US 12,135,768 B2
Adaptive language model-based solution for interactive security and safety with data privacy
Yogesh Ailawadi, Mountain House, CA (US); Navjot Singh, Chandigarh (IN); Mohit Kumar, Chandigarh (IN); and Ribhav Jain, Chandigarh (IN)
Assigned to AlertEnterprise, Inc., Fremont, CA (US)
Filed by AlertEnterprise, Inc., Fremont, CA (US)
Filed on Mar. 25, 2024, as Appl. No. 18/615,747.
Claims priority of provisional application 63/492,572, filed on Mar. 28, 2023.
Prior Publication US 2024/0330421 A1, Oct. 3, 2024
Int. Cl. G06F 21/31 (2013.01); G06F 8/33 (2018.01)
CPC G06F 21/31 (2013.01) [G06F 8/33 (2013.01)] 18 Claims
OG exemplary drawing
 
1. A method, comprising:
providing an interface to answer natural language user queries based upon real time data generated by live processes;
converting said natural language queries into a machine query syntax; and
providing said machine query syntax to a large language model without sharing underlying data that is used to satisfy the natural language query by serving the data to the end user while, at the same time, masking said data from large language model;
wherein interaction with the large language model is based upon a predefined syntax protocol;
receiving a reply from the large language model in the same syntax; and
using said reply to create an output to be served to the end user and/or to execute a functionality; and
using said large language model to generate a syntax of natural language queries posed by the end user;
wherein when a query is posed the large language model provides a syntax for use on the end user's enterprise data;
wherein to return an actual desired outcome the large language model only generates the syntax from a structure of tables or columns in the enterprise;
wherein the large language model does not know what is inside the tables or columns; and
wherein contents of the tables and columns are processed by an internal application, a large language model API, and an enterprise database, all of which operate synchronously.