Conversational Agent Engine Using API Specifications

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Traditional methods for developing conversational agents are time-consuming and resource-intensive, requiring extensive configuration of dialogue templates and intents, making it difficult to extend them into new domains and efficiently handle customer inquiries.

Innovation Solution

Creating conversational agents based on API specifications, using a conversational agent engine that includes a natural language processor, semantic data graph, and API orchestration model to translate natural language inputs into API operations, allowing businesses to access integrated system functionalities and interact with customers in a domain-specific manner.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional dialogue templates and intents are used to develop conversational agents, then the agents can handle customer inquiries, but the development process becomes time-consuming and resource-intensive

Engineering Contradiction:

Improvecustomer inquiry handling capabilityVSAvoiddevelopment time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent uses API specifications as templates to automatically generate conversational agent configurations. Instead of manually creating dialogue templates and intents, the system copies and adapts API endpoint definitions, request parameters, and response structures to automatically generate the conversational agent's dialogue flow, slot definitions, and intent recognition rules, dramatically reducing development time while maintaining reliability

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent performs preliminary actions by pre-defining API specifications, domain models, and linguistic corpora before conversational agent deployment. These pre-configured elements serve as reusable templates that can be automatically instantiated for multiple domains and use cases, eliminating the need to rebuild conversational agents from scratch for each new application

Inventive Principle:

Principle #10Preliminary action

2Adaptability or versatility

If traditional conversational agents are extended into new domains, then they can serve different industries, but the complexity of reconfiguration increases

Engineering Contradiction:

Improvedomain extension capabilityVSAvoidreconfiguration complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent creates a universal framework where a single conversational agent engine can serve multiple domains and industries. By using API specifications as the foundation for generating domain-specific configurations, the system achieves multi-functionality without requiring separate agent implementations for each domain, reducing reconfiguration complexity while maintaining high adaptability

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent segments the conversational agent into modular components: API specification parsing, domain model generation, linguistic corpus integration, and dialogue flow construction. Each component can be independently configured and reused across different domains, allowing complex domain extensions to be achieved through simple composition of standardized modules rather than complete reconfiguration

Inventive Principle:

Principle #1Segmentation

3Ease of operation

If manual configuration of conversational agents is performed, then custom behaviors can be achieved, but extensive human resources are required

Engineering Contradiction:

Improvecustom behavior configurationVSAvoidhuman resources

Core Design Contradiction:

Ease of operationVSQuantity of substance

Solution Approach 1:

The patent enables self-service by allowing conversational agents to automatically generate their own configurations from API specifications. The system parses API endpoint definitions, extracts request parameters and response structures, and automatically constructs dialogue templates, slot definitions, and intent recognition rules without human intervention, eliminating the need for extensive manual configuration while preserving custom behavior capabilities

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent introduces an intermediary layer consisting of API specifications and domain models that mediate between the conversational agent engine and the target system. This intermediary automatically translates API definitions into conversational agent configurations, serving as a bridge that eliminates the need for manual configuration while ensuring accurate integration with backend systems

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS11042707B2Conversational interface for APIs

Publication Date: 2021.06.22 SALESFORCE INC

AI summary

This disclosure relates to a mechanism to create conversational agents from API specifications based on domain-specific inputs. The conversational agents may provide the functionalities exposed by the underlying API to users engaging with the conversational agent. Thus, the user may execute actions exposed by the API specification using natural language in a conversational, comfortable, and familiar fashion.