Customer information system integrating asynchronous knowledge processing and dynamic scheduling

Abstract

This invention provides a customer information system integrating asynchronous knowledge processing and dynamic scheduling, comprising a server and a user device communicating with the server. By employing an asynchronous processing mechanism that publishes input information to a task queue, this invention enables the server to immediately return an acceptance status code upon receiving the information, effectively avoiding network connection timeouts and system congestion caused by prolonged vector operations. Furthermore, by integrating standardized documents and vector indexing technology into the dynamic scheduling program, it also solves the problem of traditional scheduling being decoupled from the knowledge base, achieving a highly efficient, knowledge-based automated customer information system.

Claims

1. A customer information system integrating asynchronous knowledge processing and dynamic scheduling, comprising: a user device that generates input information; a server comprising: a network module communicatively connected to the user device to receive the input information; a front-end processing module connected to the network module; a back-end processing module connected to the front-end processing module; and a storage module connected to the front-end processing module and the back-end processing module, storing a task queue and a vector database; wherein, When the front-end processing module receives the input information through the network module, it publishes the input information to the task queue, generates acceptance status information, and transmits an acceptance status code corresponding to the acceptance status information to the user device through the network module. The back-end processing module retrieves the input information from the task queue and performs a knowledge vectorization process based on the input information to generate multiple text blocks and corresponding high-dimensional semantic vectors. After the back-end processing module finishes the knowledge vectorization process, it stores the text blocks and high-dimensional semantic vectors in the vector database of the storage module, and the front-end processing module updates the acceptance status information. When the user device receives the updated acceptance status information, the user device refreshes its screen. When the input information contains a scheduling intent, the back-end processing module executes a scheduling program and compares existing schedules in a calendar according to an application interface. The back-end processing module performs a conflict check and transmits a suggested time period to the user device through the front-end processing module and the network module.

2. The customer information system integrating asynchronous knowledge processing and dynamic scheduling as described in Request 1, wherein, Before the front-end processing module is published to the task queue, the input information is stored as a raw document in the storage module.

3. The customer information system integrating asynchronous knowledge processing and dynamic scheduling as described in Request 1, wherein, The front-end processing module first verifies the input information, and only stores the input information after the front-end processing module determines that the input information is compliant.

4. The customer information system integrating asynchronous knowledge processing and dynamic scheduling as described in Request 1, wherein, This acceptance status code is an HTTP response status code.

5. The customer information system integrating asynchronous knowledge processing and dynamic scheduling as described in Request 1, wherein, When the user device sends a confirmation message to the server according to the suggested time period, the back-end processing module updates the existing schedule of the calendar through the application interface.

6. The customer information system integrating asynchronous knowledge processing and dynamic scheduling as described in Request 1, wherein, When the input information received by the backend processing module is a user query, the backend processing module converts the user query into a semantic vector, retrieves the associated text block from the vector database, and combines the dialogue history to generate feedback information through a large language model (LLM).

7. A customer information system integrating asynchronous knowledge processing and dynamic scheduling as described in claim 6, wherein, After the user device finishes sending the user query, the backend processing module performs natural language understanding (NLU) analysis on the user query and extracts feature data containing emotion, intent and keywords.

Images