| CPC G06Q 30/0203 (2013.01) [G06F 40/166 (2020.01); G06F 40/279 (2020.01); G06F 40/40 (2020.01)] | 23 Claims |
|
1. A method for adaptively customizing a user interaction at an artificial intelligence engine operating on a large language model, the method comprising:
at a first processor comprising a researcher natural language user interface operating a researcher application programming interface (“API”):
receiving, from a researcher, at the researcher API, an instruction to instantiate an instance of a survey;
communicating the instruction from the researcher API to the large language model, a database and an object storage;
instantiating the instance of the survey for the researcher at the large language model; and
receiving, from the researcher, in natural language via the researcher natural language user interface, a set of survey question guardrails for the survey, said set of survey question guardrails constructed in natural language, said set of survey question guardrails comprising one or more parameters comprising descriptions, examples, excluded topics, excluded questions, length limitations, instructions how to manage specific responses by topic, parameters specific to a first survey question and parameters specific to a subsequent survey question; and
inputting the set of survey question guardrails constructed in natural language as prompts to the large language model via the researcher API;
at a second processor comprising the artificial intelligence engine operating on the large language model, said artificial intelligence engine in communication with the researcher natural language user interface:
generating a discrete instance of the survey for each survey-taker included in a plurality of survey-takers;
communicating the first survey question to each survey-taker via one or more output devices associated with a processor associated with each survey-taker via a participant application programming interface, said participant application programming interface in communication with the database, the large language model and the object storage;
receiving a first natural language response from each survey-taker via one or more input devices associated with the processor associated with each survey-taker;
auto-formatting the first natural language response into a data element prompt, said data element prompt being constructed in natural language, said data element prompt conforming to artificial intelligence engine input parameters;
adaptively generating the subsequent survey question, said subsequent survey question responsive to the first natural language response and constrained by the set of survey question guardrails;
communicating the subsequent survey question to each survey-taker via the one or more output devices associated with the processor associated with each survey-taker;
receiving a subsequent natural language response from each survey-taker via the one or more input devices associated with the processor associated with each survey-taker;
auto-formatting the subsequent natural language response into a second data element prompt, said second data element prompt being constructed in natural language, said second data element prompt conforming to the artificial intelligence engine input parameters;
in response to receipt, at the large language model, of a predetermined trigger constructed in structured computer code, auto-terminating the survey for each survey-taker;
at the first processor:
extracting a plurality of data elements from a processed survey set, said processed survey set comprising a set of survey questions communicated to the plurality of survey-takers and a set of survey responses received from the plurality of survey-takers, said plurality of survey-takers comprising each survey-taker, said set of survey questions comprising the first survey question and the subsequent survey question and said set of survey responses comprising the first natural language response and the subsequent natural language response;
processing the plurality of data elements into a set of organized information; and
auto-generating a report comprising the set of organized information.
|