API Chain Plan Bank Routing for Context-Aware Digital Assistants
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional chatbots rely on predefined skill or intent models, limiting their ability to engage in natural and contextually aware conversations, and face challenges in handling complex user queries that require multiple API interactions without proper routing and planning.
Innovation Solution
A computer-implemented method using a routing engine with a plan bank of chained APIs and generative AI models to identify and execute actions based on user queries, enabling seamless and contextually aware responses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional predefined skill or intent models are used in chatbots, then the system structure is simple and easy to implement, but the ability to engage in natural and contextually aware conversations is limited
Solution Approach 1:
The system segments the complex task of handling user queries into multiple specialized components: a routing engine that directs queries to appropriate tools, a plan bank that stores predefined API chains for specific tasks, and generative AI models that create dynamic action plans. This segmentation allows each component to specialize in specific functions, improving overall adaptability while managing complexity through modular design.
Solution Approach 2:
The plan bank pre-stores multiple API chains and action plans for common tasks and query types. When a user query is received, the system can quickly retrieve and adapt these pre-prepared plans rather than constructing everything from scratch. This preliminary action enables more sophisticated conversation handling while reducing the computational complexity of real-time processing.
2Productivity
If complex user queries requiring multiple API interactions are handled without proper routing and planning, then the system can potentially access more tools, but the system becomes inefficient and requires repetitive user input
Solution Approach 1:
The routing engine continuously monitors the conversation context and the results of API executions, using this feedback to dynamically adjust the action plan. When API calls return results, the system feeds this information back to the generative AI model, which then determines the next appropriate actions. This closed-loop feedback mechanism enables efficient handling of complex multi-step queries without requiring repetitive user input, as the system autonomously tracks progress and coordinates subsequent actions.
Solution Approach 2:
The system employs dynamic action plans that can adapt their structure and content based on the specific query and evolving conversation context. Rather than following rigid predefined workflows, the generative AI model creates flexible action plans that can be modified in real-time based on API results and user responses. This dynamic approach improves productivity by efficiently coordinating multiple API interactions while reducing time loss through automated context retention and plan adjustment.
3Ease of operation
If a routing engine with plan bank and generative AI models is implemented, then the ability to understand and respond to user queries is enhanced, but the device complexity increases
Solution Approach 1:
The routing engine serves as an intermediary layer between the user queries and the complex plan bank infrastructure. It translates user intent into appropriate plan bank queries, selects relevant API chains, and coordinates their execution. This intermediary function shields users from the underlying system complexity while enhancing query understanding and response capability. The routing engine acts as a mediator that manages the complexity of coordinating multiple API chains and generative AI models without exposing this complexity to users.
4Adaptability or versatility
If multiple API chains are chained together for complex tasks, then the functionality and versatility of the digital assistant improve, but the difficulty of detecting and measuring the correct routing path increases
Solution Approach 1:
The plan bank pre-organizes and stores multiple API chains with their dependencies, parameters, and execution contexts defined in advance. Each API chain is structured with clear input-output relationships and associated metadata. This preliminary organization of routing paths makes it easier for the routing engine to detect and measure the correct path by querying the structured plan bank, rather than having to dynamically figure out complex routing logic for multi-API tasks.
Data Source
AI summary
Techniques for routing a user input to an action and associated parameters to generate a response to an utterance using a digital assistant and a plan bank for API chains are disclosed. A system can access, based on an utterance, a plan bank that includes composite tools. Each of the composite tools includes: a chain of tools including: application programming interface (API) tools or other composite tools, and associated logic. The system can generate, by a first generative artificial intelligence model, an action plan including one of the composite tools. The system can execute the action plan to obtain response data by executing actions associated with each of the API tools or the other composite tools. The system can generate, by a second generative artificial intelligence model, a response to the utterance based on the response data and can provide the response to the user.