspect above.

[0044] Fourthly, this embodiment provides a storage medium storing a computer program that, when executed by a processor, implements the steps of the accounts receivable overdue analysis method described in the first aspect above.

[0045] Compared with related technologies, the accounts receivable overdue analysis method, system, electronic device and storage medium provided in this embodiment are superior. This method parses the overdue accounts receivable data for the first year into a first structured key-value pair data in the accounts receivable intelligent agent, and parses the overdue accounts receivable data for the second year into a second structured key-value pair data in the same agent; where the first and second years are two adjacent calendar years; the accounts receivable intelligent agent is used to perform data parsing, automatic alignment, data calculation, intent recognition, and report generation; the accounts receivable intelligent agent automatically aligns the first and second structured key-value pair data, and calculates the overdue accounts receivable statistical indicators for the first and second years respectively; it obtains and parses the natural language query statements of user requests in the accounts receivable intelligent agent, and determines the type of user request by performing keyword matching on the natural language query statements; based on the type and the overdue accounts receivable statistical indicators for the first and second years, the accounts receivable intelligent agent generates an overdue accounts receivable analysis report. It automatically aligns data from heterogeneous tables across different years using an accounts receivable intelligence agent, integrates heterogeneous data, and enables multi-dimensional display of accounts receivable overdue analysis results through keyword-based traffic diversion, thereby effectively improving the efficiency and accuracy of accounts receivable overdue analysis.

[0046] Details of one or more embodiments of this application are set forth in the following drawings and description to make other features, objects and advantages of this application more readily apparent. Attached Figure Description

[0047] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0048] Figure 1 This is a hardware structure block diagram of a terminal for an accounts receivable overdue analysis method according to an embodiment of this application;

[0049] Figure 2 This is a flowchart of an embodiment of the accounts receivable overdue analysis method of this application;

[0050] Figure 3 This is a flowchart of an accounts receivable overdue analysis method in an accounts receivable smart agent according to an embodiment of this application. Detailed Implementation

[0051] To better understand the purpose, technical solution, and advantages of this application, the application is described and illustrated below in conjunction with the accompanying drawings and embodiments.

[0052] Unless otherwise defined, the technical or scientific terms used in this application shall have the general meaning understood by one of ordinary skill in the art to which this application pertains. Words such as “a,” “an,” “an,” “the,” “the,” and “these” used in this application do not indicate quantitative limitation and may be singular or plural. The terms “comprising,” “including,” “having,” and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that comprises a series of steps or modules (units) is not limited to the listed steps or modules (units) but may include steps or modules (units) not listed, or may include other steps or modules (units) inherent to these processes, methods, products, or devices. Words such as “connected,” “linked,” and “coupled” used in this application are not limited to physical or mechanical connections but may include electrical connections, whether direct or indirect. “Multiple” used in this application refers to two or more. “And / or” describes the relationship between related objects, indicating that three relationships may exist; for example, “A and / or B” can represent: A alone, A and B simultaneously, and B alone. Normally, the character " / " indicates that the objects before and after it are in an "or" relationship. The terms "first," "second," "third," etc., used in this application are merely to distinguish similar objects and do not represent a specific order of objects.

[0053] The method embodiments provided in this example can be executed on a terminal, computer, or similar electronic device with a certain computing power. For example, it can run on a terminal. Figure 1 This is a hardware structure block diagram of a terminal for an accounts receivable overdue analysis method according to an embodiment of this application. For example... Figure 1 As shown, a terminal may include one or more ( Figure 1 Only one is shown in the diagram. A processor 102 and a memory 104 for storing data are also included. The processor 102 may be, but is not limited to, a microprocessor (MCU) or a programmable logic device (FPGA). The terminal may also include a transmission device 106 for communication functions and an input / output device 108. Those skilled in the art will understand that… Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the terminal described above. For example, the terminal may also include components that are larger than... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown are illustrated.

[0054] The memory 104 can be used to store computer programs, such as application software programs and modules, like the computer program corresponding to the accounts receivable overdue analysis method in this embodiment. The processor 102 executes various functional applications and data processing by running the computer programs stored in the memory 104, thereby implementing the above-described method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0055] The transmission device 106 is used to receive or send data via a network. This network includes a wireless network provided by the terminal's communication provider. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 can be a Radio Frequency (RF) module used for wireless communication with the Internet.

[0056] This embodiment provides a method for analyzing overdue accounts receivable. Figure 2 This is a flowchart of an embodiment of the accounts receivable overdue analysis method of this application, such as... Figure 2 As shown, the process includes the following steps:

[0057] Step S210: Parse the overdue accounts receivable data for the first year into a first structured key-value pair data in the accounts receivable intelligent agent, and parse the overdue accounts receivable data for the second year into a second structured key-value pair data in the accounts receivable intelligent agent; wherein the first year and the second year are two adjacent natural years; the accounts receivable intelligent agent is used to perform data parsing, automatic alignment, data calculation, intent recognition and report generation.

[0058] Specifically, this involves exporting report files from a company's financial system or enterprise resource planning (ERP) system to obtain files containing overdue accounts receivable data for the first and second years. The files may include information such as customer or company names, accounts receivable amounts for different months within the year, aging intervals, transaction dates, and overdue status. The report files can be in the form of spreadsheets, comma-separated value files, text files, or database export files. For example, an Excel-formatted overdue data report can be exported from the company's financial system or ERP system, with the first year's overdue accounts receivable data placed in one sheet and the second year's overdue accounts receivable data placed in another sheet.

[0059] After obtaining the file containing overdue accounts receivable data for the first and second years, the file is input into the accounts receivable agent. For spreadsheet files, the accounts receivable agent can use a spreadsheet toolbox plugin to parse the file, identify the file format, extract the data content, and convert the unstructured tabular data into structured key-value pair data, such as JavaScript Object Notation (JSON). The first and second years refer to two adjacent calendar years; the first year is the base year, such as 2025, and the second year is the adjacent calendar year, such as 2026.

[0060] The Accounts Receivable Agent is an autonomous task execution unit driven by a large language model. It is deployed in a low-code workflow engine and is equipped with a spreadsheet toolbox plugin, a data alignment plugin, a large language model, a document conversion plugin, and a chart engine. It is used to autonomously perform data parsing, difference identification and automatic alignment, overdue statistical indicator calculation, intent recognition, and multimodal report generation.

[0061] Step S220: The first structured key-value pair data and the second structured key-value pair data are automatically aligned by the accounts receivable intelligent agent, and the accounts receivable overdue statistical indicators for the first year and the second year are calculated respectively.

[0062] First, the accounts receivable agent identifies the types of differences between the first and second structured key-value pairs of data. These differences include semantic differences in column names, differences in monetary units, differences in aging intervals, and differences in company name formats. Then, based on the identified differences, the accounts receivable agent performs corresponding normalization operations to align the second structured key-value pairs to the data scope of the first structured key-value pairs, creating a consistent benchmark. Next, different code execution nodes are set up within the accounts receivable agent to process the first and second structured key-value pairs separately. Based on the first structured key-value pairs, overdue accounts receivable statistics for the first year are calculated. For example, these statistics include total overdue amount, aging distribution, and overdue amounts aggregated by company or customer. Similarly, overdue accounts receivable statistics for the second year are calculated based on the aligned second structured key-value pairs. Furthermore, after obtaining the accounts receivable overdue statistics for the first year and the accounts receivable overdue statistics for the second year, it is possible to calculate the year-on-year change rate between the first and second years.

[0063] Step S230: Obtain and parse the natural language query statement of the user request in the accounts receivable intelligent agent, and determine the type of user request by performing keyword matching on the natural language query statement.

[0064] Specifically, the accounts receivable intelligent agent acquires the natural language query statement input by the user. This natural language query statement is a request made by the user in everyday natural language, such as "Please generate an annual overdue analysis report" or "Query the overdue status of Company A and generate a chart." The accounts receivable intelligent agent performs word segmentation on the natural language query statement, breaking it down into several independent keyword units. After obtaining the segmentation results, the accounts receivable intelligent agent performs semantic analysis and comparison of the keywords with several preset keyword categories. These preset keyword categories include text-image keywords (such as "text-image," "report with chart"), chart-type keywords (such as "chart," "bar chart," "pie chart," "line chart"), and company name recognition rules (identified by company suffixes such as "company," "group," "limited," etc.). Based on the semantic analysis and comparison results, the agent can then determine the type of user request, which may include mixed types, chart types, single-company types, or company-wide types.

[0065] For example, if the natural language query entered by the user contains the word "chart", then the chart-related keyword is matched, and the user request type is determined to be chart type; if the company name keyword is included, then the type is determined to be single company type; if the "text and image" category keyword is included, then the type is determined to be mixed type; if no preset keyword is matched, then the default type is determined to be full company type.

[0066] Step S240: Based on the type and the accounts receivable overdue statistical indicators for the first and second years, generate an accounts receivable overdue analysis report through the accounts receivable agent.

[0067] The accounts receivable intelligent agent generates a corresponding accounts receivable overdue analysis report based on the determined user request type and the calculated accounts receivable overdue statistics for the first and second years, as well as the year-on-year change rate between the first and second years, and other correlation data. Specifically, when the request type is mixed, the report text and chart configuration information are generated simultaneously during the same call to the large language model within the accounts receivable intelligent agent, and output in a mixed text and chart format. When the request type is chart-based, the report is output in a visual chart format by calling the chart engine within the accounts receivable intelligent agent. When the request type is company-wide or single-company, the report text is output in document format by calling the document conversion plugin within the accounts receivable intelligent agent.

[0068] In the daily financial management practices of enterprises, finance personnel responsible for accounts receivable management generally rely on the company's existing financial information system to conduct statistical work related to overdue accounts. However, due to the functional limitations of the existing financial system, it is usually unable to directly provide overdue data reports that meet the needs of multi-dimensional analysis. Finance personnel first export overdue account details data for two different statistical periods, the first and second years, from the financial system, resulting in two independent data tables. Subsequently, finance personnel need to manually check, organize, and compile accounts receivable analysis reports based on these two tables. However, the data organization structure, field naming rules, units of measurement, and aging criteria used in the first and second year tables often differ significantly, resulting in a lack of consistency and comparability in the data types of the two tables, making them unsuitable for direct year-on-year comparative analysis across years. This requires finance personnel to invest a lot of time and energy in repeatedly adjusting, aligning fields, and validating data in the two tables, directly leading to low overall efficiency and accuracy of accounts receivable overdue analysis.

[0069] Steps S210 to S240 above involve the following steps: First, the overdue accounts receivable data for the first year are parsed into first structured key-value pairs in the accounts receivable intelligent agent, and the overdue accounts receivable data for the second year are parsed into second structured key-value pairs in the same agent; the first and second years are two adjacent calendar years. The accounts receivable intelligent agent is used to perform data parsing, automatic alignment, data calculation, intent recognition, and report generation. Second, the accounts receivable intelligent agent automatically aligns the first and second structured key-value pairs, and calculates the overdue accounts receivable statistical indicators for the first and second years, respectively. Subsequently, the natural language query statements requested by users in the accounts receivable intelligent agent are obtained and parsed, and the type of user request is determined by keyword matching of the natural language query statements. Finally, based on the type and the overdue accounts receivable statistical indicators for the first and second years, an overdue accounts receivable analysis report is generated by the accounts receivable intelligent agent. It automatically aligns data from heterogeneous tables across different years using an accounts receivable intelligence agent, quickly integrates heterogeneous data, and achieves multi-dimensional display of accounts receivable overdue analysis results through keyword-based traffic diversion, effectively improving the efficiency and accuracy of accounts receivable overdue analysis.

[0070] Optionally, in one embodiment, parsing the overdue accounts receivable data for the first year into a first structured key-value pair data in the accounts receivable intelligent agent, and parsing the overdue accounts receivable data for the second year into a second structured key-value pair data in the accounts receivable intelligent agent, includes:

[0071] Obtain a spreadsheet file containing overdue accounts receivable data for the first year and overdue accounts receivable data for the second year; parse the spreadsheet file using the spreadsheet tool plugin in the accounts receivable agent to obtain the first and second structured key-value pairs of data.

[0072] Specifically, the spreadsheet file can be an Excel file (including .xls or .xlsx format) exported by the user from the enterprise's financial system or enterprise resource planning system, or a comma-separated value file or other spreadsheet file with a row and column structure. The spreadsheet file includes the following fields: customer or company name, accounts receivable amount, aging range, transaction date, and overdue status. After obtaining the spreadsheet file, it is input into the accounts receivable intelligent agent. The accounts receivable intelligent agent calls its configured spreadsheet toolbox plugin to parse the spreadsheet file. The spreadsheet toolbox plugin first identifies the format type of the spreadsheet file, then reads the data content in the table row by row and column by column, extracting field names as keys and field values ​​as corresponding values, converting the unstructured tabular data into a structured key-value pair data format. Through the above parsing operation, the first structured key-value pair data corresponding to the first year and the second structured key-value pair data corresponding to the second year are obtained respectively. The first and second structured key-value pair data are saved in JSON format, forming structured data objects that can be automatically aligned and calculated later.

[0073] The spreadsheet tool plugin in the accounts receivable agent automatically parses spreadsheet files, converting unstructured tabular data into structured key-value pairs, providing a data foundation for subsequent alignment and calculations, and avoiding the inefficiency and error-prone problems caused by manual data parsing.

[0074] Furthermore, in one embodiment, the accounts receivable intelligent agent automatically aligns the first structured key-value pair data and the second structured key-value pair data, and calculates the accounts receivable overdue statistical indicators for the first year and the second year, respectively, including:

[0075] Perform column name mapping, unit conversion, company name normalization, and aging consolidation operations on the second structured key-value pair data to automatically align the first and second structured key-value pair data. Based on the first structured key-value pair data, calculate the accounts receivable overdue statistics for the first year. Based on the automatically aligned second structured key-value pair data, calculate the accounts receivable overdue statistics for the second year. Based on the accounts receivable overdue statistics for the first and second years, calculate the year-on-year change rate between the first and second years.

[0076] First, column name mapping, unit conversion, company name normalization, and aging consolidation operations are performed on the second structured key-value pair data. Column name mapping unifies the field identifiers of the first and second structured key-value pair data; unit conversion unifies the units of measurement for amounts; company name normalization unifies the names of customers or companies; and aging consolidation unifies the aging intervals. Second, after processing the second structured key-value pair data, it is aligned with the first structured key-value pair data. Next, after alignment, the first year's overdue accounts receivable statistics are calculated based on the first structured key-value pair data, and the second year's overdue accounts receivable statistics are calculated based on the aligned second structured key-value pair data. These statistics include total overdue amount, aging distribution, and overdue amount at the company or customer level. Finally, based on the first and second year's overdue accounts receivable statistics, the year-on-year change rate between the two years is calculated.

[0077] By automatically performing column name mapping, unit conversion, company name normalization, and aging consolidation operations on the structured data of the second year, the data scope of the first and second years is automatically aligned. Based on the aligned data, the overdue statistical indicators and year-on-year change rates for the two years are calculated respectively. Accurate analysis of cross-year data can be completed without manual calculation, which can significantly improve the efficiency and accuracy of data analysis.

[0078] In one embodiment, column name mapping, unit conversion, company name normalization, and aging consolidation operations are performed on the second structured key-value pair data to achieve automatic alignment of the first and second structured key-value pair data, including:

[0079] By using semantic similarity matching, the original column names of the second year are mapped to the standard column names of the first year; by using a preset conversion factor, the amount values ​​of the second year are scaled to the same unit as the first year; by using a fuzzy matching algorithm, the company names of the second year are normalized to the same standard names as the first year; by using interval accumulation and merging, multiple aging segments of the second year are merged into the same aging interval as the first year; and by using the structure of the first structured key-value pair data as a benchmark, the second structured key-value pair data is automatically aligned.

[0080] First, semantic similarity matching maps the original column names from the second year to the standard column names from the first year. Specifically, all column names in the first structured key-value pair data are extracted to form a standard column name set; all original column names in the second structured key-value pair data are extracted; for each original column name in the second structured key-value pair data, a semantic similarity matching algorithm is used to match its semantic similarity with each standard column name in the standard column name set; the standard column name with the highest semantic similarity exceeding a preset threshold is used as the mapping target for that original column name, and the column data in the second structured key-value pair data is associated with the corresponding standard column name; original column names that cannot be matched are retained or discarded according to preset rules. Through the above column name mapping operation, the column name field identifiers of the second year are kept consistent with those of the first year.

[0081] Secondly, by using a preset conversion factor, the monetary values ​​for the second year are scaled to the same unit as those for the first year. Specifically, the first unit (e.g., "ten thousand yuan") used for monetary values ​​in the first structured key-value pair data is identified; the second unit (e.g., "yuan") used for monetary values ​​in the second structured key-value pair data is identified; it is determined whether the first and second units are consistent; if they are inconsistent, a preset unit conversion factor is obtained, and all monetary values ​​in the second structured key-value pair data are scaled according to the conversion factor, i.e., the converted amount for the second year is the original amount for the second year divided by the conversion factor; the converted amount replaces the original amount in the second structured key-value pair data, so that the monetary unit for the second year is consistent with that for the first year.

[0082] Next, a fuzzy matching algorithm is used to standardize the company names for the second year to match those for the first year. Specifically, all company names in the first structured key-value pair data are extracted to form a standard name set. For each original company name in the second structured key-value pair data, a fuzzy matching algorithm, including edit distance, cosine similarity, and pinyin-based matching algorithms, is used to calculate the string similarity between the original company name and each standard name in the standard name set. The standard name with the highest similarity exceeding a preset threshold is identified as the matching result, and this standard name replaces the original company name in the second structured key-value pair data. For company names that cannot be matched, the original name is retained and marked for manual review. Through this company name standardization process, the name representation of the same customer or company is made consistent across the two years of data.

[0083] Then, by interval summation and merging, multiple aging segments of the second year are merged into the same aging segments as the first year. Optionally, in one embodiment, the first aging segment interval set used by the first structured key-value pair data and the second aging segment interval set used by the second structured key-value pair data are obtained; the second aging segment interval set is compared with the first aging segment interval set to identify adjacent intervals in the second aging segment interval set that are more finely divided than those in the first aging segment interval set; for the identified adjacent intervals that need to be merged, the overdue amounts belonging to these intervals are summed and aggregated by interval summation to generate merged aging segment data; the merged aging segment data is updated in the second structured key-value pair data so that the aging segment intervals of the second year are consistent with those of the first year.

[0084] For example, the first year uses the following four aging intervals: "1-30 days", "31-60 days", "61-90 days", and "over 90 days"; the second year uses the following five aging intervals: "1-30 days", "31-60 days", "61-90 days", "91-120 days", and "over 120 days". A comparison shows that the intervals in the second year are more granular than those in the first year. Therefore, the "91-120 days" and "over 120 days" intervals in the second year are merged into the "over 90 days" interval, and the overdue amounts in the corresponding intervals are summed up to ensure that the merged aging intervals are consistent with those in the first year.

[0085] Finally, using the structure of the first structured key-value pair data as a benchmark, the second structured key-value pair data is automatically aligned. After all the above column name mapping, unit conversion, company name normalization, and aging consolidation operations are completed, the data dimensions and indicators of the second structured key-value pair data are fully aligned with those of the first structured key-value pair data. That is, using the structure of the first structured key-value pair data as a benchmark, the second structured key-value pair data is automatically aligned, forming a unified data benchmark that can be used for year-on-year analysis.

[0086] In one embodiment, the natural language query statement of a user request in the accounts receivable agent is acquired and parsed, and the type of user request is determined by keyword matching of the natural language query statement, including:

[0087] Obtain the natural language query statement requested by the user; perform word segmentation on the natural language query statement to obtain keywords; perform semantic analysis and comparison on the keywords with preset keywords to obtain matching results; the preset keywords have a corresponding relationship with the type; determine the type of the user request based on the matching results, and route the request to the corresponding path in the accounts receivable agent according to the type.

[0088] Specifically, a user may input a query request through an interactive interface in the form of daily natural language, such as "Query the overdue situation of Company A and generate a chart". After obtaining the natural language query sentence, the accounts receivable agent performs preprocessing and word segmentation on it. Preprocessing includes removing stop words (such as "please", "of" and "has / have"), removing special symbols and punctuation marks; word segmentation refers to splitting a continuous query sentence into several independent keyword units, for example, splitting "query the overdue situation of Company A" into keywords such as "query", "Company A", "overdue" and "situation".

[0089] After word segmentation, the accounts receivable agent performs semantic analysis and comparison between the obtained keywords and a plurality of preset keyword categories. The preset keyword categories include image-text keywords, chart keywords and company name identification rules. There is a preset corresponding relationship between the preset keyword categories and the types of user requests. Specifically, image-text keywords correspond to the mixed type, chart keywords correspond to the chart type, company name keywords correspond to the single-company type, and when no preset keyword is matched, it corresponds to the whole-company type.

[0090] Through semantic analysis and comparison, the keyword category matched by the keyword is determined, thereby obtaining a matching result. According to the matching result, the type of the user request is confirmed, and the user request is distributed to the corresponding processing path in the accounts receivable agent according to the type.

[0091] Through word segmentation processing and keyword semantic matching on the user's natural language query sentence, the user's intention is automatically identified and distributed to the corresponding processing path, which realizes natural language interactive intelligent query, lowers the threshold for user operation, and improves the intelligent level of query response.

[0092] In one embodiment, the type of the user request is confirmed according to the matching result, and the user request is distributed to the corresponding path in the accounts receivable agent according to the type; the types include: mixed type, chart type, single-company type or whole-company report type; wherein:

[0093] When the natural language query sentence contains an image-text keyword, the type is determined as the mixed type, and the request is distributed to the mixed path; when the natural language query sentence contains a chart keyword, the type is determined as the chart type, and the request is distributed to the chart path; when the natural language query sentence contains a company name keyword, the type is determined as the single-company type, and the request is distributed to the single-company path; when the natural language query sentence does not contain any image-text keyword, chart keyword or company name keyword, the type is determined as the whole-company type, and the request is distributed to the whole-company report path.

[0094] Specifically, in one embodiment, when a natural language query contains text-image keywords, the accounts receivable agent determines that the user request is a mixed type and routes the request to a mixed processing path. Text-image keywords include phrases such as "text and image," "report plus chart," "text plus image," "rich in both text and images," and "chart + report," indicating a simultaneous need for both text reports and visual charts. For example, when a user enters "Please generate a text-image analysis report for Company A," which contains the keyword "text and image," the type is determined to be mixed, and the request is routed to a mixed path. Subsequently, the large language model will synchronously generate the report text and chart configuration information during the same call process.

[0095] When a natural language query contains chart-related keywords, the accounts receivable agent determines that the user's request is for a chart and routes it to the chart processing path. Chart-related keywords can include words such as "chart," "bar chart," "pie chart," "line chart," "scatter plot," "visualization," and "data graph," indicating that a visual chart needs to be generated. For example, when a user enters "generate a bar chart of overdue amounts for each company," which contains the keyword "bar chart," the query is determined to be for a chart, and the request is routed to the chart path. Subsequently, the chart engine will be invoked to generate the corresponding visual chart.

[0096] When a natural language query contains company name keywords, the accounts receivable agent determines that the user request is a single-company type and routes it to the single-company processing path. Company name keywords include specific company names (such as "Company A" or "Group B") and common company name suffix patterns, such as "Company," "Group," "Limited," "Shareholding," and "Limited Liability," which are identified using regular expressions or fuzzy matching algorithms. For example, when a user enters "Query the overdue status of Company A," which includes the company name "Company A," the type is determined to be single-company, and the request is routed to the single-company path for report generation.

[0097] When the natural language query does not contain image / text keywords, chart / graph keywords, or any company name keywords, the accounts receivable agent determines the user request type to be "company-wide" and routes the request to the company-wide report path. This rule serves as the default matching rule for handling general query requests that do not explicitly specify a query scope. For example, when a user enters "Generate this month's overdue analysis report," which does not contain any of the aforementioned keywords, the type is determined to be "company-wide," and the request is routed to the company-wide report path to generate a company-wide overdue analysis report.

[0098] Based on the above rules, the accounts receivable intelligent agent can automatically identify user intent and route requests to the corresponding processing paths according to the keyword features in the user's natural language query, thereby achieving differentiated report generation.

[0099] Furthermore, in one embodiment, based on the type and the accounts receivable overdue statistics for the first and second years, an accounts receivable overdue analysis report is generated by the accounts receivable agent, including:

[0100] When the type is either company-wide or single-company, the analysis report text in structured markup format is generated through the large language model, and then the analysis report text is converted into document format for output. When the type is chart type, the chart engine is called to generate a visual chart based on the overdue accounts receivable statistics. When the type is mixed, the analysis report text in structured markup format and chart configuration information are generated synchronously in the same call process through the large language model, and the chart configuration information is embedded in the analysis report text for output.

[0101] Specifically, when the user request type is either company-wide or single-company, the accounts receivable agent calls the large language model, takes the accounts receivable overdue statistics for the first year, the accounts receivable overdue statistics for the second year, and the year-on-year change rate between the first and second years as input data, generates a report text in a structured markup format (e.g., Markdown format), and then uses the document conversion plugin in the accounts receivable agent to convert the structured markup format report text into a Word document format for output.

[0102] When the user request type is a chart, the accounts receivable agent determines the chart type and aggregation dimension based on the chart keywords (such as bar chart, pie chart and line chart) in the user request, performs data filtering and aggregation based on the overdue accounts receivable statistics, and then calls the chart engine in the accounts receivable agent to generate a visual chart for output.

[0103] When the user request type is mixed, the accounts receivable agent synchronously generates a structured markup format report text and chart configuration information (such as a JSON format chart configuration object) during the same large language model call process, and embeds the chart configuration information into the specified position of the report text to form a comprehensive analysis report with mixed text and graphics for output.

[0104] By automatically generating analysis reports in document, chart, or mixed format based on user intent, the efficiency of information delivery and the intuitiveness of analysis results can be improved.

[0105] Figure 3 This is a flowchart of an accounts receivable overdue analysis method in an accounts receivable smart agent according to an embodiment of this application. Figure 3 As shown, users input requests via natural language, which are then parsed and calculated for the first year's data, as well as for the second year's data. After processing the data, the accounts receivable agent determines the type of user request, identifying whether it falls under the categories of mixed type, chart type, single-company type, or company-wide report type, and then routes the user request to the corresponding processing path based on the identified request type.

[0106] When the data is identified as a chart type, the process proceeds to the chart path. First, the Cico-LLM large language model is used to identify the chart intent, determining the type of chart the user needs and how to label it. Next, a configuration extraction step is performed to extract the necessary configuration parameters for chart generation. Then, data filtering is conducted, extracting data relevant to chart generation from first and second year data. Further data aggregation is performed, summarizing and formatting the filtered data according to the chart display requirements. Finally, the intelligent chart generator produces the chart report.

[0107] When identified as a single-company type, the system proceeds to the single-company path. First, it filters companies based on the company name keywords in the user request, identifying overdue data for the corresponding companies from both the first and second year data. Then, it performs year-on-year calculations to determine the rate of change between the first and second years. Next, it uses the Cico-LLM large language model to generate a company-level report, generating a structured, marked-up report text based on the filtered data and the year-on-year change rate. Finally, it performs format conversion to generate the company-level report and outputs the single-company report.

[0108] When a mixed type is identified, the process proceeds to the mixed path. First, the Cico-LLM large language model is invoked to obtain the analysis report, which is then used as the text content in the graphic report. Simultaneously, the Cico-LLM large language model is used for chart planning, preprocessing the chart data and outputting multiple charts, including "Chart 1 - Company Comparison," "Chart 2 - Current Year's Aging," "Chart 3 - Last Year's Aging," "Chart 4 - Year-on-Year Change," and "Chart 5 - High Risk." These chart contents are then used as the image content in the graphic report. Finally, the text report and chart content are merged to form a comprehensive analysis report that integrates text and graphics.

[0109] When identified as the "entire company" type, the system enters the "entire company" path. First, it calls the "entire company year-on-year calculation module" to obtain the full statistical indicators of the first and second year's data and calculate the year-on-year change rate. Then, it calls the Cico-LLM large language model to generate a year-on-year analysis report text based on the full data and the year-on-year change rate. Finally, it performs format conversion to generate and output the "entire company" report.

[0110] Through the above process, the accounts receivable intelligent agent can automatically route users to the corresponding processing paths based on their different query intentions, thereby generating differentiated analysis reports.

[0111] This embodiment also provides an accounts receivable overdue analysis system, including: terminal equipment, transmission equipment, and server equipment;

[0112] In this embodiment, the terminal device is connected to the server device via a transmission device; the terminal device is used to obtain files uploaded by the user containing overdue accounts receivable data for the first and second years, as well as to obtain natural language query statements requested by the user; the transmission device is used to transmit the files and natural language query statements from the terminal device to the server device, and to send the overdue accounts receivable analysis report generated by the server device back from the server device to the terminal device; the server device is used to execute the steps in any of the above method embodiments.

[0113] Specifically, the terminal device establishes a communication connection with the server device through a transmission device. The terminal device is used to acquire files uploaded by the user containing overdue accounts receivable data for the first and second years, such as an Excel-formatted overdue data report exported from the company's financial system, and to acquire natural language queries entered by the user, such as "Query the overdue status of Company A and generate a chart." The transmission device is used to transmit the files and natural language queries from the terminal device to the server device, and to send the overdue accounts receivable analysis report generated by the server device back from the server device to the terminal device, thus realizing data interaction between the user terminal and the server. The server device is used to execute the steps in any of the above method embodiments.

[0114] This embodiment also provides an electronic device including a memory and a processor, the memory storing a computer program and the processor being configured to run the computer program to perform the steps in any of the above method embodiments.

[0115] Optionally, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor and the input / output device is connected to the processor.

[0116] Optionally, in this embodiment, the processor can be configured to perform the following steps via a computer program:

[0117] S1, parses the overdue accounts receivable data for the first year into a first structured key-value pair data in the accounts receivable intelligent agent, and parses the overdue accounts receivable data for the second year into a second structured key-value pair data in the accounts receivable intelligent agent; where the first year and the second year are two adjacent calendar years; the accounts receivable intelligent agent is used to perform data parsing, automatic alignment, data calculation, intent recognition, and report generation;

[0118] S2, automatically align the first structured key-value pair data and the second structured key-value pair data through the accounts receivable intelligent agent, and calculate the accounts receivable overdue statistical indicators for the first year and the accounts receivable overdue statistical indicators for the second year respectively;

[0119] S3: Obtain and parse the natural language query statement of the user request in the accounts receivable intelligent agent, and determine the type of user request by performing keyword matching on the natural language query statement;

[0120] S4 generates an accounts receivable overdue analysis report based on the type and the accounts receivable overdue statistics for the first and second years through the accounts receivable agent.

[0121] It should be noted that the specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementations, and will not be repeated in this embodiment.

[0122] Furthermore, in conjunction with the accounts receivable overdue analysis method provided in the above embodiments, this embodiment can also provide a storage medium for implementation. This storage medium stores a computer program; when executed by a processor, the computer program implements any of the accounts receivable overdue analysis methods in the above embodiments.

[0123] It should be understood that the specific embodiments described herein are merely illustrative of the application and not intended to limit it. All other embodiments derived by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application.

[0124] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0125] Obviously, the accompanying drawings are merely some examples or embodiments of this application. Those skilled in the art can apply this application to other similar situations based on these drawings without any creative effort. Furthermore, it is understood that although the work done in this development process may be complex and lengthy, for those skilled in the art, certain design, manufacturing, or production modifications made based on the technical content disclosed in this application are merely conventional technical means and should not be considered as insufficient disclosure of this application.

[0126] The term "embodiment" in this application refers to a specific feature, structure, or characteristic described in connection with an embodiment that may be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily imply the same embodiment, nor does it imply that it is mutually exclusive with or independent of other embodiments. It will be clearly or implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0127] The above embodiments merely illustrate several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of patent protection. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the appended claims.

Claims

1. A method for analyzing overdue accounts receivable, characterized in that, include: The overdue accounts receivable data for the first year are parsed into first structured key-value pair data in the accounts receivable intelligent agent, and the overdue accounts receivable data for the second year are parsed into second structured key-value pair data in the same agent; wherein the first year and the second year are two adjacent calendar years. The accounts receivable intelligent agent is used to perform data parsing, automatic alignment, data calculation, intent recognition, and report generation; The accounts receivable intelligent agent automatically aligns the first structured key-value pair data and the second structured key-value pair data, and calculates the accounts receivable overdue statistical indicators for the first year and the accounts receivable overdue statistical indicators for the second year, respectively. The system acquires and parses the natural language query statements of user requests in the accounts receivable intelligent agent, and determines the type of user request by performing keyword matching on the natural language query statements. Based on the type and the overdue accounts receivable statistics for the first year and the second year, the accounts receivable overdue analysis report is generated by the accounts receivable agent.

2. The accounts receivable overdue analysis method according to claim 1, characterized in that, The step of parsing the overdue accounts receivable data for the first year into a first structured key-value pair data in the accounts receivable intelligent agent, and parsing the overdue accounts receivable data for the second year into a second structured key-value pair data in the accounts receivable intelligent agent, includes: Obtain a spreadsheet file containing overdue accounts receivable data for the first year and overdue accounts receivable data for the second year; The spreadsheet file is parsed using the spreadsheet tool plugin in the accounts receivable agent to obtain the first structured key-value pair data and the second structured key-value pair data.

3. The method for analyzing overdue accounts receivable according to claim 1, characterized in that, The automatic alignment of the first structured key-value pair data and the second structured key-value pair data by the accounts receivable intelligent agent, and the calculation of the accounts receivable overdue statistical indicators for the first year and the second year respectively, includes: Perform column name mapping, unit conversion, company name normalization, and aging consolidation operations on the second structured key-value pair data to achieve automatic alignment between the first and second structured key-value pair data. Based on the first structured key-value pair data, calculate the overdue accounts receivable statistics for the first year; Based on the automatically aligned second structured key-value pair data, calculate the accounts receivable overdue statistics for the second year; Based on the overdue accounts receivable statistics for the first year and the overdue accounts receivable statistics for the second year, calculate the year-on-year change rate between the first and second years.

4. The accounts receivable overdue analysis method according to claim 3, characterized in that, The process of performing column name mapping, unit conversion, company name normalization, and aging consolidation operations on the second structured key-value pair data to achieve automatic alignment of the first and second structured key-value pair data includes: By matching semantic similarity, the original column names of the second year are mapped to the standard column names of the first year; By using a preset conversion factor, the amount in the second year is scaled to the same unit as in the first year; By using a fuzzy matching algorithm, the company name for the second year is standardized to be the same as the name for the first year; By merging intervals, multiple aging segments of accounts receivable in the second year are merged into the same aging segment as in the first year; Using the structure of the first structured key-value pair data as a benchmark, the second structured key-value pair data is automatically aligned.

5. The method for analyzing overdue accounts receivable according to claim 1, characterized in that, The process of acquiring and parsing the natural language query statements of user requests in the accounts receivable intelligent agent, and determining the type of user request by performing keyword matching on the natural language query statements, includes: Obtain the natural language query statement requested by the user; The natural language query statement is segmented to obtain keywords; The keywords are semantically analyzed and compared with preset keywords to obtain matching results; the preset keywords have a corresponding relationship with the type. The type of the user request is determined based on the matching result, and the request is routed to the corresponding path in the accounts receivable agent according to the type.

6. The method for analyzing overdue accounts receivable according to claim 5, characterized in that, The process involves determining the type of the user request based on the matching result and routing it to the corresponding path in the accounts receivable agent according to the type; the types include: mixed type, chart type, single company type, or company-wide report type; wherein: When the natural language query statement contains image and text keywords, the type is determined to be a mixed type, and the query is routed to a mixed path; When the natural language query contains chart-related keywords, the type is determined to be chart type, and the query is routed to the chart path; When the natural language query contains company name keywords, the type is determined to be a single company type, and the query is routed to the single company path. When the natural language query does not contain image / text keywords, chart / graph keywords, or company name keywords, the type is determined to be the company-wide type, and the query is routed to the company-wide reporting path.

7. The accounts receivable overdue analysis method according to claim 6, characterized in that, The step of generating an accounts receivable overdue analysis report through the accounts receivable intelligent agent based on the type and the accounts receivable overdue statistical indicators of the first year and the second year includes: When the type is either company-wide or single-company, a structured markup-formatted analysis report text is generated using a large language model, and the analysis report text is then converted into a document format for output. When the type is a chart type, the chart engine is called to generate a visual chart based on the overdue accounts receivable statistics; When the type is a mixed type, the analysis report text and chart configuration information in a structured markup format are generated synchronously in the same call process through the large language model, and the chart configuration information is embedded in the analysis report text for output.

8. An accounts receivable overdue analysis system, characterized in that, include: Terminal equipment, transmission equipment, and server equipment; wherein the terminal equipment is connected to the server equipment through the transmission equipment; The terminal device is used to acquire files uploaded by users that contain overdue accounts receivable data for the first year and overdue accounts receivable data for the second year, as well as to acquire natural language query statements requested by users. The transmission device is used to transmit the file and the natural language query statement from the terminal device to the server device, and to transmit the overdue accounts receivable analysis report generated by the server device back from the server device to the terminal device; The server device is used to execute the accounts receivable overdue analysis method according to any one of claims 1 to 7.

9. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to run the computer program to perform the accounts receivable overdue analysis method according to any one of claims 1 to 7.

10. A 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 steps of the accounts receivable overdue analysis method according to any one of claims 1 to 7.

Images