Data asset reutillization engine for enhancing LLM agent-based data systems

1 . A method for generating a query response using a data asset reutilization engine, the method comprising:

receiving a first user query from a user via a client;

extracting contextual metadata from the first user query, wherein the contextual metadata includes creation descriptions and artificial intelligence (AI) conversation logs, and wherein the AI conversation logs include a natural language request and a natural language conversation between a second user and a large language model (LLM) agent used to obtain or create a first data asset;

embedding the contextual metadata as a first contextual vector;

making a first determination, using a vector database and the first contextual vector, that the first data asset matches the first user query;

obtaining, in response to the first determination, the first data asset;

generating a first response based on the first data asset; and

causing to display the first response to the user.

2 . The method of claim 1 , wherein prior to generating the first response;

making a second determination that additional processing is needed on the first data asset; and

performing additional processing on the first data asset to generate a processed data asset,

wherein generating the first response is based on the processed data asset, and

wherein the additional processing comprises at least one selected from one of the following: sorting the data asset, making predictions using the data asset, and finding trends using the data asset.

3 . The method of claim 1 , wherein the method further comprises:

making a second determination that the first user query is not in structured query language (SQL) form, wherein the embedding is in response to the second determination;

converting the first user query into SQL form to receive a SQL query; extracting operational metadata from the SQL query;

embedding, prior to the first determination, the operational metadata into the first contextual vector.

4 . The method of claim 3 , wherein the first determination is based upon a similarity score between the first data asset and the first contextual vector.

5 . The method of claim 1 , wherein the method further comprises updating the vector database using the first contextual vector to obtain an updated vector database.

6 . The method of claim 5 , further comprising:

receiving a second user query from the user via the client;

extracting second contextual metadata from the second user query;

embedding the second contextual metadata as a second contextual vector;

making a second determination, using the updated vector database and the second contextual vector, that a second data asset matches the second user query;

obtaining, in response to the second determination, the second data asset;

generating a second response based on the first data asset; and

causing to display the second response to the user.

7 . The method of claim 1 , wherein prior to receiving the first user query:

obtaining at least one data asset;

extracting operational metadata from the at least one data asset;

assigning a task ID to the at least one data asset;

storing the task ID in a task ID index;

extracting contextual data from the at least one data asset;

embedding the contextual metadata and the operational metadata into a third contextual vector; and

storing the third contextual vector in the vector database.

8 . A non-transitory computer readable medium (CRM) comprising computer readable program code, which when executed by a computer processor, enables the computer to perform a method for generating a query response using a data asset reutilization engine, the method comprising:

receiving a first user query from a user via a client;

extracting contextual metadata from the first user query, wherein the contextual metadata includes creation descriptions and artificial intelligence (AI) conversation logs, and wherein the AI conversation logs include a natural language request and a natural language conversation between a second user and a large language model (LLM) agent used to obtain or create a first data asset;

embedding the contextual metadata as a first contextual vector;

making a first determination, using a vector database and the first contextual vector, that the first data asset matches the first user query;

obtaining, in response to the first determination, the first data asset;

generating a first response based on the first data asset; and

causing to display the first response to the user.

9 . The non-transitory CRM of claim 8 , wherein prior to generating the first response;

making a second determination that additional processing is needed on the first data asset; and

performing additional processing on the first data asset.

10 . The non-transitory CRM of claim 8 , wherein the method further comprises:

making a second determination that the first user query is not in structured query language (SQL) form, wherein the embedding is in response to the second determination;

converting the first user query into SQL form to receive a SQL query; and

extracting operational metadata from the SQL query, wherein the operational metadata is embedded in the first contextual vector.

11 . The non-transitory CRM of claim 10 , wherein the first determination is based upon a similarity score between the first data asset and the contextual vector.

12 . The non-transitory CRM of claim 8 , wherein the method further comprises updating the vector database using the contextual vector to obtain an updated vector database.

13 . The non-transitory CRM of claim 12 , further comprising:

receiving a second user query from a user via a client;

extracting contextual metadata from the second user query;

embedding the contextual metadata as a second contextual vector;

making a second determination, using the updated vector database and the second contextual vector, that a second data asset matches the second user query;

obtaining, in response to the second determination, the second data asset;

generating a second response based on the first data asset; and

causing to display the second response to the user.

14 . The non-transitory CRM of claim 8 , wherein prior to receiving the first user query:

obtaining at least one data asset;

extracting operational metadata from the at least one data asset;

assigning a task ID to the at least one data asset;

storing the task ID in a task ID index;

extracting contextual data from the at least one data asset;

embedding the contextual metadata and the operational metadata into a third contextual vector; and

storing the third contextual vector in the vector database.

15 . A system for generating a query response using a data asset reutilization engine, the system comprising:

a computing device programmed to:

receive a first user query from a user via a client;

extract contextual metadata from the first user query, wherein the contextual metadata includes creation descriptions and artificial intelligence (AI) conversation logs, and wherein the AI conversation logs include a natural language request and a natural language conversation between a second user and a large language model (LLM) agent used to obtain or create a first data asset;

embed the contextual metadata as a first contextual vector;

make a first determination, using a vector database and the first contextual vector, that a first data asset matches the first user query;

obtain, in response to the first determination, the first data asset;

generate a first response based on the first data asset; and

cause to display the first response to the user.

16 . The system of claim 15 , wherein prior to generating the first response the computing device is further programmed to:

make a second determination that additional processing is needed on the first data asset; and

perform additional processing on the first data asset.

17 . The system of claim 15 , wherein the computing device is further programmed to:

make a second determination that the first user query is not in structured query language (SQL) form, wherein the embedding is in response to the second determination;

convert the first user query into SQL form to receive a SQL query; and

extract operational metadata from the SQL query, wherein the operational metadata is embedded in the first contextual vector.

18 . The system of claim 15 , wherein prior to receiving the first user query the computing device is further programmed to:

obtain at least one data asset;

extract operational metadata from the at least one data asset;

assign a task ID to the at least one data asset;

store the task ID in a task ID index;

extract contextual data from the at least one data asset;

embed the contextual metadata and the operational metadata into a third contextual vector; and

store the third contextual vector in the vector database.

19 . The system of claim 15 , wherein the first user query comprises unstructured data.