Systems and methods for ai-assisted generation of data transformation maps

A non-transitory, computer-readable medium storing computer code thereon, the computer code comprising instructions executable by a processor for:

• accessing a message implementation guideline (MIG) specifying a first data format for an electronic document for use within an electronic data interchange (EDI) environment, the MIG comprising information associated with electronic data interchange (EDI) components including segments and data elements;
• processing the MIG with a trained AI-MIG router that is trained to output a predicted MIG class for the MIG, the trained AI-MIG router trained on a corpus of training data, the corpus of training data comprising examples of a plurality MIG classes;
• selecting normalization logic based on the Predicted MIG class for the MIG;
• converting the MIG from a first MIG format to a normalized MIG using the selected normalization logic, wherein the normalized MIG comprises rules text determined from the information included in the MIG and associated with particular fields in a structured format;
• automatically generating a map from the normalized MIG, the map comprising mapping rules for transforming messages formatted according to the MIG to a second data format, wherein generating the map comprises:
• generating a prompt to a large language model (LLM), the prompt including the rules text and requesting that the large language model convert the rules text to respective mapping logic;
• inputting the prompt to the LLM;
• receiving a response from the LLM, the response comprising an AI-generated mapping rule;
• and
• adding the AI-generated mapping rule to the map.

An electronic document interchange (EDI) system comprising:

• computer systems of a plurality of entities;
• a data transformation computer system coupled to the computer systems of the plurality of entities, the data transformation computer system comprising:
• a data store storing a plurality of data transformation maps, the plurality of data transformation maps automatically generated from message implementation guidelines (MIGs) of the plurality of entities, the data transformation maps comprising data mapping rules for mapping from EDI data formats specified in the MIGs to a canonical data format;
• a processor;
• a non-transitory, computer-readable medium storing computer code thereon, the computer code comprising instructions executable by the processor for:
• receiving a first MIG and an EDI message from a first entity of the plurality of entities, the EDI message destined for a second entity of the plurality of entities, the EDI message including a document of a first document type;
• determining a particular format corresponding to the first MIG based on processing the first MIG with a trained AI-MIG router, wherein the AI-MIG router is trained to classify MIGs based on formats of the MIGs to be classified, the trained AI-MIG router being trained on a corpus of training data comprising examples of a plurality MIG classes associated with a plurality of known formats corresponding to a plurality of MIG editing tools;
• in response to determining that the particular format of the first MIG corresponds to another format other than the plurality of known formats:
• processing the first MIG with an AI-MIG converter to generate a normalized first MIG, wherein the AI-MIG converter is trained to convert new MIG formats to a normalized MIG format, and wherein the normalized first MIG comprises first rules text determined from business logic or first processing data logic included in the first MIG and associated with particular fields in a structured format;
• generating a new map for the first entity from the normalized first MIG, wherein generating the new map for the first entity comprises:
• generating a prompt to a large language model (LLM), the prompt including the first rules text and requesting that the LLM convert the first rules text to respective first AI-generated mapping logic;
• inputting the prompt to the LLM;
• receiving the respective first AI-generated mapping logic from the LLM;
• and
• adding a first mapping rule to the new map for the first entity, the first mapping rule comprising the respective first AI-generated mapping logic;
• selecting a second data transformation map from the plurality of data transformation maps, the second data transformation map associated with the second entity and the first document type and including second data mapping rules, the second data mapping rules comprising a second AI-generated data mapping rule;
• executing the first data mapping rules to transform message data from the EDI message from a first EDI message format to the canonical data format;
• executing the second data mapping rules to transform the message data from the canonical data format to a second EDI message format;
• and
• forwarding the message data to the second entity using the second EDI message format.