# US 2026/0093574 A1 · Intelligent test data troubleshooting in erp systems > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2026/0093574 A1 - Document ID: US20260093574A1 - Status: Published application - Kind: A1 - Application number: 18/902,467 - Filed: 2024-09-30 - Published: 2026-04-02 - Earliest claimed: 2024-09-30 - Assignees: Sap SE - Applicants: Sap SE - Inventors: Prashant Telkar; Shubham Saklani - Web page: https://patentgrep.com/patent/US20260093574A1 ## CPC - G06F 11/0793: {Remedial or corrective actions (recovery from an exception in an instruction pipeline G06F9/3861; by retry G06F11/1402; for recovering from a failure of a protocol instance or entity H04L69/40)} - G06F 11/0766: {Error or fault reporting or storing} - G06F 16/3347: {using vector based model} - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) ## Abstract A computer-implemented method for improving test data troubleshooting in an enterprise resource planning (ERP) system is disclosed. The method can receive an error message generated by an application programming interface of the ERP system when processing a set of test data, embed the error message into a message vector, search a vector database comprising a plurality of object vectors to identify a target object vector matching the message vector, retrieve a target object document associated with the target object vector, and generate an answer including instructions to modify the set of test data based on the error message and the target object document. Related systems and software for implementing the method are also disclosed. ## Claims ### Claim 1 A computing system for improving test data troubleshooting in an enterprise resource planning (ERP) system, the computing system comprising: memory; one or more hardware processors coupled to the memory; and one or more computer readable storage media storing instructions that, when loaded into the memory, cause the one or more hardware processors to perform operations comprising: receiving an error message generated by an application of the ERP system when processing a set of test data; embedding the error message into a message vector; searching a vector database comprising a plurality of object vectors, wherein the searching identifies, among the plurality of object vectors, a target object vector matching the message vector; retrieving a target object document associated with the target object vector, wherein the target object document contains text descriptions of the error message; and generating an answer including instructions to modify the set of test data based on the error message and the target object document. ### Claim 2 (depends on 1) The computing system of claim 1, wherein searching the vector database comprises measuring similarities between the message vector and the plurality of object vectors. ### Claim 3 (depends on 2) The computing system of claim 2, wherein searching the vector database further comprises determining a maximum of the measured similarities. ### Claim 4 (depends on 1) The computing system of claim 1, wherein generating the answer comprises generating a prompt based on a prompt template, wherein the prompt template includes a first placeholder for receiving the error message and a second placeholder for receiving context from the target object document. ### Claim 5 (depends on 4) The computing system of claim 4, wherein generating the answer further comprises prompting a large language model using the prompt. ### Claim 6 (depends on 1) The computing system of claim 1, wherein the operations further comprise creating the vector database based on a group of object documents. ### Claim 7 (depends on 6) The computing system of claim 6, wherein creating the vector database comprises dividing the group of object documents into a plurality of text segments and embedding the plurality of text segments into the plurality of object vectors, respectively. ### Claim 8 (depends on 6) The computing system of claim 6, wherein creating the vector database comprises filtering a document corpus using a set of rules to identify the group of object documents. ### Claim 9 (depends on 6) The computing system of claim 6, wherein the operations further comprise periodically detecting whether there is an update to the group of object documents. ### Claim 10 (depends on 9) The computing system of claim 9, wherein the operations further comprise updating the vector database responsive to detecting that the group of object documents have had the update. ### Claim 11 A computer-implemented method for improving test data troubleshooting in an enterprise resource planning (ERP) system, the method comprising: receiving an error message generated by an application of the ERP system when processing a set of test data; embedding the error message into a message vector; searching a vector database comprising a plurality of object vectors, wherein the searching identifies, among the plurality of object vectors, a target object vector matching the message vector; retrieving a target object document associated with the target object vector, wherein the target object document contains text descriptions of the error message; and generating an answer including instructions to modify the set of test data based on the error message and the target object document. ### Claim 12 (depends on 11) The computer-implemented method of claim 11, wherein searching the vector database comprises measuring similarities between the message vector and the plurality of object vectors. ### Claim 13 (depends on 12) The computer-implemented method of claim 12, wherein searching the vector database further comprises determining a maximum of the measured similarities. ### Claim 14 (depends on 11) The computer-implemented method of claim 11, wherein generating the answer comprises generating a prompt based on a prompt template, wherein the prompt template includes a first placeholder for receiving the error message and a second placeholder for receiving context from the target object document, and prompting a large language model using the prompt. ### Claim 15 (depends on 11) The computer-implemented method of claim 11, further comprising creating the vector database based on a group of object documents. ### Claim 16 (depends on 15) The computer-implemented method of claim 15, wherein creating the vector database comprises dividing the group of object documents into a plurality of text segments and embedding the plurality of text segments into the plurality of object vectors, respectively. ### Claim 17 (depends on 15) The computer-implemented method of claim 15, wherein creating the vector database comprises filtering a document corpus using a set of rules to identify the group of object documents. ### Claim 18 (depends on 15) The computer-implemented method of claim 15, further comprising periodically detecting whether there is an update to the group of object documents. ### Claim 19 (depends on 18) The computer-implemented method of claim 18, further comprising updating the vector database responsive to detecting that the group of object documents have had the update. ### Claim 20 One or more non-transitory computer-readable media having encoded thereon computer-executable instructions causing one or more processors to perform a method for improving test data troubleshooting in an enterprise resource planning (ERP) system, the method comprising: receiving an error message generated by an application of the ERP system when processing a set of test data; embedding the error message into a message vector; searching a vector database comprising a plurality of object vectors, wherein the searching identifies, among the plurality of object vectors, a target object vector matching the message vector; retrieving a target object document associated with the target object vector, wherein the target object document contains text descriptions of the error message; and generating an answer including instructions to modify the set of test data based on the error message and the target object document.