System and method for graph-based conversational-flow editing

A computer-implemented method for controlling a conversational agent modification system using natural-language input, the method comprising:

• receiving a natural-language instruction from the user describing a desired change relating to the conversational agent;
• interpreting the natural-language instruction to determine at least one operation to be performed on a stored representation of the conversational agent and one or more parameters for the at least one operation;
• generating, based on the determined at least one operation and the one or more parameters, at least one structured operation request identifying a function selected from a plurality of predefined functions and the one or more parameters for the function;
• executing the identified function based on the one or more parameters for the function to modify the stored representation of the conversational agent, wherein the executing the identified function comprises:
• treating the at least one structured operation request as a function call;
• encapsulating the function call as a job;
• placing the job into an asynchronous job queue for processing;
• and
• executing the identified function based on the asynchronous job queue;
• receiving an execution result indicating a modification to the stored representation of the conversational agent;
• and
• updating a graphical sequence builder canvas displayed to the user based on the modification, by:
• determining, using a layout routine, updated coordinates for at least one of a new, a deleted, and a modified node and/or edge in a visual graph representation of the conversational agent,
• re-rendering, in real-time, the graphical sequence builder canvas based on the updated coordinates.