Entity management for language models

A method comprising:

• receiving a first natural language request comprising a reference to a first entity;
• determining, in response to receiving the first natural language request, first resolved entity data comprising structured metadata indicating a plurality of attributes of the first entity;
• generating first identifier data associated with the first resolved entity data;
• storing, in a session cache, the first resolved entity data in association with the first identifier data;
• generating a first prompt based on the first natural language request, wherein the first prompt comprises the first identifier data;
• generating, based on the first prompt, a data request applicable to the first natural language request, wherein the data request comprises the first identifier data as an input parameter;
• receiving at least a portion of the structured metadata of the first resolved entity data from the session cache using the data request;
• generating a second prompt comprising at least the portion of the structured metadata of the first resolved entity data;
• and
• performing, based on the second prompt, a first action with respect to the first entity in response to the first natural language request.

A system comprising:

• at least one processor;
• and
• non-transitory computer-readable memory storing instructions that, when executed by the at least one processor, are effective to:
• receive a first natural language request comprising a reference to a first entity;
• determine, in response to receiving the first natural language request, first resolved entity data comprising structured metadata indicating a plurality of attributes of the first entity;
• generate first identifier data associated with the first resolved entity data;
• store, in a session cache, the first resolved entity data in association with the first identifier data;
• generate a first prompt based on the first natural language request, wherein the first prompt comprises the first identifier data;
• generate, based on the first prompt, a data request applicable to the first natural language request, wherein the data request comprises the first identifier data as an input parameter;
• receive at least a portion of the structured metadata of the first resolved entity data from the session cache using the data request;
• generate a second prompt comprising at least the portion of the structured metadata of the first resolved entity data;
• and
• perform, based on the second prompt, a first action with respect to the first entity in response to the first natural language request.