Ai based modernization of legacy applications through ontology driven knowledge extraction and relearning

A method for artificial intelligence-based modernization of legacy applications through ontology driven knowledge extraction and relearning, wherein the method is implemented by a processor executing program instructions stored in a memory, the method comprising steps of:

• processing a first input type and a second input type using a pre-configured Large Language Model (LLM) agent, wherein the first and second input types are retrieved via an Input/Output unit from different data sources;
• generating a knowledge graph representing semantics and implemented rules of the legacy application using the first input type and the second input type and first and second summaries of the first input type and second input type;
• performing a relearning operation to determine implemented rules from the first input type using a first prompt type;
• generating a data dictionary comprising variables present in the first input type by executing a multi-step workflow;
• deriving workflows from the first input type and second input type by extracting syntactic and semantic relationships across first input type and second input type;
• and
• employing the knowledge graph, the implemented rules, the workflows and the data dictionary for enabling downstream modernization activities associated with the first input type.

A computer program product comprising:

• a non-transitory computer-readable medium having computer readable program code stored thereon, the computer-readable program code comprising instructions that, when executed by a processor, cause the processor to:
• process a first input type and a second input type using a pre-configured Large Language Model (LLM) agent, wherein the first and second input types are retrieved via an input/output unit from different data sources;
• generate a knowledge graph representing semantics and implemented rules of the legacy application using the first input type and the second input type and first and second summaries of the first input type and second input type;
• perform a relearning operation to determine implemented rules from the first input type using a first prompt type;
• and
• generate a data dictionary from the first input type by executing a multi-step workflow;
• derive workflows from the first input type and second input type by extracting syntactic and semantic relationships across first input type and second input type;
• and
• employ the knowledge graph, the implemented rules, the workflows and the data dictionary for enabling downstream modernization activities associated with the first input type.