A closed-loop NLP query pre-processor and response synthesizer
A patent application on synthesizing accurate, engaging, contextually relevant, and personalized query responses
A closed-loop natural language query pre-processor and response synthesizer architecture accepts natural language queries and dynamically synthesizes query results. The query results may be in the form of data stories.
The architecture identifies, selects, and composes candidate response elements into a coherent and meaningful query result. The architecture also implements an adaptable delivery mechanism that is responsive to connection bandwidth, query source preferences, query source characteristics, and other factors. Feedback from multiple sources adapts the architecture for handling subsequent queries
The architecture implements technical solutions to many difficult technical problems in the field of automatically generating meaningful query responses given extensive and impossible to manually search data stores of potentially relevant information. A few examples of the technical solutions are summarized next. The architecture provides a personalization mechanism for answering questions, responsive, as examples, to: the role and perspective of the person asking the question; timing considerations; context; session history, including prior queries and responses, query and response history from others with similar characteristics to the querying entity, such as other enterprise engineers or managers; and other factors. The architecture may also identify explicitly and implicitly referenced entities in the input query and use the identified entities in its search for candidate response elements.
The architecture also implements query prediction to determine, in advance, likely subsequent queries to follow, given a starting input query or sequence of input queries and contexts.
The architecture understands which metrics, key performance indicators (KPIs), and other data are relevant to the substance of the input query, responsive to configurable ontologies and other models whose content provides a pre-defined context for the substance of the input query. For instance, the context may describe a particular enterprise, its markets, its products, workflows, metrics, and its enterprise activities.
The architecture also identifies the type of question asked in the input query, and correlates the input query and candidate response elements with enterprise activities, targeting, planning, and other goals. The technical solutions in the architecture further identify the time frame of reference in the input query, its positioning within, e.g. a pre-defined fiscal year for the enterprise or competitor enterprises, and other timing data. The architecture responds to enterprise structural data, e.g., organizational structures, and enterprise dynamics to differentiate the query responses.
The technical implementation of the architecture further implements a recommendation engine for suggesting intelligent actions following a session of input queries and query results. The recommendation engine provides the further benefit of encouraging additional interactive sessions via suggestions, questions, and data stories that follow any given query and response.
