Generative Model Routing for Multi-Bot Communication

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In computing environments with multiple bots, user communications that do not unambiguously identify the intended recipient can result in overwhelming responses from all bots, leading to inefficiencies and excessive computing resource usage, especially when generative models are involved.

Innovation Solution

A system that utilizes a generative model to determine the intended recipient of a user communication by constructing prompts that include user input, previous communications, and bot identities, allowing the generative model to identify the appropriate bot to respond.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If each video game bot responds to every user communication, then all bots can potentially provide relevant responses, but the user becomes overwhelmed with responses and computing resources are wasted

Engineering Contradiction:

Improvebot responsivenessVSAvoidcomputing resource consumption

Core Design Contradiction:

Adaptability or versatilityVSLoss of energy

Solution Approach 1:

The system performs preliminary action by having the generative model predict the intended recipient bot before actual response generation occurs. This advance determination allows the system to filter which bots need to generate responses, preventing wasted computing resources on bots that won't respond while ensuring the correct bot is prepared to respond efficiently.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The generative model serves as an intermediary between the user communication and the bot response system. It analyzes the communication content and user context to determine which bot is the intended recipient, acting as a mediator that filters communications before they reach individual bots, thereby reducing unnecessary response generation while maintaining appropriate bot responsiveness.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Loss of energy

If a generative model is used to determine the intended recipient bot, then computing resource consumption is reduced by avoiding unnecessary responses, but additional computing resources are consumed by the generative model itself

Engineering Contradiction:

Improvecomputing resource consumptionVSAvoidgenerative model computing resources

Core Design Contradiction:

Loss of energyVSUse of energy by moving object

Solution Approach 1:

The system extracts the recipient determination function from the individual bot response generation process and places it in a separate generative model. This extraction allows the system to perform recipient identification as a preliminary filtering step, preventing multiple bots from consuming resources to generate responses they wouldn't send, thereby net-reducing total computing resource consumption despite the added generative model overhead.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The generative model performs preliminary analysis of user communications to determine intended recipients before the actual response generation process begins. This preliminary action prevents unnecessary response generation by multiple bots, reducing total computing resource consumption by avoiding the much more expensive operation of generating full bot responses for communications that won't be answered.

Inventive Principle:

Principle #10Preliminary action

3Reliability

If all bots receive and process every user communication, then no relevant responses are missed, but response accuracy decreases due to irrelevant responses from incorrect bots

Engineering Contradiction:

Improveresponse completenessVSAvoidresponse relevance

Core Design Contradiction:

ReliabilityVSMeasurement precision

Solution Approach 1:

The generative model acts as an intermediary that accurately identifies the intended recipient bot by analyzing communication content and user context. This intermediary function improves response relevance by ensuring communications are routed to the correct bot, eliminating irrelevant responses while maintaining response completeness through accurate recipient determination.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system performs preliminary recipient identification using the generative model before routing communications to bots. This preliminary action ensures that only the intended recipient bot processes and responds to each communication, improving response relevance by preventing incorrect bots from generating irrelevant responses while maintaining completeness through accurate routing.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS20250184290A1Determining the intended recipient of undirected utterances in a multi-bot context

Publication Date: 2025.06.05 MICROSOFT TECHNOLOGY LICENSING LLC

AI summary

Technologies described herein relate to a computer-implemented environment that includes multiple bots with which a user can interact. At least one generative model is employed to identify which bot of the multiple bots is to respond to a user communication set forth by the user in the computer-implemented environment.